AD-777 747 ARPANET MANAGEMENT STUDY Paul Baran et el Cabiedata Associates Incorporated Prepared for Advanced Research Projects Agency 20 January 1974 DISTRIBUTED BY National Technical Information Service U S DEPARTMENT OF COMMERCE 5285 Port Royal Road Springfield Va 22151 ' ' -v ' - J l V -i - i - -- Vrt iv - -' THIS DOCUMENT IS BEST QUALITY AVAILABLE THE COPY FURNISHED TO DTIC CONTAINED A SIGNIFICANT NUMBER OF PAGES WHICH DO NOT REPRODUCE LEGIBLYo UNCLASSIFIED SECURITY CLASSIFICATION OF TMIS PAGE f kmn Dim r mmrmd READ INSTRUCTIONS BEFORE COMPLETING FORM REPORT DOCUMENTATION PAGE I COVT ACCESSION O RECIPIENT'S CATALOG NUMim E OIT NUMBER DI-S-3591 S-117-1 M o TITLE i Sumtllltt AP PANET MANAGEMENT STUDY TYPE OF c RT pr oo COVERED Final Technical Report 1973 March-December t PERFORMING t-IG REPORT NUd'ER R-123 7 AuTMORf Paul Baran David C Caulkins Vinton G Cerf Ronald C Crane CONTRACT OR GRANT NUMBERf Paul Goldstein Robert S O'Brien Edwin B Parker Marc U Porat Contract Number FO8606-73-C-0036 10 PROGRAM ELEMENT PROJECT AREA WORK UNIT NUMBERS PERFORMING ORGANIZATION NAtjE AHO AOORESS 'ASK Cabledata Associates Inc 701 Welch Road Palo Alte CA 94304 12 REPORT DATE II CONTROLLING OFFICE NAME AHO AODRES4 1974 January 20 Defense Advanced Research Projects Agency 1400 Wilson Boulevard Arlington VA 22209 IS NUMBER OF PAGES U IS SECURITY CLASS ol thlt nport 312 MONITORING AGENCY NAME A ADORES dUlormnt ttom Controlling Ollle Range Measurements Laboratory Building 981 Air Force Eastern Test Range Patrick Air Force Bast FL 32925 Unclassified IS OECLASSIFlCATION'OOWNGRADttIG SCHEDULE It DISTRIBUTION STATEMENT ol Mi Ropotl Distribution of this document is unlimited 17 DISTRIBUTION STATEMENT el Wit mbttrmct anl i In Sleek 3 II dlllormnl horn Ropati Distribution of this document is unlimited ' SUPPLEMENTARY NOTES Reproduced by None NATIONAL TECHNICAL INFORMATION SERVICE U S Department of Commerce Springfield VA 2 151 I f V VOROS Ccnllnu on o oo Id II n e ry mid ivenllty by block nunbar ARPANET Communications Computer Netting Divestiture Packet Switching Regulation 10 AUSTRAGT Contlnno en rvror olde II nocottmnr mnl IdtntH by block nuntierj Examines isSUeS affecting th long range development of the ARPANET in general Considers the near-term question of possible divestiture of the ARPPwvr specific and reviews the background of Proposes specific steps to meet t the network to better describe present context primary ARPA objectives for the future Basically r lls for encouragement of interaction and cooperation by organizations providing packet switching services If successful inter-network connection could become economically and operationally feasible this would permit ARPA to divest on incremental test basis these portions of the network not required for experimental purposes DD 'Zn 1473 COITION OF 1 NOV oo It OBSOLETE UNCLASSIFIED SECURITY LAl F CATION OF THIS PACE hon Dim Kil r d CABLEDATA ASSOCIATES INC 701 Welch Road ' Palo Alto California 94304 ARPANET MANAGEMENT STUDY FINAL TECHNICAL REPORT DI-S-3591 S-117-1M This research was supported by the Advanced Research Projects Agency of the Department of Defense and was monitored by the Range Measurements Laboratory under Contract No FO8606-73-C-0036 The views and conclusions contained in this document are those of the authors and should not be interpreted as necessarilv representing the official r -ilicies either expressed or implied of the Advanced Research trojects Agency of the U S Government ARPA ORDER NUMBER 2317 PROGRAM CODE NUMBER 627O6E 3P10 NAME OF CONTACTOR Cabledata Associates Inc EFFECTIVE DATE OF CONTRACT 1973 April 30 CONTRACT EXPIRATION DATE 1974 January 24 AMOUNT OF CONTRACT $118 965 00 CONTRACT NUMBER FO8606-73-C-0036 PRINCIPAL INVESTIGATOR HONE NC Paul Baran - 415 328-241' PROJECT SCIENTIST PHONE Vinton G Cerf - 415 328-2411 SHORT TITLE OF WORK ARPANET Management Study DATE OF REPORT 1974 January 14 CONTRACT PERIOD COVERED BY REPORT 1973 March - December This document approved for public release a Tc' sale distribution is unlimited SUMMARY AND CONCLUSION This study examines some of the issues affecting the long range development of packet switching in general It specifically considers the role of the ARPANET in this development The near- t rm question of transfer of a portion of the communications portion of the ARPANET is examined in detail and the background of the development cf the network is discussed to place the key issues in perspective A plan for the transfer of the communications portion cf the ARPANET and not the host resources is proposed which is based upon the encouragement of interaction and cooperation among organizations providing pocket switching services If successful effective internetwork connection could become economically and operationally feasible This would permit ARPA the option of transferring on an incremental test basis the portion of the network not required for exnerimental purposes In the proposed plan described in Sections 3 and 4 of the report 1 ARPA would transfer the service aspects of the network r ot needed to carry out experimentb in packet switching technology 2 ARPA would retain or create an experimental packet switching subnet on which it would test satellite communications methods packet radio network interconnection methods and other ideas The transfer itself would provide new means for sharing ox thi packet switching subnet between private and public sectors urdei the aegis of a consuitiua in a l gal and harmonicas fashion with minimal need for FCC control The initial reasons behind the development of the ARPANET remain valid today Those reasons include 1 The desire to use ARPA-owned or -funded resources more effectively resource sharing 2 The desire to obtain low cost computer communication facilities necessary for resource sharing This requires both high bandwidth e g file transfer and low delay e g interactive traffic to be serviced 3 An interest in applying experimental packet switching techniques to communication development to overcome limitations of conventional data c- ounications high error rates low bandwidths inflexible topologies and limited reliability 4 The need to develop alternatives for military communication systems having lower cost lower delay and hijhev bandvidth capabilities than those currently in use while still providing the end-to-end securitv and reliability needed The ARPANET project has been successful in several ways 1 The technical feasibility of packet switching for termii al-to-computer and computer-to-con puter communication has bten demonstrated at marginal costs lower than any present alternative 2 Common protocols which allow diverse host computers to communicate with one another have been designed and implemented at virtually all sites in the network The network has provided a good test bed for exploring solutions to problems of interprocess communication distributed operating system design interfacing diverse operating systems security and privacy accounting and reliability 3 Effective sharing of the network's resources among users and host computers has been achieved This sharing has permitted closer interaction among researchers in the network community made better use of limited computer resources and has demonrtrated new capabilities in computer science and project management 4 Research into new communication methods based on packet switching e g packet radio packet broadcast satellite is now under way largely spurred on by the initial success of the ARPANET packet switching experiment After reviewing the status of the ARPANET we then considered a set of major issues now facing the network r These issues included 1 The continuing need to provide ARPANET-based services to ARPA contractors on a high reliability basis li J 2 The need for similar services by other governmental agencies 3 The desire for ARPANET-type services by the civilian sector 4 The desirability and problems of interconnection with other national and international networks 5 The proper role of ARPA as a research organization committed to the concept of technological transfer when research matures into proven feasibility 6 The on-going role of ARPA in developing the computer resource sharing concept In this study we have reached a number of conclusions which we state below as recommendations 1 We recommend that the commercial packet switching industry be encouraged to provide the additional capacity that ARPA and new governmental applications will be seeking from ths present ARPANET rather than permit an open-ended expansion of the ARPANET communications network 2 We recommend that ARPA continue full ownership and control of those parts of the ARPANET needed for experimentation in improvements in packet switching 3 We recommend that the nation develop a unif led packet switching service accessible to all users on an equitable basis rather than encourage a collection of isolated packet networks that cannot share specialized computer resources 4 We believe that more effective use of limited national communicatior s resources would occur if all packet networks were built so as to permit interconnections with one another and recomnand that it be encouraged by ARPA 5 We belxeve that the healthy development of the packet switching industry will be of significant importance to the development of the computer resource sharing capability of the country and recommend that it be encouraged 6 We believe that the transfer of the ARPANET communications facilities should not as a matter of public policy lead to the creation of any monopoly on future packet sv itching by any potential bidder To this end we recommend that new means be created t o permit the suppliers of packet switching to work together to create and maintain a healthy competitive environment while supplying competitive services in 7 Inasmuch as no presently suitable arrangement exists for accomplishing these objectives we recommend the specific plan which is deferibed in detail in this report This plan is based upon the formation of an industry group or consortium The form and name of the institutional structure is secondary provided that it contains effective provision for the following three essential functions A A clearinghouse mechanism for transferring payments among cooperating entities B A mechanism for creating and enforcing common industry standards C A mechanism to allow continuously free and open entry to avoid formation of any closed oligopolistic structure that will demand close governmental supervision or regulation iV CONTENTS SUMMARY Section 1 2 INTRODUCTION Contents Goals Emphasis Host Economics ' PROBLEM STAT-MEWi Content Initial Objectives Status Services vs Experimentation Conflict between Service and Network Experimentation Needs Separating Resource Sharing and Switching Experiments Charter Restrictions Demand for ARPANET Services Nature of the Demand Experienced Demand Expected Demand Reduction of Management Burden The Problem Interim Administration Commercial Interest in Packet Switching 3 4 r 1 1 4 5 5 7 7 7 8 8 8 9 9 10 10 10 1C 11 11 11 11 TOWARDS A RECOMMENDATION General Policy Statements A Proposed Transitional Transfer Strategy The Concept of the Packet Switching Industry CoopeT cior Arrangement New Options Establishing the Consortium 14 14 15 SPECIFIC PROPOSED PLAN FO A PARTIAL TEST TRANSFER Request for Industry Comments Purpose of Disposition Request for Comments - Operations Range of Comments Desired Importance of Industry-Group Arrangements 19 19 19 22 27 28 15 16 17 Appendix on Legal Issues A LEGAL BACKGROUND ANALYSIS Appendices on Institutional Issues - Preface B INSTITUTIONAL ALTERNATIVES C ON FORMATION OF A COMMON INTEREST CONSORTIUM OF PACKET SWITCHING ENTITIES D A DELPHI EXERCISE EXAMINING FOUR ALTERNATIVE COMMUNICATIONS POLICY OPTIONS E INDEPENDENT NODES ECONOMICS SIMULATION MODEL F USER'S MANUAL TO INEC INDEPENDENT NODES ECONOMICS SIMULATION MODEL Appendices on Facilities Economic Issues - Preface G ARPANET INVENTORY LISTING PROGRAM RCNA H ARPANET INVENTORY I FINANCIAL ANALYSIS OF THE ARPANET PROGRAM RON 1 RON 2 J USING THE PROGRAM RON i FOR COST AND VALUE DETERMINATION VI j INTRODUCTION CONTENTS This report summarizes a ten-month study on aspects of the future evolutionary development of the ARPANET In specific we consider the question of transfer of ownership of the communications portions of the ARPANET to meet the growing demand for these types of services The introductory section of this report describes the goals of this study the emphasis and purpose the methodology used and the location of the component sections of this study Section 2 of this report considers the near-term question of transfer reviews the initial objectives of the ARPANET considers its present status and differentiates between the conflicting network needs for providing reliable services and for providing a vehicle for experimentation Section 2 also con- siders ARPA's charter in providing policy guidance as to allowable directions of alternative policies The question of in- creasing demand for ARPANET services is next discussed and is followed by a consideration of the desirability of reducing the management burden to ARPA Lastly in this section is a discus- sion of commercial interest in packet switching and some of the implications of the expected availability of this new capability Section 3 of this report starts with some of the basic postulates underlying a possible proposed course of action A proposed transitional divestiture strategy is suggested based heavily upon the concept of a packet switching industry cooperative arrangement This approach opens some new options to ARPA which are considered Because the idea of a consortium is -1- relatively new it is described briefly in Section 3 and in very much greater detail in Appendices E through F The last section of the report Section 4 proposes a specific action plan for a partial test transfer of ownership of parts of the ARPANET under tight control to protect other users of the existing network This section of the report is not intended as a take-it-or-leave-it proposition Rather it is intended as a proposed plan that would benefit from industry review and modification where necessary to provide an eventual plan that all parties would find acceptable in the belief that it would accomplish its desired objectives Section 4 is written in question and answer form to facilitate the reader's skipping over sections of detail that are only of minor interest We sought to keep the basic report shc-t If the reader has little time reviewing the summary will tell much of the story If the reader has a little more time then the full report can be read as it is only about 30 pages in length However much of the report resides in the appendices But these are arranged to be read in stand-alone fashion The appendices of this report are arranged in three groups with yellow divider sections used to facilitate the reader in pinpointing individual sections of interest Blue dividers are used to separate the appendices The first Appendix A was prepared by Paul Goldstein It stands alone and provides legal background analysis important to the question of the transfer o the ARPANET This appendix shows among other matters that governmental regulation is a substitute for competition and is needed only where the open marketplace is unable to achieve effective competition This appendix also views regulation and non-regulation not as binary concepts but as shades of gray issues Some of these fine shadings are of importance to the question of transfer of ownership of government resources to private ownership present in possible transfer of the ARPANET -2- The next group cf appendices Appendices B through F relate to institutional alternatives and were prepared primarily by Marc U Porat The first Appendix B reviews alternative industrial structures possible describes the present trajectory of development towards one of these possibilities as being most likely unless active reconsideration is taken and describes the writer's preferred course of action together with the reasons for his position In support of his arguments useful background infor- mation is presented reviewing some of the most recent changes taking place in communications regulation Appendix C is a detailed description of the operation of a possible consortium or industry association entities including suppliers and users of packet switching This provides fine qrain detail discussion of the day-to-day procedural operation of an imaginary consortium It provides a flavor of how such an imaginary organization might work Much detail is included in this report since the concept of a consortium has not been considered before in this application The detail is intended to aid discussion about possible organizational arrangements The third appendix in this series Appendix D is a Delphi exercise prepared early in this study In this the staff con- sidered a spectrum of alternative options narrowed them down to four major ones and then expressed their subjective judgments Considered were differences in the characteristics and expected operational behavior This appendix suggests the broad range of alternative institutional arrangements initially considered and some of the reasons why certain arrangements were narrowed for further investigation Appendix E is a simulation also performed early in the project It examines the expected behavior patterns of competing organizations in the hypothetical situation where such organizations owned different segments of a single network and where strict rules of behavior specified in advance were followed -3- This appendix addresses the question as to whether actions beneficial to tne entire network would result if each separate owner made decisions solely in its own best interest The appendix shows how one might go about programming this behavior to predict performance in advance of a real world situation Lastly in this set of appendices on institutional arrangements is Appendix F which is the Users Manual for the simulation model The last four appendices G through J relating to ARPANET communications facilities economic issues were prepared by Ronald C Crane They describe a cost model structure for estimating the costs involved in the ARPANET and provide a do-it-yourself kit of tools and a data base They allow the use to consider any combination of ARPANET elements in place at any point in time producing output analyses under a wide set of depreciation assumptions and costing bases GOALS Below we list the factors that form the impetus and goals of this study 1 ARPA is facing major decision - during the next few years on the growth development and possible ownership of the ARPANET There are major implications for state-of-the-art of computer system development that hinge upon some of these decisions 2 Much of the work in this project consisted of detailed consideration of the alternatives viewed from different dimensions including technical desirability regulatory constraints management effectiveness legal constraints economic factors and the specific impacts upon each affected community of interest 3 The basic goal sought is that of creating a specific plan of action that will provide the greatest long range benefit to the nation in return for its past present and future involvement and investment in the ARPANET 4 All policy decisions that result from this effort will be the responsibility of ARPA IPT This work is intended only to provide ARPA with background policy research to facilitate its policy making role -4- EMPHASIS Unlike most research studies we sought not a single correct answer but rather we explored a large set of alternatives each alternative considered has been evaluated but the final choice is left to the sponsor of the study In any examination of this type the ideal degree of detail is almost open-ended and probably depends more upon the amount of time that the researcher has available than any other factor Thus at the initiation of this study we prepared a systematic effort tree or effort-weighted outline of the dimensions of the issues that form the context of the topic We questioned the sponsor the Information Processing Techniques Office of the Defense Advanced Research Projects Agency as to what it fsit were the most important topics to be considered and how should the limited effort best be expended This procedure provided a formal structure for selecting a part of J broad initial menu for narrower analysis The narrowing down of the menu was performed in discussions with Dr Lawrence G Roberts In brief emphasis was to be given primarily to those questions that related directly to transfer and how it might be accomplished HOST ECONOMICS Although the resources represented in the ARPANET are primarily in the host computer installations we were specifically asked not to consider the economics of the ho-st installations in order to provide emphasis to the communiccvions network matters per se This was done as only the network itself was being considered for divestiture at that time This limita- tion was very important because the investment in the host computers in connection with the ARPANET is much greater than the cost of the communications network subsystem considered in detail in this study Of course some examination of host economics was investigated because the cost for communications -5- I processing within ths TENEX operating system appears in excess of other communications cost Because of this disparity tl e reader should be cautioned that detailed attention should be given to the economics of the host computers in estimating total costs Host costs posed a bit of a problem because the actual use of the host computers is neither rigorously monitored nor in cur opinion adequately understood -6- 2 PROBLEM STATEMENT CONTENT This section of the report considers the near-term question of possible transfer of the ARPANET and reviews the background of the network to better describe its present context The following section proposes steps to meet this ARPA objective This report proposes a plan which calls for the encouragement of interaction and cooperation by organizations providing packet switching services with the goal that widespread internetwork connection could become economically and operationally feasible This would in turn allow ARPA to transfer on an incremental test basis those portions of its network not required for experimental purposes and thus allow shared internetwork usage INITIAL OBJECTIVES It is helpful to review the initial reasons behind the development of the ARPANET so that the longer term goals are kept in mind The initial impetus came from several directions including 1 The desire to use ARPA-owned or -funded resources more effectively resource sharing 2 The desire to obtain low cost computer communication facilities necessary for resource sharing This requires both high bandwidth e g file transfer and low delay e g interactive traffic to be serviced 3 An interest in applying experimental packet switching techniques to communication development to overcome limitations of conventional data communications high error rates low bandwidths inflexible topologies and limited reliability -7- 4 The need to develop alternatives for military communication systems having lower cost lower delay and higher bandwifth capabilities than those currently in use while still providing the end-to-end security and reliability needed STATUS The ARPANET project has been successful in several ways 1 The technical feasibility of packet switching for terminal-to-computer and coinpufrer-to-computer communication has been demonstrated at marginal costs lower than any present alternative 2 Common protocols which allow diverse host computers to communicate with one another have been designed and implemented at virtually all sites in the network The network has provided a good test bed for exploring solutions to problems of interprocess communication distributed operating system design interlacing diverse operating systems security and privacy accounting and reliability 3 Effective sharing of the network's resources among users and host computers has been achieved This sharing has permitted closer interaction among researchers in the network community made better use of limited computer resources and has demonstrated new capabilities in computer science and project management 4 Research into new communication methods based on packet switching e g packet radio packet broadcast satellite is now under way largely spurrpu on by the initial success of the ARPANET packet switch j experiment SERVICES VS EXPERIMENTATION Conflict Between Service and Network Experimentation Needs The success of resource sharing and the building of intersite protocols has required that the packet switching network offer a stabilized service with good reliability and low error rates Experiments with the packet switching subnet have been limited to some extent by the constraint that the network must remain operational Some small scale experimenting can and has been done at Bolt 3eranek and Ntwman with IMPS or are under construction and testing TTPS which In-house networks of three or four nodes can be created from equipment scheduled for shipment -8- However full-scale experiments on the ARPANET have been curtailed owing to th demand for a functional network As an example for the problems that occur when the functions of service provision and network improvement compete consider the following case In late 1971 it became apparent that serious problems were being encountered with the flow control mechanism in the IMP system system BB N set about to design a new flow control By mid-1972 the revised system was ready for installation after undergoing txtensive testing in the laboratory on a small scale network 3-4 nodes In the ensuing several months attempts to install this system met with unexpected disasters The BB N staff were limited to one try per week Tuesday mornings and it took about 2 1 2 months before the new version was stable enough to be used operationally The usual flaky period followed with minor bugs discovered as the system was exercised e g collecting statistics caused the net to crash Separating Resource Sharing and Switching Experiments In the belief that many experiments are yet to be tried on the net it becomes timely to plan to separate these network experiments from the resource sharing experiments Provision for experiments with the packet switching network e g satellite usage packet radio very high multi-megabit bandwidths interconnection o packet switching networks can be met by forming an experimental subnetwork distinct from the service network Of course a host may be on both nets but the nets should be independent of one another CHARTER RESTRICTIONS There are other important reasons for making a clear distinction between experimentation and service By its charter ARPA is not in the service business it is a research agency Of course it can and must purchase services to carry out its research program so that ARPA will always need services to support its research By separating and distinguishing its -9- L anticipated needs for various experiments ARFA will help pave the way for transfer in some form of the part of the network which it can no longer justify managing without disruption to its netting research ARPA can exercise long term leverage on the evolution of commercial packet switching through a carefully planned transfer which acts beneficially on the development of commercial packet switching services DEMAND FOR ARPANET SERVICES Nature of the Demand The payoff of the ARPANET'S unique capabilities for resource sharing has been sufficiently visible to interest many non-ARPA supported groups in connecting their computing facilities to the network This interest for the most part cannot always be met because of the present restrictions on access to the ARPANET This demand for access is multi-dimensional Sometimes it is sought by computer center directors seeking to sell unused computer time Sometimes it is sought as a low cost answer to the requirement for stable computer communication service spanning the continental U S and Hawaii Experienced Demand M iny groups in the private public and military sectors have requested access to the ARPANET Some of these requests have been accommodated through issuing ARPA contracts Others in the government sector have access by direct transfer of funds co ARPA and many have simply been turned down or have not met the DoD guidelines under which access could be granted Expected Demand The demand for interconnection is likely to continue and more likely to increase The benefits of the unique national computer communication capability offered by the ARPANET among the connected sites include 1 Better shared computer interaction among researchers -10- 2 Rapid sharing of results via file transfers and convenient message exchange 3 Better sharing of software computation and data resources The T yof s for these features appear to be sufficiently well understood so that the pressure for access will not go away of its own accord by those not now connected with ARPA's research program The political pressure on ARPA for access will increase while extending ARPA support for all these interested groups is impractical Thus the question that must be addressed is how to respond to a real nend without disrupting ARPA's on-going interests Therefore we shall seek ways to allow both ARPA and non-ARPA groups to share packet switching communications resources REDUCTION OF MANAGEMENT BURDEN The Problem Since 1968 ARPA IPT has borne almost all of the cost of maintaining improving and operating the network The most time- consuming aspect of network management to the small IPT staff has been the allocation of computing resources on the network to the research projects sponsored by ARPA Provision of computing resources to one network site from another requires a conscious policy decision and paperwork authorizing expenditure of funds as each request is unique The issvie is not a matter of a tangible dollar saving since transfer oi any part of the network is not expected to reduce the immediate out-of-pocket costs to ARPA for services Rather the administrative issues represent a drain on ARPA's management resources which might better be spent on research management Interim Administration ARPA IPT has already reduced - he management impact on its administrative staff to some extent by funding the following organizations for administrative tasks a BBN NCC Network Control Center -11- b RML operation and management of the network c MITRE facilitation of new attachees d SRI NIC network Information Center Nevertheless each of the administrative groups must still be coordinated by ARPA IPT In comparing alteinative transfer plans all things being equal we would tend to favor those approaches which reduce ARPA IPT administrative responsibility for the service functions of the net to the greatest extent commensurate with ARPA long range goals This in turn causes us to give our attention to commer- cial availability for the service functions desired and their control by the marketplace For example as an extreme we might imagine turning to a free market in which the research sponsor provided rav dollar - for each project's computational requirements with the proviso that each site spend its money as it sees fit for computing resources and computer communications Such a strategy delegates the funding policy decisions to each site freeing ARPA IPT from this task However effective resource sharing would still require a close awareness of available resources at each site by every other site and there may be practical factors that will limit the effectiveness of this approach The next section discusses some of these constraints in detail COMMERCIAL INTEREST IN PACKET SWITCHING There are several companies interested in entering the general packet switching business as purveyors of services to all comers as well as being interested in supplying networks or parts of a common network for specific applications Could ARPA buy the services it needs from such companies in lieu of the ARPANET Of course this is a real possibility but as a minimum ARPA should be prepared to spend more for such services than it is presently paying partly because some of the costs of the ARPANET are buried in other budgets partly because of -12- the extra marketing and overhead costs involved in operating as well as possibly more expensive line costs A key question regardless of cost is whether the ARPA network itself offers a commercially viable nucleus around which a packet switching industry can develop The present topology of the ARPA network does not span the center of commercial computing usage in the United States even represented New York Chicago and Texas are not Many more nodes would have to be added in Los Angeles San Francisco Boston Chicago Dallas Houston retroit St Louis Seattle New York etc before adequate access could be had to network resources by commercial computing centers IMP and TIP equipment presently in use by the ARPANET employs Honeywell DDP-516 and 316 machines These computers are 10-year old technology with severe memory size limitations obsolete architecture and expensive components Newer minicomputers which utilize solid state meuories LSI logic and microprogramming offer lower cost and increased flexibility Thus the ARPA network is seen as a small nucleus focused oa support of a research community rather than service to large commercial markets ARPA network technology is aging fast and its topology is not ideally suited to the support of a nationwide commercial service Given the new technologies emerging HSMIMP multiprocessor SUE satellite IMPS packet radio the present ARPANET in toto as a closed system is not an ideal business venture Initial ownership is desirable mor 2 for providing momentum to a new company than in its tangible value If however the price of the net is sufficiently low and the price for providing services to ARPA customers is sufficiently high it would be of interest -13- 3 TOWARDS A RECOMMENDATION GENERAL POLICY STATEMENTS Given the background situation as described the following postulates form the basis upon which we shall propose a specific recojnmer ded course of action 1 It is in the nation's interest in best using resources tc encourage computer resource sharing 2 The development of the packet switching industry will aid resource sharing 3 Better use of national communication and computation resources could occur if all packet networks were built so as to readily interconnect with one another The nation could develop a unified service accessible to all users on an equitable basis rather than isolated networks 4 ARPA should not dispose of the ARPANET merely to underwrite the funding of a commercial service Not only would it be inappropriate to use ARPA funds as venture capital for the support of any single packet switching service entrepreneur but also a sale of the entire network to a sole bidder could conceivably imptde that bidder's ability to adjust to or introduce new technology 5 A conservative policy could be to adopt a wait and see attitude observe which comrercial offerings survive trial by fire and purchase service from those which appear to be technically and economically sonnd 6 A more active policy has much to recommend itself In such a plan ARPA would stimulate the commercial development of packet switching technology for example by release of all technical details of present system design paid for by ARPA funding and actively extend the present understanding of network design and its performance 7 The government can aJso influence this vivified packet switching development by the magnitude of its demand for computing power It could for example require a commitment to meet all present and future network interconnection standards from firms supplying packet switching services to -14- the government Separate commercial networks may be nervous about interconnecting the primary barrier being reluctance of Company A to guarantee quality of service to those customers dependent on the performance of both A's net aid that of autonomous Company B Unless there is a mechanism for the enforcement of performance standards and transfer payments the goal of easy interconnection may remain elusive A PROPOSED TRANSITIONAL TRANSFER STRATEGY The Concept of the Packet Switching Industry Cooperation Arrangement As a matter of public policy we would prefer to see the packet communications industry encouraged to develop in a manner non-conducive to monopoly As such it would do well to have the charac- teristic of low cost of entry free competition and enforceable interfacing standards to aid the harmonious interconnection of private public and military networks Appendices B and C which deal with organizational and institutional matters consider the establishment of a non-profit mechanism for cooperation perhaps a government-industry activity or consortium which could administer the interconnection of participating networks provide for a clearinghouse operation for use made by one network of another and insure standards of performance At pr oent the government is both a supplier and user of packet communications Therefore entry should be open to all private public or military agencies having a packet network Networks as small as one IMP and one TIP could be eligible This industry association or cooperative or consortium would function as a settlements clearinghout 3 and ns a coordinator We believe that it would be appropriate that ARPA be one of the founding members along with any of the fledgling packet switching firms Interconnection standards hardware and software would be developed and agreed upon by the consortium membership and internally administered However recommendations by CCITT or ISO might also be adopted and enforced by the consortium as well as NBS standards -15- Upon establishment of the consortium ARPA at its conven- ience could divest those parts of its network which are not directly related to ARPA research goals e g AEC HEW und other DoD sites by having them join the consortium and by transferring ownership of their IMPS or TIPS in exchange for funds and or services Under the rules of tha consortium distinct networks could bilaterally or multi laterally arrange interconnections ARPA may choose to allow some members to connect directly to its network via packet switches IMP-IMP and others via gateways In those cases where ARPA allows direct connection it would dictate points cf interconnect thereby controlling its own topology All other interconnections could be via gateways which are attached to IMPS designated by ARPA If there are any initial problems encountered with the joint use of AT T telephone circuits by members of the consortium the procedure maring all subscribers co-leasors of the lines such as used by Tymnet could be used The consortium network will need facilities similar to those on the ARPANET a NIC NCC These could be supplied from the commercial sector as in the case of Tymshare's NLS for ARPA and funded out of consortium fees In the initial periods of operation it is likely that consortium members will all be owners of IMPS and TIPS Eventually other whole packet switching networks could join NEW OPTIONS Once ARPA has accomplished its two short-term objectives of 1 separating the SERVICENET from the RESEARCHNET and 2 using the SERVICENET or some portion of it as the vehicle for catalyzing the consortium it can decide what to do with the remaining network 1 ARPA could sell the network and lease it back from a private firm as control and interconnect management problems are solved V Cerf and R Kahn A Prc-ocol for Packet Network Inter- communication to appear in IEEE Transactions on Communications May 1974 -16- 2 ARPA could sell the network equipment and purchase services from one or more consortium companies for its contractors 3 ARPA could write off the network use it for experimentation etc and simply allocate funds to its contractors who can then choose from companies U V W W Y the computing services they desire If U and Y are members of the consortium then resource sharing can continue among the ARPA research contractors ESTABLISHING THE CONSORTIUM A number of options have been considered In the main it is most in keeping with ARPA's historic role that it be a catalytic gent and fade out of the picture of the consortium as it would no longer be supplying services even to itself One possible strategy could have RML sponsor a small industry group to create the shell of the consortium as a non-profit industry organization In this case RML would over the short term continue to administer the network -- but bypassing non-ARPA expansion -- to other members of the con Ttium Transfer of nor -ARPA portions of the network could occur in an orderly fashion Service centers now on the network could purchase their IMPS and join the consortium expanding their markets in the process if they wished Over the long term ARPA couM consider the various transfer options for the service portion of the current ARPANET while retaining as separate and distinct an experimental research network which is not part of the consortium Host computers would be allowed to reside on more than one network RML could continue to serve ARPr in the administration of the experimental network but would relinquish direct responsibil ty for the SERVICENET upon its transfer Therefore in summary of the proposed Transfer Plan 1 ARPA would divest itself of th service aspect of the ARPANET 2 ARPA would retair or create an experimental subnet on which it can test satellite eommunicatio packet radio notwork intercom action and other ideas and -17- 3 the transfer would provide for sharing of the packet switching subnet among the private government and public sectors under the aegis of a consortium in a legal and harmonius fashion with minimal need for FCC control o -18- SPECIFIC PROPOSED PLAN FOR A PARTIAL TEST TRANSFER REQUEST FOR INDUSTRY COMMENTS To this point we have discussed a very general Transfer Plan and shall now consider the more specific steps required We believe that it is timely to consider reviewing the details of proposed test transfer with industry on a completely open basis Unless the eventual arrangement is acceptable to one or more responsible organizations competent in packet switching then the divestiture plan cannot serve its intended purpose Any proposed arrangement must be fair to all parties be workable and must lead to the desired end objective We recommend that the items and comments below be presented to industry for comment and feedback as an aid to planning These are arranged in question and answer form to aid in collation of comments PURPOSE OF DISPOSITION Why is the ARPANET important to DoD The ARPANET is a nucleating seed of a major potential national resource whose continuing operation is deemed to be in the public interest What is the purpose of the proposed disposition The proposed disposition of the facilities is to accelerate the commercialization of technology developed by the Department of Defense and to permit the provision of such services to the Department at comparable cost wider availability and greater -19- effectiveness tha i the alternative arrangement of having to supply the same service under closely managed Department of Defense control What is DoD's immediate interest It is the interest of the Department of Defense to have such facilities continue in operation and continue to be made available for use by ARPA and other parts of the Department ttiat could thg transfer accomplish The Advanced Research Projects Agency ARPA of the Department of Defense us operator and custodian of the ARPANET seeks to transfer a portion of this network to better meet long term demands for growth of the network improved versatility survivability reliability and usefulness to resource sharing in detail these objectives are To meet the needs for system growth The ARPANET has grown to about a forty-seven node net in several years and may grew at a similar rate for the near term It is inappropriate for the DoD to sponsor the growth of any communications network beyond its own needs especially if the private sector can accomplish the same end Therefore to respond to the pressures for growth on the network in both the number of connected sites tnd volume of traffic the private sector is invited to share in the growth in lieu of openended governmental sponsorship To improve system versatility As a DoP entity the ARPANET is highly limited in the ease with which it can connect other governmental users and is precluded from adding purely commercial users At present the ARPANET facilities are limited to servina those with an ARPA contractual relationship and universities Private organizations performing research in behalf of ARPfi and government agencies may be served but only those with a research requirement appropriate to ARPA's interests Thus the present rapidly growing community of interest represents only the tip of the iceberg in -20- the new demand for services Such a demand cannot be filled without private participation Private packet switching networks are coming but in their early state can probably serve only groups of dedicated users Gain economy of scale There are economies of scale in several aspects of communications networks Better economy results if many users share a conunsn resource than each providing his own under-utilized network While there is the prospect that many packet networks will be built in the next decade it will be in the public interest that these networks be able to interconnect to one another ii a reasonably effective manner and that artificial barriers not be erected at the interface between these networks to prevent such flows whenever it is economically desirable to do so Improve system reliability A small thin network cannot be as reliable or handle as heavy peaks as a larger one with more redundant paths Access to such larger facilities will be bene- ficial to ARPA ARPA wishes to develop a rational set of rules to define and determine reliability for an oyerall network or for subnets Since the government is relatively protected from the disciplinary forces of the marketplace such rules for the protection of the overall network become mandatory Improve overall survivability for critical users The larger and more highly interconnected a network the more survivable it can be to enemy attack or natural disruptions Thus the ability of networks to interconnect with one another aids survivability Of cjurse the problem is complicated as each would operate under independent and autonomous managements where the present ARPANET or the divested form of the present ARPANET is but one part of an eventual composite Combined Network What is sought is a set of independent networks that can operate together Accelerate the development of resource sharing The initial motivation for the ARPANET was in part to aid the development of large scale computer resource sharing In keeping with this goal ARPA wishes to encourage potential shared usage of facilities t -21- between the ARPANET and any future firm wishing to acquire a portion of the ARPANET REQUEST FOR COMMENTS - OPERATIONS Having described ARPA's goals and objectives ARPA welcomes comments and suggestion that would help achieve these goals and is not limited by preconception or prejudgement ARPA welcomes new private sector initiatives here and wishes to encourage all such efforts What is meant by Combined Network Definition of the Combined Network The Combined Network consists of parts of the ARPANET devoted to providing non-experimental services plus one or more independent networks which are interconnected by one or more gateways Each member of the Combined Network is a Combined subnet Will more than one Combined Subnet be Permitted es ARPA is interested in considering responses from interested parties in helping to achieve the stated goals evolving toward development of a Combined Network comprised of the remaining ARPANET plus new participants While the ARPANET is composed of a homogeneous set of assets e g IMPS TIPS leased line arrangements only a minimum of coordinated management services connecting the components together must be centralized Aside from the provision of some minimum overall management services for network control the ARPANET could evolve into multiple ownership provided operation of the overall network is not jeopardized How can packets flow from one subnet to another The art of connecting different packet switching networks is still in a primitive state of development -22- Among the most care- fully thought out proposals is that of V Cerf and R Kahn While the Cerf Kahn protocol appears to be a workable method of interconnection even with entirely dissimilar networks and goes far towards the solution of building the Combined Network comments from industry are particularly desired While it would be highly desirable that the interconnection continue to be made at the packet level provided this can be done without jeopardizing overall network-performance interconnection with gateways is clearly feasible at this time While no uniform packet level transnetwork protocol is presently defined this arrangement will be regarded as an open possibility if a suitable proposal is made showing methods of insuring it workability What does the ARPANET consist of The total facilities of the ARPANET including those under consideration for possible divestiture consist of the elements shown in Appendix H Equipment on this list are in as is state and no statement implied or otherwise can be given as to condition or operating performance These elements are connected together with communication lines provided by common carriers These lines are not the property of the ARPANET and are leased as shown How much did the facilities cost Equipment elements that might possibly be diver ted have been purchased over a several-year period Appendices G I and J describe a program used to estimate the present value based upon various costing a d depreciation schedules Whatever amounts are shown for V Cerf and R Kahn op cit IEEE Transactions on Communications to appear May 1974 23- the con'onience of the reader shot Id not be misconstrued as being an implicit statement of the value of these facilities What is the 516 IMP 316 IMP and TIP The 516 IMP The work IMP refers to the unit described in insert specification in Formal IIFP and includes a Honeywell DDP-516 minicomputer Its characteristics are sixteen-bit word size 96 microsecond memory cycle 16 000 words of memory sixteen multiplex channels and sixteen priority interrupts The DDP-516 provides a throughput rate of about 850-kilobits per second as used in the ARPA network The 316 IMP The 316 IMP is similar in operation to the Honeywell 516 described above except that it is a lower cost unit also made by Honeywell and produces a throughput of 650-kilobits per second The TIP The TIP is constructed of a Honeywell DDP-316 com- puter plus an additional 12 000 words of memory and a specialpurpose multi-line controller built by Bolt Beranek and Newman If the ARPANET is divested what minimum standards will be required The ARPANET has established through its various management services a procedure for testing the network and keeping daily records on the network's operations Generally the overall reliability of the system has been on the order of 98% up-time The target minimum reliability is that the overall network be operational twenty-four hours per day every day of the year What allowance should be considered for downtime In computing reliability it can be assumed that any individual IMP or TIP may be taken out of service as required by routine or emergency maintenance provided that the fractional amount of time involved in all such maintenance shall not exceed five percent 5% of all time computed as follows -24- Routine maintenance and program changes which interrupt service shall generally be restricted to a Scheduled Maintenance Period The Scheduled Maintenance Period shall extend between 0100 and 0600 Eastern Standard Time Any failure which interrupts on-going computation from any TT or IMP occurring outside a published Scheduled Maintenance Period as defined above shall be counted as Emergency Down Time Emergency Down Time shall be measured from the first detection of failure until the failing unit is restored to full service The duration of Emergency Down Time shall be multiplied by ten 10 when computing the time involved in maintenance discussed above Any transient failure which interrupts on-going computation frtyn any TIP or IMP for three 3 minutes or less shall be counted as a thirty 30 minute Down Time failure Any transient failure logger than three 3 minutes shall be considered as Emergency Dovn Time Failure of a single TIP input modem from the user shall count as a failure of one-tenth 1 10 its time duration in computing TIf statistics Hc w would performance be reported The owner of any transferred facilities would bo expecv ed to provide failure statistics for each TIP and each IMP rwonthly and certify such reports as being correct Signed copies of the performance reliability report would be sent to all active nodes on the combined network and be used as the basis for reaching a subsequent decision for converting from a tentative to a final transfer of divssted facility ownership What are the rights and duties for interconnection It would be expected that privilege would be granted to allow any IMP in the transferred part of the ARPANET where desired to be able to interconnect with any other IMP in the non-transferred part -25- Such a connection right would be subject to the requirement that any such interconnection not jeopardize the reliability or performance of either node or reduce the combined network reliability and the incremental costs would be borne by the owner wishing interconnection Requests for interconnection meeting this require- ment would be granted and effected within 60 days not counting delays involved in installation of required common carrier facilities Who would pay for the cost of leased lines used for interconnection Leased line or other communication arrangements would be negotiated by the nodes seeking interconnection The share of costs borne by each node would be subject to negotiation Would all parties be treated the same Yes All conditions and terms stated above would apply in an equal and reciprocal fashion to all participants in the combined network How would a prospective connecting subnetwork go about seeking ownership of a part of the network Thio of course would be controlled by allowable government contracting procedures but in general requests for proposals would be issued to encourage specific proposals where each such proposal would indicate the particular elements sought together with a proposed timetable Will there be a trial period Each connected subnet would be given a 12-month test period to allow ARPANET and other members of the combined network an opportunity to stabilize operating and management problems before final acceptance is granted During this 12-month transitional period ARPANET representatives and other members of the combined network would work together and review the technical managerial and financial problems that require resolution and negotiate policy matters regarding future operations -26- % How will cross payments be made One of the key aspects of the proposed cooperative sharing I I arrangement is that a satisfactory mechanism for cross payments be developed Details of one proposed arrangement are described in Appendix C on formation of a common interest consortium of packet switching entities o One of the activities during the first year trial period is to test and refine the payments exchange mechanism The participants of the combined network will on occasion use ARPANET facilities for switching packets in cases where the chosen route requires relay through ncno transferred nodes In such cases the ARPANET should be re- imbursed a per-packet fixed charge Similarly the transferred ARPANET nodes could use the transferred facilities to relay ARPA-originated terminated traffic o In such cases a reverse payment should be made upon an equivalent basis How much traffic will ARPANET continue to flow through its possibly transferred IMPS This is of course difficult to assess but it seems evident that the traffic should be no less than is presently being carried It is conceivable that ARPA could enter into contracts guaranteeing a minimum level of ARPA traffic that will equal or I exceed either the average of the six months preceding the transfer or the average of the busiest six months in the year precid ng '-ransfer whichever is greater I RANGE OF COMMENTS DESIRED To this point we have described how the transfer might be carried out But this transfer cannot be done in a vacuum It will require close cooperation between ARPA and other members of ' the possible combined network To this end comments are enter- taired trom all interested parties What i'j sought is broad-based consensus of the most effective way of achieving the end objective namely the long-term gradual o conversion of a DoD-owned experimental facility to an operationally -27- integrated but independently-owned network of networks providing excellent quality guaranteed service to fill ARPA's needs But whatever alternative is suggested it should also contain provisions to protect the government including return of facilities to ARPA with financial penalty if benchmark performance standards are not met And agreement by new owners of the transferred portions to interconnect to all other packet switching networks that meet agreed-upon standards provided such other packet switching networks wish to engage in such interconnection In reviewing the comments three test criteria should be 1 Does the proposed arrangement guarantee that the operational performance of the ARPANET will be equal to or better than is presently experienced 2 Will ARPA retain the freedom to regain the transferred portions if the trial does not work well 3 Does the proposed approach avoid the present trajectory by which the nation will probably find itself with a set of packet switching networks that cannot talk to one another preventing optimum reliability load and resource sharing capability inherent in large packet networks IMPORTANCE OF INDUSTRY-GROUP ARRANGEMENTS Basically the detailed plan proposed here is a way tc guarantee a supply of milk without owning the cow The idea is here is our cow which we could rent or sell to a buyer at a low price but we insist that the buyer guarantee to take good care of this cow because we shall want it back if it isn't producing good quality milk As a matter of simple prudence it is reasonable to expect that the would-be buyer demonstrate a knowledge of one end of the cow from the other ARPA's interests would bt further served if the buyer were willing to join his local dairy farm cooperative so that it can take on much of the burden of quality control This not only reduces concern about having to worry whether the milk is fit to drink it will also permit access to a greater supply of milk through the cooperative if needed -28- And conversely it solves the problem of what to do with an excess of a highly perishable ucnmodity which is occasionally in surplus Of course there is no farm cooperative for packet switching But as described in Appendix C on the formation of a common interest consortium of packet switching entities there are good reasons for ARPA to aid in establishing one The possible transt r should not hinge on the existence of a cooperative Rather the transfer is made much easier if the members of the combined network vrere working in a cooperative manner as will be described in Appendix C 29- ATl-ESDIX ON LEGAL BACKGROUND ANALYSIS Appendix A The following appendix was prepared by Paul Goldstein to describe the legal considerations of divesting the ARPANET regulatory considerations and the alternatives to regulation Appendixn mm BACKGRUJND ANALYSIS bY PAUL GOLDSTEIN Appendix A LEGAL BACKGROUND ANALYSIS PREFACE Dispositions of public wealth have historically raised issues more pregnant and complex than those attending dispositions of private wealth Policy in the form of established legal rules naturally shapes both dispositions private and public In the public disposition however policy occupies an added special place it is made The sale of public lands is an early example urbar renewal a more modern one of the disposition of public wealth to achieve specific goals--civilization of the frontiers or more recently of the cities Complexity stems in part from the fact that the implications of any public disposition will invariably exceed its avowed objectives Thus large scale trans- fers may by inundating supply influence market prices both short and long term second order consequences usually include extended distributional effects t Issues of yoveinment disposition become even more complex when the public wealth to be conveyed takes the form not just of realty or personalty but of a functioning public institution that possesses many if not all the attributes of the firm The com- plexity attending divestiture of public firms-stems from the nature of the public firm--a firm all of whose operational decisions are vested in or made the responsibility of a governmental body--and of divestiture's consequences for th3 regulatory process consequences flowing from the transfer to the private sector of the power to make some or all of the firm's operational decisions decisions previously made in the public sector Stated in its broadest terms--terms that will be refined in thi course of this article--divestiture of the public firm involving as it does the transfer of decisions from public to private hands represents the A-l converse of the regulatory process which involves the transfer of private decisions to public hands Because it is so thoroughly imbued with regulatory implications the decision to divest a public firm deserves at least the level of attention paid the decision to regulate The central most difficult task is determining which functions should remain under public control--regulated--which should be divested--deregulated--and the extent of divestiture for any component The determination is complicated by the fact that public ownership may entail concessions not immediately available to private owners-- reduced government rates on telephone lines for example or the governmental capability of continuous below marginal cost pricingconcessions likely to generate false signals respecting the prospects for the firm's success in competitive markets These and related considerations underlie the federal government's deepening evaluation of plans to divest the ARPANET an experimental venture of the Advanced Research Projects Agency of the Department of Defense ARPA designed to test the efficiency reliability and economy of a packet switched network for computer communications Expansion of ARPANET'S present structure and technology is expected to accelerate developments in the already burgeoning computer communications industry indeed the network may eventually form the nucleating seed of a major international and domestic data communications system The success of the ARPANET experiment as measured by the satisfaction of present users and the increasing demands of prospective users for admission to the system has raised the question of the institutional form that a fully operational cost efficient network should take Part I of this article considers in the context of an analysis of the public firm generally whether the decidedly commercial cost of the network's future role excludes the public firm as a fitting candidate for the network's continued management Part II describes ARPANET'S nature and origins and identifies an ideal set of characteristics for future operations If as concluded full government control of network operations A-2 through management of the system as a public firm would confine network performance to a point far short of the ideal stated in Part II the relevant question becomes to what extents should exercises of government control and the discipline of market forces influence the network's management for the ideal to be approached This question is considered in succeeding sections Part III examines regulation through government retention of certain network components Part IV direct regulation on the common carrier or public utility model Part V summarizes the probable effects on network operations of largely unfettered markets and explores two market alternatives to regulatory techniques A-3 I TOE PUBLIC FIRM TOWARD A REGULATORY CONCEPT OF TOE PUBLIC FIRM The behavior of all firms private or public regulated or unregulated can be descr ed in terms of the operative decisions the firm routinely me kes With respect to the goods or services supplied by the firm these decisions embrace price quality marketing techniques materials and labor The decisions also involve judgements respecting the level of investment to be committed to plant and research and development and the rate of return to be derived from investment While these decisions are common to all firms the conditions under which they are made will vary with the character of the firm and with its regulatory setting In the private firm decisions are largely left to managers and boards of directors to be made according to the objectives for the firm set by them or the firm's stockholders Even for the private firm however there are some regulatory constraints on decision Antitrust strictures for example may affect firm decisions respecting growth and caution against setting prices differentially or below average cost no matter how profitable either strategy may appear A pharmaceutical company's decisions on the quality of its drugs may be importantly confined by Food and Drug Administration rules and its marketing decisions limited by Federal Trade Commission rules on deceptive advertising in these last two cases the ambit for decision will be further circumscribed by the threat of private actions brought This catalogue of firm decisions is an abridgement and condensation of a more extensive though summary list set out in McKie Regulation and the Free Market The Problem of Boundaries 1 Bell J of Econ Man Sei 6 7 1970 A-4 by injured consumers and competitors In the regulated industries--occupied by public utilities and common carriers--some decisions are like the decisions made by private firms lodged with private managers bat hedged by general legal rules Other decisions are preempted altogether The com- mission charged with overseeing the regulated firm's operations may be empowered to determine the firm1 overall revenue needs and governmental determination of revenue needs will in turn affect other of the firm's decisions the firm's managers will be dis- suaded both from incurring expenses that they know the commission will disallow and from paring expenditures that no matter how inefficiently applied they expect the commission to tolerate Also while the regulated firm's revenues are prescribed in the aggregate tvc aggregate figure is not the only source of limitation on the firm's pricing decisions Thus -for example under the Communications Act's requirement that carrier rates be just 2 and reasonable the FCC enjoys the power to require alterations in tariffs that may in its judgement be too high or too low whether measured by the cost to the carrier of providing the tariffed service or by the value of the service to the user Finally the regulated firm's decisions respecting capital expenditures or alterations in service are limited by the requirements of commission approval accounting procedures too must be compatible with commission needs This comparison of private and regulated firms sheds some analytic light on the regulatory process generally and on the place of public firms in a regulated economy It should be clear from the case of the private firm and even more so from that of the regulated one that the process of regulation involves little more than tlie removal from firm to government of part of the power to make some decisions and in some instances of the power to decide altogether This suggests that differences in behavior between the private and the regulated firm are not of kind but 2 47 USC Sec 201 b 1970 A-5 degree the degree to which decisions have been transferred from the private to the public sector This further suggests that the difference between private and regulated firms on the one hard and the public firm on the other is also importantly one of degree in the case of the public firm all operational decisions are governmentally made Decisionmaking in publ'c firms as in private and regulated firms will of course be influenced by consumer preference Public firm decisions may additionally be affected by perceived voter preference There is to be sure a difference of kind implicit in the regulatory process a difference that is a function not of where decisions are made--in the private or public sector--but rather of the objectives toward which decisions are directed The determina- tion to regulate at any level implies a judgement that the performance or more accurately the effects of performance of the unregulated firm maximizing its internal economic objectives will not correspond with government's chosen social and political as well as economic objectives Thus the unregulated firm may con- sider that it serves its interests better by hoarding gold bullion than by purchasing pollution control equipment and that it would serve them better still by larding the campaign coffers of malleable legislators At the least stringent level of regulation laws establishing air quality standards curtailing traffic in gold and proscribing corporate gifts to political campaigns are intended to confine private decisionmaking to a range more consonant with 3 To the significant extent that public firms must compete with private firms in product labor and capital markets factors affecting public firm input decisions approximate those affecting private firm input decisions While government's power of eminent domain might appear to give the public firm an edge on inputs unavailable to the private firm the significance of the erlg is limited by the facts tha' government must pay fair market value for property condemned nd that private firms are increasingly coming to enjoy the substance if not always the form of eminent domain power A-6 perceived societal needs 4 More significant systemic departures of private firm behavior from governmental objectives may call for the imposition of public utility or common carrier status The reasons for regulation are even more apparent at regulation's extreme when the performance of private firms sufficiently departs from public needs to warrant the formation of public firms either from scratch as in the case of ARPANET or through the nationalization of existing private firms Wartime needs if satisfied neither by the operation of free markets nor by the incremental process of regulation represent at least the most dramatic predicate for the nationalization of private firms 4 It may be objected that the force of this distinction between public and private goals depends upon the view that firms pursuing private goals will be exclusively profit maximizing and that the distinction is blunted if as has been argued planning not profits constitutes the objective of at 'least the larger firms Compare J Galbraith The New Industrial State 1967 with W Mueller A Primer on Monopoly and Competition 160-175 1970 See generally G Stocking M Watkins Monopoly and Free Enterprise 491-529 1951 See also Hearings on Planning Regulation and Competition before the Subcosun on Retailing Distribution and Marketing Practices and the Subcomm on Monopoly of the Senate Select Conm on Small Business 90th Cong 1st Sess 1-45 1967 debate between W Adams J Galbraith W Mueller and D Turner However it is not at all clear that in adopting planning as its goal a firm is forsaking profits in any but the most limited sense What is more likely is that it is shirking immediate profits for profits in the longer term A firm or industry may for example voluntarily curtail its contaminant emissions and suffer diminished present profits in the hope that it will thus avoid public hostility and forestall future regulation that would cut more deeply into its operations The force of the distinction --and indeed the case for regulation--might appear weakened by this last observation to the extent that the firm plans with the objective of currying public and legislative favor its operations can bo expected to comport with public goals The problem is however circular absent regulation--or perhaps more important the threat of regulation--the firm would have no incentive to plan in these directions If anything complicates the distinction between private and puolic objectives it is that the pursuit of private goals through unfettered markets is in this country itself a cardinal public goal A-7 j Other functions--defense the administration of justice are examples--are viewed as so central to the political system that the need for exclusive government control has been treated as self-evident Regulatory expedience though a less obviously compelling reason than war or politics also accounts for the formation of public firms A commission that assumes control over a public utility's rate of return may soon find that the rate established has produced untoward effects on the utility's pricing and investment decisions requiring the firm to relinquish these decisions too the commission may discover that as a consequence wrong decisions are being made on still other fronts a phenomenon that McKie calls the tar-baby effect At some point full public control of the firm's decisions which is to say full public proprietor chip of the firm may appear the most efficient solution Efficiency more than any othec reason perhaps accounts for a situation in which it is not unusual to find that extensively regulated sectors of the economy succumb to complete socialization with governmer ownership and operation In the last century privately-owned roads canals and bridges passed from regulated activities to government operations The socializa- tion of water supply and urban public transit is nearly complete Any regulatory commission that tries to control these effects by regulating additional variables such as cost performance executive salaries - nd prerequisites choice of technical methods and rates of innovation will quickly find its hopes to economize the means of regulation evaporating As it extends further into the network of enterprise decisions it may discover that still other compensatory changes partly frustrate its efforts and there are always more just over the horizon Extension of control in response to perpetually escaping effects of earlier regulation may be called the 'tar-baby effect ' since it usually enmeshes the regulatory authority in a control effort c f increasing complexity with little gain in efficiency but a growing feeling of frustration McKie Regulation and the Free Market The Problem of Boundaries 1 Bell J of Econ Man Sei 6 8-9 1970 Jones An Example of a Regulatory Alternative to Antitrust New York Utilities in the Early Seventies 73 Colum L Rev 462 465 1973 A-8 FORMING THE PUBLIC FIRM If the market's failure to satisfy public needs is not the exclusive reason for the formation of public firms it is at least a predominant one The existence of three such public needs-- reliability of service equality of access and innovation in techniques--underlies ARPA's decision initially to structure ARPANET as a public enterprise and it may be helpful at this point to compare these needs generally with the nature of governmental response This is not to suggest that needs of this sort are best met by public firms but only that they are said to be 7 by those whose word is law That the public firm is not the only or necessarily the most efficient means for satisfying compelling public needs should be evident from reflection on the performance of some public endeavors designed to achieve reliability equality and innovation Reliability Together with related historical and political factors the need for a high degree of reliability is popularly perceived to underlie the de ision to operate the functions of national defense as a public firm While the need to internalize in v overnment the power to make decisions respecting the uses of the defense establishment should be self-evident it does not necessarily follow that the production and deployment of material and services are also best accomplished within the public sector indeed the military presently relies heavily on private firms for the production of material That the military service function has largely been kc_- t internal to the government can be ascribed to a factor not reproducible in the market at least not since passage of tha thirteenth amendment government's power to compel For one comparative study see Davies The Efficiency of Public versus Private Firms The Case of Australia's T'O Airlines 14 J L Econ 149 1971 A-9 its citizens to perform military or alte native service Although this power might appear particularly attuned to the level of reliability represented by a ciptive readily mobilized labor force its exercise indicates only that the military is paying its servants less than they could command in the marketplace Recent moves to abolish the draft and to replace it with schedules of compensation more nearly enjoying parity with labor's market value suggest that the conscription power is not a prerequisite to reliability and security Equality The problem of equality in access arises in its most graphic form when the cost of vital services for which demand is relatively inelastic--municipal transit and postal service are two--exceeds what an important segment of the public can reasonably be expected to pay To avoid undesired distributional effects government could permit provision of the needed services on a competitive basis and achieve equalization through direct payments to the poor either in cash or in vouchers as is done with food stamps Alter- natively government could channel its subsidy directly to the private entrepreneur requiring in return pricing that though uniform is at a level the poor could afford Government could also give the private enterprise a wide latitude for price discrimination prohibiting resale by low-price buyers and exacting as a condition for its permission the firm's agreement to price services for the poor at an affordable level below the firm's average and even marginal cost presumably the firm would make up its losses in these markets by capturing consumer surplus in more affluent markets The first two of these approaches are generally shunned because government seems to prefer covert to overt subsidies While the third approach roughly approximates the one employed in the differential pricing of business and residential telephone service there is no evidence that residential service is provided at less than marginal cost and in any event the latitude allowed A-10 has not been sufficiently broad to provoke undesired political g reaction The governmental solution motivated by the need for both equality as actually enjoyed and equality as perceived has in many areas been the public firm setting a single below cost price for its services subsidizing its activity covertly from tax revenues Innovation The formation of public firms may also be prompted by the existence of areas of significant technological need the resolution of which will not for one reason or another be achieved in the private sector The very magnitude of the problem to be solved may be thought a sufficient condition to goad the profitseeking firm to its solution particularly if the firm is abetted by the promise of patent protection for its discovery Yet the anticipated profits to be derived from marketing a discovery do not necessarily correspond with the magnitude of ics neeA a phenonmenon that may go far to explain the poverty of innovation in instructional materials for public and private schooling At the same time elements of risk associated with any research and development venture may for the private firm render the opportunity costs of research expenditures unbearably high Finally the patent law never a particularly efficient system for encouraging needed innovation has in recent years revealed itself to be an increasingly creaky device its promise hedged on all cornors It is in these areas of great unmet needs where the calculus of anticipated profits risk and patent protection weighs against the private commitment of resources to innovation that more direct government intervention becomes appropriate Intervention may take the form of direct subsidy as in ARPA's o At the same time significant income-based disparities in access have been successfully avoided in the pricing of residential telephone services See Bureau of the Census Current Population Characteristics Characteristics of Households with Telephones Table I Series P-20 No 46 1965 A-ll dramatically successful program of support for research in advanced computer capabili zies or it may take the form of a public innovative enterprise like ARPANET DISSOLVING THE PUBLIC FIRM Because it is an instance not an exception of government regulation the public firm is subject to many of the same stresses that affect regulation generally Thus just as for regulated private firms the tug of the marketplace may first be felt in assertions that the firms will more efficiently achieve relevant public goals without a particular legal rule than with it so in the case of public firms the demonstrated superiority of private firms in reaching public goals may call for the transfer of all or parts of the firm from the public to the private sector The mission-oriented enterprise like ARPANET once having marshalled resources to initiate major change may if called on to market the services it has developed be expected quickly to fall into a pattern of resisting change a particularly undesirable posture in fields where continued flexibility and invention is essential ARPANET'S nature objectives and underlying technology described in the next section strongly suggest that an optimally functioning network will have to be highly responsive to the needs of users in private firms academic institutions and government agencies and that to a significant extent responsiveness will require the commitment of resources to incremental innovation This informing need responsiveness to consumer demand particularly as responsiveness takes the form of innovation directed toward demand suggests the public firm's incapacity to provide the desired kind and level of services Lacking a price mechanism sensitive to competitive forces lacking any basis foxreceiving accurate signals as to performance and consumer needs lacking any spur to business-oriented innovation or any road map identifying the proper direction for innovation to take the public firm seems poorly placed to operate an optimally functioning network This is not to say that the public firm's structure A-12 for decision-making could not with some work be fashioned to simulate the more responsive decisional structure possessed by private competitive firms--indeed it could The point Lner is that if a structure disciplined by the market would function best then it would seem more efficient to bypass simulation and get ehe real thing In assessing the most efficient means for government to shed the public firm it will be important to keep in mind that although the transfer of public firms to private and competitive markets is commonly chzr cterized as divestiture it may be both more accurate and helpful to treat the transaction not as a sale but as an instance of deregulation Just as the regulatory loop can be run torward from the largely unregulated private firm to the regulated firm to the public firm so it can be run in reverse shedding some of its components retaining others the public firm can be introduced into the market as a private firm regulated to varying degrees This point suggests that the decision to divest involves judgements respecting not only the extent to which previously regulated components should be deregulated but also by implication the extent to which the firm's behavior should remain regulated taking ir co account the effect on regulated components of the newly deregulated components t A-13 II THE NATURE AND OBJECTIVES OF THE FIRM Under present generally prevailing coiiditions a computer center if it wishes to use distantly located data files or software must first reproduce the data or programs internally As elsewhere this redundancy stems from inefficiencies in communication specifically from the application of existing communications systems to unanticipated and largely incompatible computer communications uses quality cf long-haul service over telegraph and voice grade communication lines is far below what computer users can reasonably be expected to tolerate and the cost of national interconnection through leased lines or dial-up facilities is prohibitively high Widespread differences iu local facilities also impede fluent computer communications hardware programs and formats at one site may be incompatible--and hence uncommunicative--with their counterparts at a distant site from which information is desired The central objective of packet communication networks is to reverse these inefficiencies and capture the significant unrealized scale economies represented by multiple widely distributed use of a single computer resource hardware software and data which now must be replicated to be used at distant sites would be directly accessible to any system in the network wherever located The informing innovation of network packet technology lies in its conversion of existing communication modes to efficient communication use Added innovation has focused on resolving the second need inter-system compatibility Linking a number of autonomous nationally dispersed computer centers packet communications networks would facilitate interactive communications between any two systeius within the network A-14 Network operation would rely upon a geographically distributed set of switching centers Each center would be hooked up through leased landlines or through microwave or satellite to a small number--subset--of other centers centers within any subset would be linked to centers in other subsets with the result that taken together the centers would form a fully distributed richly interwoven network enabling communication through linked centers between any two points in the system Because several linked paths would be available between any two centers a high degree g of reliability would be assured Situated at each switching center would be technology functionally similar to that now employed at centers in the ARPANET an Interface Message Processor IMP --a small general purpose computer designed to route messages in the form of packets of c bits check for errors and provide links to the network's computer resources HOST'S TIPS--terminal IMPs performing all IMP functions and also interfacing up to 64 individual terminals within the network--are also scattered throughout the ARPANET As a goal future networks would seek to offer a wide variety of services tailored to meet the information processing needs of the broadest range of users Access between any two points in the world would be possible at high data rates and with communications and maintenance costs errors and dealy low in comparison to other system costs The system would be fully distributed and autonomous so that malfunction or disaster at any node need not affect the rest of the system users could enjoy sufficient security to exchange messages with only minimal concern for sabotage or other interference by unauthorized oursiders These are to be sure goals and not 9 This description of ARPANET and its background is drawn from Roberts Wessler Computer Network Development to Achieve Resource Sharing 36 APIPS Conference Proceedings 1970 Spring Joint Computer Conference 543 1970 which contains a good general discussion of the network Other succeeding papers in the same volume present more specific studies of aspects of network operation A-15 o A present reality but they do suggest the contours of an ideal toward which the system's operation might be directed It was in response to long-range goals of this sort as well as to short-range interests in sharing ARPA-owned or funded resources more effectively that ARPA initiated the ARPANET experiment in September 1969 In the experiment's first phase the network interconnected 14 sites primarily university and non-profit research centers each involved in ARPA-supported research widely scattered across the country The second phase which began in 1970 involved the interconnection of additional sites engaged in a broader range of research activity However tempting entry into a commercial service operation might at any point have appeared ARPA chartered as a research not a service agency avoided these potentially lucrative markets Largely because ARPA's mission in designing the ARPANET was not to develop an ideally or even adequately functioning 10 Recognition that this is an optimum not likely to be realized in practice does not necessarily imply that all steps taken in the direction of attaining the optimum will place the network in a better position than would have obtained had the steps not been taken By way of comparison economies' general theory of second best states that when one or more of a set of optimal conditions are not fulfilled there is no reason to believe that the optimum can be approached by fulfilling or approximating more closely more of the conditions rather than fewer See e g Lipsey Lancaster The General Theory of Second Best 24 Rev Econ Studies 11 1956 Whether or not there is a technological counterpart of the general economic theory of second best any such theory would appear inapplicable to the ARPANET which as a service institution comprised of many discretely operative components would seem effectively placed to enjoy incremental advances in internal efficiency This is not to say however that steps toward the network's optimum will not produce disproportionate second order effects-- both technological and economic--generally or even to say that the network's operation at its own technological optimum will more likely tend to advance social welfare than operation short of the optimum Thus it is entirely possible that a system enjoying far fewer internal efficiencies than the on - proposed will be preferable from an overall welfare perspective as measured by allocative effects on other sectors of the economy Analysis at this level however critical is beyond the scope of this paper and when the term welfare is employed it is intended only to represent the limited public interest in a system working at the described technical optimum A-16 commercial network but rather was confined to exploring ways in which the technology for such a network might most efficiently be assembled the ARPANET as now structured lies far from the network's long-range commercially based goals Thus the network's present topology merely traces the location of ARPA contractors and grantees and does not follow the pattern of commercial com puter use in this country New York Chicago Dallas and Houston are not represented and many more nodes--Los Angeles San Francisco Boston Detroit Seattle are some--would have to be added before commercial computing centers would have sufficient access to network resources to make the network a paying and fully efficient proposition Also the IMP and TIP equipment presently employed--Honeywell DDP 516 and 316 computers--represent technology that is now 10 years old and suffers important limitations on memory capacity obsolete architecture and relatively expensive components The gap between ARPANET'S subsidized and highly experimental present and its commercially profitable well-functionirg future can be expressed in terms of traditional firm decision In the ARPANET today research and development focused on quality forms the predominant objective of firm decision-making Decisions respecting price marketing techniques materials and labor for network services are almost totally absent decisions respecting rate of return to be derived from investment are made if at all only in the most abstract sense All of these characteristic firm decisions--decisions other than those embracing research and development--can however be said to exist in embryo awaiting full network development in either the public or private sector The gap between the network's present and its future can also 1 be described functionally Under this approach ARPANET'S future operations can be divided into a the provision of computer r communications services on a commercial basis and b research and development focused en first overall basal network advances and second on improvements in resource sharing echnology Given ARPA's posture as a research agency the service function A-17 A MM would appear to be best divested Given the Agency's research posture and the market's probable inability to stimulate innovation along the first basal lines it would seem appropriate to retain this functicn in ABPA Innovation of the second incremen- tal sort however because it must be directly responsive to the needs of network users would on balance seem properly lodged with the service function Obviously the bulk of firm decisions--those respecting price quality marketing materials labor rate of return--are connected with the service aid service-related innovation functions and would naturally flow with divestiture of these functions Divestiture might as noted be to another public firm--one already existing as say the Postal Service or one to be established If however for reasons already given private firms appear best placed to receive the divested service functions and their connected firms decisions some further questions arise can market decisions privately made be relied on for the maintenance of adequate quality of network service If not through what techniques should private decision be curbed ARPA of key quality decisions by less direct means By direct FCC regulation Or These questions are considered in the sections that follow See supra pp A-12-A-13 A-18 by retention in III REGULATION BY RETENTION As presently structured ARPANET is a vertically integrated public firm which though it has contracted with private firms for the performance of some functions has retained control of all While the public interest may press for ARPANET'S disintegration and for the transfer to private firms of institutional parts and functions the public welfare may also command the retention of some components in the public firm or in government generally Thus if it is felt that left to the discretion of private firms the quality of network services will deteriorate to a less than acceptable level ARPA may decide to retain institutional functions incorjxjrating critical quality decisions--basic system programming functions for example Or if perpetual private ownership of network components is seen as too chancy a route the agency could impose a specific time limitation on the transfer as in typical leasehold transactions Alternatively ARPA could hedge its grants with performance criteria requiring as a condition of continued private ownership the provision of service at some specified level J of quality THE GENERAL SETTING Although there arj important differences between the sale of government resources iv jd divestiture of the public firm it may be helpful for comparative purposes to consider two well-developed instances of government transfer of resources--the conveyance of public lands and the grant of licenses for use of ti e electro- A-19 12 magnetic spectrum The Public Lands Analogy The United States government has disposed of the bulk of its land in hopes of achieving objectives that could not as well be met if the land were retained In some cases the purposes of the government grant were general to stimulate the settlement of sparsely populated areas and to encourage private ownership of land In other cases the purposes were more spcific to en- courage the development of railroads for instance or highways and schools The federal grants have commonly been made in fee simple absolute with only the most narrow and immediate strings attached duration cf ownership has been perpetual save of course for the always prevalent prospect of eminent domain and the conditions imposed on the grant have rarely been burdensome-- development and cultivation of the ceded land for three years 14 under the Homestead Act for example These characteristics of public lands policy--perpetual ownership encumbered by few conditions--seem particularly fitting for divestit re of ARPANET'S components To begin with the general argument for perpetuity is especially persuasive in the case of the network the market tends to operate more efficient- ly when the time-tenure of the property interest is of long duration since predictions about the usable life of specialized 12 Both instances are closely and imaginatively explored the first by analogy the second directly in DeVany Eckert Meyers O'Hara Scott A Property System for Market Allocation of the Electromagnetic Spectrum A Legal-Economic Engineering Study 21 Stan L Rev 1499 1969 See generally B Hibbard A History of the Public Land Policies 1939 Davidson Government Role in the Economy 48 J Urban L 1 3-4 197U 14 43 USC Sec 164 1970 Unvler the original act 12 Stat 392 1862 the required term was five years A-20 o I capital equipment investments are thereby made less critical --although to be sure a limited but specified term would increase calculability The network owner for whom loss of the operating franchise is a relatively imminent prospect will be less disposed to take needed capital and research and development expenditures than one whose ownership is perpetual particularly since the short-term owner can expect that loss of his franchise will be attended by sharp reductions in value of plant and of knowhow The same factors together with the need for lexibility in a DeVany Eckert Meyers O'Hara Sco - - A Property System for Market Allocation of the Electromagneiic Spectrum A Legal ' conomic Engineering Study 21 Stan L Rev 1499 1531 1969 There are two arguments for the position that a renewal system will exert a contrary pro-innovation effect or at the least will be no less conducive to innovation than a scheme of perpetual ownership First if renewal is made to depend explicitly or implicitly upon some level of commitment to innovation the desire for renewal can be expected to goad the firm to undertake a desirable amount of innovation This assumes however that an administrative agency can safely prescribe level of commitment with a fair degree of certainty--an unlikely enterprise given the unpredictable nature of innovation--for to the extent that the prospective standard is uncertain the firm will discount the value of renewal by the risk of non-renewal And to the risk factor must be added the transaction costs of the renewal process The argument assumes too that both the direction and level of innovation prescribed by the agency charged with administering the renewal program will be at least as desirable as the direction and level identified by consumer decisions in the marketplace Second it may be argued that the firm whose license is not renewed is in any event in no worse a position than the firm which enjoying a position of perpetual ownership decides to sell off its assets The problem here stems from the significant extent to which the value of a firm's assets will lie in the relative modernity of the firm's software An agency decision that the firm's programs possess an insufficient degree of innovative thrust to qualify for renewal can be expected to depreciate the assets in the eyes of a prospective purchaser which to gain agency approval for its operation would probably be inclined to proffer an entirely new system or at least one that bore few of the characteristics of its predecessors' A-21 vigorously competitive setting would appear to counsel against imposing serious restrictions on permissible use The Spectrum Analogy Demonstrated widespread inefficiencies in FCC management of the electromagnetic spectrum under which portions of the spectrum are allocated to private users on a durationally limited and heavily conditioned basis weigh against the adoption of a similar approach in the ARPANET divestiture Moreover persuasive argu18 ments both for placing spectrum management on a market footing 19 and for designing experiments to test the market hypothesis share a view of the spectrum resource that is particularly applicable to ARPANET Acknowledging that spe -i_rum is a resource this view maintains that While it is true that the Government appropriated the resource in 1927 it did so not on the ground that the Government was entitled to the wealth created by use of the resource but rather on the ground that regulation was necessary 20 for the resource to be useful at all The relevant point is that ARPANET like government management of the spectrum originated in a context 'in which the market alone would have produced undesired results In the case of the spectrum unregulated uses would have overlapped to an intolerable degree in the case of packet switching technological and regulatory uncertainties would See supra pp A-12-A-13 18 See generally H Levin The Invisible Resource Use and Regulation of the Radio Spectrum 1971 Coase The Federal Communications Commission 2 J L s Econ 1 1959 cf President's Task Force on Communications Policy Final Report ch 8 at 28-40 1968 19 See IViVany Jckert Meyers O'Hara Scott A Property System for AllocctuVjn of the Electromagnetic Spectrum A LegalEconomic Engraeerin Study 21 Stan L Rev 1499 1959 20 See DeVany Eckert Meyers O'Hara Scott A Property Syt Lern for Allocation of the Electromagnetic Spectrum A LegalEconomic Engineering Study 21 Stan L Rev 1499 1531 1969 A-22 have discouraged any private firm from makii g the necessary investment in plant and research and development Once it is shown that the original market dilemma can be meliorated--as in the case of spectrum--or that it has disappeared entirely--with ARPANET'S production of the needed basal innovation--largely unconstrained divestiture to the private sector becomes timely Guidelines for Divestiture All of this suggests two principles that could guide the disintegration and parcelling out of the ARPANET First com- ponents of the network could be made available to bidders on a basis that will enhance innovation and flexibility within their firms this means specifically ownership unlimited in time and bounded only by the most slender expedient conditions Second if the market is to be relied on to encourage the most efficient use of the network's components then bidders for these components must be treated on the terms most consonant with maintenance of competitive conditions simply no single bidder should be given from ARPANET assets a competitive advantage not available to all others on equal terms The second principle bears on the question how the network is to be divested and the first on the question how much It is to the question of how that discussion now turns THE ASSETS AND THEIR DIVESTITURE ARPANET possesses two institutional elements of value to private bidders one easily appropriable the other less so The first appropriable class of assets consists of hardware--IMPS TIPs and interfacing equipment situated at the IMPS for connection with the hosts--and of software--undisclosed proprietary data in 21 programs essential to operation of the system The second less 21 Omitted from this list of assets are the main computers situated at each HOST which under varying arrangements have been provided to the HOSTS by ARPA The disposition of these facilities has been excluded from this study A-23 o- appropriable class consists of the network's going value which for present purposes can be described as a congeries giving its possessor some competitive advantage in making those operational 22 decisions central to any firm private or public Transfers of assets in the first class can be vieweo as involving the products of decisions already made--investment decisions respecting hardware and innovation Transfers in the second claso involve instead the power to Fake decisions in the sense that some or all of the decision-making powers divested are for successful bidders enhanced in a way that they are not for unsuccessful bidders and nonbidders This is simply to say that if as between A and B com- petitors in the packet switching business A receives ARPANET'S goir f concern value and B does not A will with respect to some or all of its operational decisions--pricing and marketing for instance--enjoy advantages that B_ will not The Appropriable Assets The decision whether ARPA should retain existing network hardware--TIPS IMPS and interfacing equipment--or whether the hardware should be divested and if so on what terms calls for some relatively straightforward judgements respecting desired levels of barriers to entry and of innovation Thus if relatively free entry is desired ARPA could retain title to the hardware and lease the needed equipment to successful bidders on terms less forbidding than those entailed by either initial capital outlay or indebtedness Alternatively if some hurdle to entry is percieved as appropriate--possibly to separate the serious and resourceful entrepreneurs from those that are undercapitalized--then outright sale of the present stock of hardware may prove to be a desirable screening technique And ARPA may under either the sale or lease approach employ the price mechanism to modify the conditions for hardware innovation by adjusting hardware prices up or down the agency can increase or decrease the relative desirability to users 22 See supra p A-4 A-24 ' - IMMI MJM of purchasing or committing research and development resources to network hardware f I Technical knowhow concerning the network's most efficient operation under present conditions might appear generically indistinguishable from ARPANET'S hardware assets vitally different In fact it is in some instances the knowhow is completely inappropriatee--because already in the public domain--while in others it is too fully appropriable--and sequestered by private ARPANET contractors is unavailable even to ARPA In the first inappropriable class of knowhow there is of course nothing to be divested all the information involved being freely and publicly available The second class on the other hand raises a number of serious questions for divestiture policy Should ARPA let this proprietary data remain in their present hands for the exclusive use of private contractors Should compulsory licensing at specified rates be required Should ARPA appropriate the proprietary data to itself and if so should it license their use or should it inject the data into the public domair as it has done with other network knowhow However the other questions are resolved it seems clear that the questions whether network proprietary data are to be left to r the exclusive use of ARPANET contractors who become network bidders should be answered i - the negative To allow presert contractors who become future bidders to retain as their own proprietary data developed in the course of ARPANET'S developmental stage would confer on these firms a competitive advantage over entrants not occupying this privileged position and possibly deter entry by disadvantaged bidders altogether While for reasons already given some barriers to entry may be appropriate this will be the case only if the height of the barriers is uniform It can of course be argued that present contractors who become network participants may be expected to behave like prudent patent and trade secret proprietors generally licensing the use of sequestered information to others at the profit-maximizing price that the network infermation owner is a vertically integrated producer-supplier gives it no reason to discriminate A against non-integrated suppliers by entirely refusing them access to the information product Under this argument the question then becomes one of the price charged The general substituta- bility of this sort of information indicates that the price would not be one from which much in the way of monopoly profits could be extracted Perhaps because it imputes rational economic behavior to network managers and because it ignores widespread but erroneous assumption respecting the anticompetitive behavior of vertically integrated firms this argument appears to make considerable 23 sense ' Moreover there would seem to be no significant economies of integration--no internal savings of external transaction costs-- peculiarly attributable to transfers of proprietary data and consequently no reason for the integrated network to favor its 24 own branches co the exclusion of outside firms What may make a difference in terms of long run monopoly effects is not so much the fact that possession of proprietary data will give network contractors a competitive edge as the fact that because the underlying research and development was government financed the edge was obtained risklessly Lacking the need to finance past investment from current data revenues the advantaged contractors may be more inclined to hold the data off the market particularly if they believe that the exclusionary tactic together with their vertically integrated posture will pose 25 significant barriers to entry 23 See generally Bork Bowman The Crisis in Antitrust 65 Colum L Rev 363 366-368 1965 24 See Allen Vertical Integration and Market Foreclosure The Case of Cement and Concrete 14 J L S Econ 251 255-272 1971 A network developer of proprietary data may however respona to the general danger that through leakage his trade secrets will lose their secrecy and consequently their legal protectability To the extent that leakage appears less likely to occur in the internal transfer of data than through their licensing to outsiders this factor may be seen to produce one integration economy 25 Compare Blake Jones In Defense of Antitrust 65 Colum L Rev 377 392 1965 A-26 The Less Appropriable Assets Whatever the market value of ARPANET'S appropriable hardware and software assets it is the less appropriable assets that probably possess the greatest attraction for proa ctive network bidders This is particularly ironic since once hardware and software are excluded there is little or nothing left in the ARPANET inventory that can be characterized as assets in traditional terms The significant research and development that has brought ARPANET from an idea to an operational entity is with the relatively limited exception of what is being withheld by private contractors all in the public domain Going concern value is scarcely discernible particularly because if the network is structured along competitive lines there will be no single firm that can properly call itself a successor to ARPANET What is left is a customer base and goodwill of a highly fractionated sort far less say than would be involved in the sale of assets of a popular periodical The customer base component of the ARPANET inventory has two aspects The first derives from the needs of network users in quality and reliability of service These needs which will be particularly pressing during the period of the network's transition from public to private ownership might be met by a program of ARPA endorsement with customers naturally drawn to those firms that 4 I bear some imprimatur of ARPANET affiliation Yet while proprie- tary data can be effectively subjected to licensing schemes reputation cannot be so easily marketed The benefits associated with trademark licensing popularly thought to serve a reputational guarantee function are largely unavailable in this context And in any event introduction of a franchising system even if it could be mounted efficiently might well contradict demonstrated interests in genuine competition Because of the difficulties and imperfections associated with a trademark licensing scheme it can be expected that present ARPANET contractors who successfully bid for a share of the network will because of the history of their association attract the bulk A- 27 T of prospective users particularly during the critical interim period This means of course that present contractors will by reason of their past work for ARPANET enjoy a competitive advantage with respect to new entrants not unlike the one they would enjoy were they permitted to retain proprietary data Here however it would seem wasteful and not at all consonant with the needs of users during the interim period to achieve uniformity in barriers to entry by requiring present contractors to disqualify themselves as bidders and to set parity at the lowest common denominator of performance A-28 IV IRECT REGULATION The delivery of telecc-mmunications is in this country left to private enterprise Elsewhere in the world these services are 26 commonly provided by government through public firms The international comparison underscores the boldness of American policy and the thrust of its governing presumption wherever feasible the provision of goods and services no matter how vital should be left to private firms permit a corollary Both the policy and presumption wherever feasible private firms should be allowed to make their operational decisions unconstrained by government regulation In the case of the ARPANET this means that though presently an embryonic public firm there is every reason for it not to develop as such It means further that if the network is handed over to private firms government control of firm decisions should be kept to the necessary minimum The general presumption against regulation gains compelling force from some historical and economic factors surrounding communications regulation generally and network conditions specifi27 cally If the divested ARPANET is to be regulated at all regulation is most likely to come from the FCC which in the area of telecommunications characteristically works on an all or nothing basis if a firm is regulated it is as a common carrier with application to telecommunications common carriers like the 26 Several of the Canadian provinces stand with the United Status as the important exception to this general approach 27 Sne Posner Natural Monopoly and its Regulation 21 Stan L Rev 548 592-620 1969 for a generally balanced assessment of the costs and benefits associated with regulation of public utilities and common carriers A-29 -A the Bell System and Western Union 28 Little if any room is left for incremental regulation tailornade to fit the needs of the regulated form room that is vitally needed if the emerging network is to adapt effectively to market conditions FCC control of the budding network seems inappropriate for another reason Often inhospitable to newly emerging industries and technologies that appear to threaten the economic security of entrenched already regulated firms the Commission has been known to regulate prospective entrants to a point at which entry itself is all but impossible hindsight This concern may be softened somewhat by the protective stance has most frequently been taken in the broadcast context in the form for example of rebuffs to 29 CATV's perceived assaults on the integrity of VHF operations defenses on the common carrier side have not in recent years been nearly so high 28 The prospect that regulation will or can be structured under Title III Special Provisions Relating to Radio appears sufficiently unlikely not to warrant consideration here 29 See generally Goldstein Information Systems and the Role of Law Some Prospects 25 Stan L Rev 449 461-470 1973 Similarly the prospects for UHF's growth were early stunted by the Commission's failure to put the new industry in a position to compete with the already established UHF system See generally Note The Darkened Channels UHF Television and the FCC 75 Harv L Rev 1578 1962 Webbink The Impact of UHF Promotion The All-Channel Television Receiver Law 34 L Contemp Prob 535 1969 For example in Microwave Communications Inc 18 FCC 2d 953 1969 petitions for reconsideration denied 21 FCC 2d 190 1970 and the ensuing rulemaking First Report and Order in Docket No 18920 29 FCC 2d 870 1971 the Commission granted free entry to specialized carriers which in competition with Bell and Western Union proposed to offer point-to-point microwave relay services specially tailored to meet the needs of the business and data transmission communities In neither proceeding was the Commission persuaded by the existing carriers argument that entry would enable the specialized carriers to reap the rewards available in highly profitable markets an argument rooted in regulated industries' common practice of differential pricing employing supranormal profits from one area to subsidize average cost While on balance systems of cross-subsidy pricing generally may he demonstrated to do more harm than good the Commission A-30 REGULATION CONDITIONS AND CONSEQUENCES Together with these other considerations there are two particularly salient reasons for FCC abstention from network regulation First the market to be occupied by the divested network possesses few if any of the nature - monopoly contours that traditionally justify imposition of public utility or common carrier treatment Second in an area critical to ARPANET'S success--innovation--the performance'of firms under regulatory constraint has been seriously questioned Natural Monopoly A natural monopoly is said to exist in markets where demand can most efficiently be met by a single firm The cost effi- ciency of the single firm in natural monopoly markets is a function of significant economies of scale unit costs declining as production scale increases and of relative capital intensity skirted the basic welfare question and rested its decision and order instead on the view that to permit entry would not only promote satisfaction of presently unmet needs but would also spur the existing common carriers to provide improved more competitive service in the areas to be served by the new specialized carriers For the cross-subsidizer the obvious competitive response to a specialized carrier's cream-skimming is the one subsequently made by Bell and Western Union--dropping prices in the formerly highly profitable markets to meet or undercut those of the newcomer subsidizing these drops through increased prices elsewhere Compare Baumol Walton Full Costing Competition and Regulatory Practice 82 Yale L J 639 1973 with Noll Rivlin Regulating Prices in Competitive Markets 82 Yale L J 1426 1C73 Particularly if as has been shown competition will best serve the interests of network users there is every reason for the MCI rationale to apply with at least equal fore 1 to answer any charge that a divested ARPANET would improperly be skimming the cream from established common carrier operations A-31 with the ratio of fixed to variable costs being continually high To the extent that these factors are present a natural monopoly condition exists and a single firm is recommended regulation of the firm as a public utility or common carrier is in turn seen as required to prevent the abuses popularly associated with monopolies Tne market to be occupied by the divested ARPANET appears to possess none of the characteristics of natural monopoly in a sufficient degree to warrant divestiture to a single firm Two classes of capital outlay will be essential to the network's operation a national transmission system consisting of telephone lines microwave and communications satellites and terminal to network interface hardware and software While the capital costs are high in both classes the critical point for the ARPANET is that the necessar-' capital outlays have been and will continue to be made outside the network industry The transmission facilities to be employed by the network are either already in place--as in the case of the Bell System--or are being developed by non-network firms--specialized common carriers for example or domestic 32 satellite entrepreneurs Hardware and software costs while 31 Essentiality to the community of the service in question has been cited as a third factor indicative of natural monopoly conditions Irwin The Computer Utility Competition or Regulation 76 Yale L J 1299 1313 1967 This factor appears however to be not so much a predicate for natural monopoly as a description of some of the services provided by some public utilities and common carriers A number of essential services and products--health care and food are examples--are provided under truly competitive conditions while many of the services provided under natural monopoly conditions many of Bell's regulated offerings for example can only be characterized as nonessential 32 One result to be expected from the FCC's MCI decision and the ensuing rulemaking see supra n 30 is the proliferation of special function transmission systems throughout the country frequently existing side-by-side Especially as augmented by domestic satellite transmission capabilities see generally Mathison Walker Regulatory and Economic Issues in Computer Communications 60 Proceedings of the IEEE 1254 1264-1268 1972 the future transmission picture reveals a multiplicity A-32 tk they may be incurred completely within the network need not be Hardware is produced exclusively by large firms outside the network industry and software by firms both inside and out lease and licensing mechanisms are available to spread out costs for the two items and to reduce entry and exit barriers At the same time while packet technology will push the most significant variable cost of remote data services--communications--well below present levels there is every reason to expect that particularly in view of the low fixed costs the fixed to variable cost ratio can be expected to be far below the level at which natural monopoly characteristics begin to surface Innovation While the argument that monopolistic firms are characteristically disinclined to innovate in their operations and in their of competing transmission services including Bell and Western Union services from which the user of transmission facilities will be able to select the one best priced and situated to meet his individual needs % What is critical about ARPANET'S place in this picture is that it will be an entity consisting of users of transmission facilities--flexibly employing telephone microwave and satellite links--and not a carrier--providing the necessary links This suggests not only that the network market lacks the natural monopoly characteristics that traditionally call for common carrier treatment but also that 1 existing common carriers would have good reason to encourage entry by a large number of network firms which in comp t-ing for the sale of new communications services can be expected to increase overall use of common carrier facilities and 2 to the extent that an existing carrier's objection to the entry of new firms is grounded in its own hopes of entering the network business the argument reveals a carrier attempt to reach into untapped fields rather than a network attempt to enter the already tapped common carrier industry A-33 products and services has been persuasively answered 33 this does 34 not mean that regulated monopolies will behave similarly The regulatory practice of tying revenue to costs naturally produces seme disincentive to innovate toward efficient operations And while innovation directed toward the development of new products and services need not be similarly deterred even here the requirement of regulatory agency approval may be a dissuasive force effect There are some counters to this general disincentive Thus because regulated rates are almost always based on the firm's past performance and are set periodically rather than continually the firm has some reason to innovate and cut costs The argument that firms enjoying a monopoly position will be counter-innovative or at least will invest in the least efficient forms of innovation rests on a number of assumptions-- among them that monopolists because they are less cost-conscious than competitive firms will be less concerned with cost-reducing and efficiency-promoting innovations that the monopolist will either underinvest in research and development generally or will over-invest to forestall entry when part of its monopoly narket is competitively threatened and that the monopolist will invest in research nd development designed to buttress its monopoly position by extending scale economies and reinforcing other barriers to entry See for example Shepherd The Competitive Margin in Communications in W Capron ed Technological Change in Regulated Industries 86 1971 For a particularly effective rebuttal of the argument see Posner Natural Monopoly and Its Regulation 21 Stan L Rev 548 577-584 1969 34 Although it has been commonly supposed that regulation tends to inhibit innovation an apparent paradox is also recognized--if regulation has inhibited the pace of innovation why have all the regulated industries enjoyed long-term productivity increases that are above the national average and certainly higher than those in most manufacturing industries Capron ed Technological Change in Regulated Industries 3 1971 In part however this general level of performance can be attributed not to regulation but to the surrounding natural monopoly conditions that called for its exercise--capital intensity economies of scale--conditions that themselves would appear to enhance innovation IcL at 221 Regulation's counter-innovative effects in the communications industry have been carefully documented in Shepherd The Competitive Margin in Commun cations in id at 86 which concludes that more rather than less competition in the industry will best conduce to a desirable pa e and direction of innovation A-34 IM in the inter in before new rates are set thus the presence of regulatory lag may exert some pressures toward economy in operations While innovation doubtless occurs in the rate base regulated firm especially with respect to the creation and capture of new markets this does nor mean that the level of innovation will be optimal or far more important that the innovation produced will be of the proper kind Thus for example because its revenues are tied to its level of investment the regulated firm can be expected to seize every available opportunity to enlarge its rate base a capital intensive bias that may lead it to prefer research and development directed toward capital intensive but 35 comparatively inefficient production and service processes Misdirection of investment in innovation may also occur if it is general public relations not specific consumer needs that supply the motive for invention This indicates only that the rate and direction of innovation are likely to be suboptimal in regulated industries whose firms are vertically integrated Where there is some disintegration and the firms supplying the regulated firms operate in a competitive environment the degree of innovation with respect to the goods and services supplied that is present in other competitive The general bias commonly called the Averch-Johnson effect is considered in Averch Johnson Behavior of the Firm Under Regulatory Constraint 52 Am Econ Rev 1052 1962 Bell's expenditures on the Picturephone--tremendous when compared with its commitment to the development of digital transmission services see Mathison Walker Regulatory and Economic Issues in Computer Communications 60 Proceedings of the IEE 1254 1255 1972 --provides a good example of wrongly directed investment in innovation The quality of Picturephone service is far from adequate not because the research was done on the cheap nor because the system lacks sophistication and considerable ingenuity but rather because the extensive network of transmission facilities to which the Bell System is tied are just not set up for two-way video transmission Much less research and development would have been needed to produce a system able to -Tk better on some other transmission basis A-35 Ma Ma lMHlfl sectors may be expected to prevail 37 The question then is whether if certain components of the divested ARPANET are placed on a regulated common carrier basis other components critical to innovation can be isolated and left in a competitive setting Unfortunately because it is an entirely new system that is involved and because it is in the nature of significant innovativ n that its outcomes and contours cannot be known at the outset little can be accurately said about the proper locus of innovation until consumer needs become more defined and the outlines for responsive innovation become more clear What can be said with somewhat more certainty is that other things being equal more in the way of appropriate innovation stands to be lost by placing any segment of the divested firm on a regulated footing rather than on one that is competitive REGULATION THE AUTHORITY OF THE FCC Arguably Federal Communications Commission has the power to 37 To some extent this phenomenon was credited by the FCC in Use of the Carterfone Device in Message Toll Telephone Service 13 FCC 2d 420 1968 Issuing a sweeping condemnation of the carriers' foreign attachment tariffs which prohibited use of hardware not obtained from carrier affiliates to the extent that they were unnecessary to the maintenance of system integrity the Commission immediately stimulated competition and innovation in the attachment hardware industry See generally Irwin The Telecommunications Equipment Market--Public Policy and the 1970's Fall Joint Computer Conference 269 270-272 1970 Carterfone is at least indirectly relevant to the prospects for a divested ARPANET on two counts first in its broadest aspect the decision reflects a policy that when faced with new technological entities seeking connection with present carrier facilities the Commission will place the burden on the carrier to establish that connection would materially impair the carrier's services and not on the proponent to establish that it would not Second by stimulating competition and innovation in the attachment hardware industry Carterfone has dramatically increased the technological options available to network participants in terms of both their freedom to fashion equipment to meet their special requirements and to purchase needed systems and devices in the market at competitive prices A-36 regulate the components of a divested ARPANET 38 Title I of its enabling legislation which empowers the Commission to perform any and all acts make such rules and regulations and issue such orders not inconsistent with this chapter as may be necessary 39 in the execution of its functions may be interpreted to justify the regulation of institutions whose activities not falling squarely within the scope of the Act nonetheless impinge upon the Commission's regulation of activities that do--telephone common ccirriage and radio transmission Perceiving a network threat to the integrity of the telephone system--by sloppy interconnection or diversionary pricing for example--the Commission might fine in Title I the required authority to regulate the network's activities generally Similar reasoning formed the predicate for the Commission's early regulation of CATV functions 40 and although there are important differences between the CATV context and the present one it is significant that the rationale was expressly sustained by the Supreme Court in one of its in41 frequent reviews of an FCC decision 38 The FCC's authority under Title II is of course confined to iiterstate and foreign carrier operations It is entirely possible that the states probably through their public utility commissions and even some municipalities will attempt regulation of intrastate network activities as they have done to varying degrees with CATV systems See generally Barnett State Federal and Local Regulation of Cable Television 47 Notre Dame Lawyer 685 1972 39 47 USC Sec 154 i 40 Second Report and Order in Dockets 14895 15233 15971 2 FCC 2d 725 1966 41 United States v Southwestern Cable Co 392 U S 157 178 1968 There is no need here to determine in detail the limits of the commission's authority to regulate CATV It is enough to emphasize that the authority which we recognize today under Sec 152 a is restricted to that reasonably ancillary to the effective performance of the Commission's various responsibilities for the regulation of television broadcasting A-37 Whatever the case for the existence of an incidental statutory power to regulate the divested network it is by no means clear that the statutory language and the underlying legislative history command or even warrant an exercise of the specific power to characterize the network as a common carrier The Act's definition of common carrier is singularly unhelpful--' common carrier' means any person engaged as a common carrier for hire - and the legislative history is only slightly more enlightening the statutory definition was said not to include any person if 43 not a common carrier in the ordinary sense of the term The ordinary r ense of the term as it was understood by the Aci ' framers may be generalized from three early instances of common carriage ferryboats railroads and the telephone system Elements common to the three include a service available to the public generally for transporting persons things or Kiesiges in unaltered for a from OuC place to another ARPANET would depart from these traditional contours in all important particulars use of the system will in tne near term at least be confined to commercial and gover'iment buyers the very reason for use will be to obtain some sig iificant alteration of the message conveyed often with additional data returned and though to be sure messages will travel from one site to another the ticket will as likely as not be round trip with processing not switching the significant function at the distant end 42 47 USC Sec 153 h The full definition reads 'Common carrier' or 'carrier' means any person engaged as a common carrier for hire in interstate or foreign communication by wire or radio or in interstate or foreign radio transmission of energy except where reference is made to common carriers not subject to this chapter but a person engaged in radio broadcasting shall not insofar as such person is so engaged be deemed a common carrier Statement of Managers ou the Part of the House Conference Report on Communications Act of 1934 H R No 1918 73d Cong 2d Sess 45-46 1934 A-38 The Computer-Communications Inquiry Although the FCC's own interpretations of its statutory mandate lend few more guidelines for answering the questions of whether and to what extent the divested network is to be regulated 44 its Computer-Communications Inquiry initiated in 1966 at least provides a starting point The Inquiry which culminated in a 45 final order in 1971 explored some of the knotty issues raised at the junctures between the telecommunications industry and the computer and data processing industries Among other questions the Inquiry considered whether services combining data processing-- previously unregulated--and communications functions--pervasively regulated--should be regulated by the FCC Avoiding the broader questions raised the Commission decided only that regulation would be inappropriate for certain of the new forms of service some of which--the so-called hybird data processing services -- are markedly akin to the services that will bo p- ovided by the divested ARPANET 4 Defining hybrid service as an offering of service which 47 combines Remote Access data processing and message-switching 48 to form a single integrated service the Commissicn irew the v 44 Notice of Inquiry Docket 16979 7 FCC 2d 11 1966 45 In the Matter of Regulatory and Policy Problems Presented by the Interdependence of Computer and Communication Services and Facilities Docket No 16979 Final Decision and Order 28 FCC 2d 267 1971 '- t o f I 46 I n view of all the foregoing evidence of an effective competitive situation we see no need to assert regulatory authority over data processing activities whether or not such services employ communications facilities in order to link the terminals of subscribers to centralized computers 28 FCC 2d 291 298 1970 47 ' Remote Access Data Processing Service' is an offering of data processing wherein communications facilities linking a central computer to remote customer terminals provide a vehicle for the transmission of data between such computer and customer terminals 47 C F R Sec 64 702 4 1971 48 47 C F R Sec 64 702 5 1971 A-39 regulatory line between Hybrid Data Processing Services and Hybrid Communication Services The hybrid communication service defined as a hybrid service offering wherein the data processing capability is incidental to the message-switching function or 49 purpose would under the Order be subjected to regulation The hybrid data processing service described as a hybrid service offering wherein the message-switching capability is incidental to 50 the data processing function or purpose would for the present remain unregulated Expressly rejecting the argument raised in several quarters that it was obligated by statute to regulate the 'hybrid service' as defined insofar as such service contains a communication component the Commission appears to have confirmed some rules of thumb it had earlier formulated for distinguishing between the two types of hybrid service If the package offering is oriented essentially to satisfy the communications or message-switching requirements of the subscriber and the data processing feature or function is an integral part of and incidental to message-switching the entire service will be treated as a communications service for hire whether offered by a common carrier or non-common carrier and will be subject to regulation under the Communications Act One applicable test will be whether the service by virtue of its message-switching capability has the attributes of the point-to-point services offered by conventional communications common carriers and is basically a substitute therefore Another test will be the extent to which the message-switching feature of the service facilitates or is related to the data processing component or whether such message-switching is essentially independent of such data processing 52 How the ARPANET would be characterized under these tests is far from clear Two commentators have concluded though without 49 47 C F R Sec 64 702 5 ii 1971 50 47 C F R Sec 64 702 5 i 51 Final Order in Docket No 16979 28 FCC 2d 267 277 1971 2 Tentative Decision in Docket No 16979 28 FCC 2d 291 305 1971 A-40 1971 much further explanation that the ARPANET ''if offered on a commercial basis to the public at large would under the Commission's present rules have to operate as a common carrier The issue is however more slippery than this assertion might indicate for though to be sure point-to-point service is the essence of the network the service is hardly a substitute for those offered by conventional common carriers and as an economic matter the message-switching feature is at best secondary to the data processing aspect with which it is closely related ARPANET The Overlooked Threshold Questions The real conceptual difficulty steins from the fact that ARPANET is generically distinct from the types of systems the Commission envisioned in the Inquiry Where the Final Order con- templates a longitudinal division in a system's services entailing say first transmission then message-switching then processing ARPANET contemplates a division along latitudinal lines Point-to-point communication is the crux of the system but from start to finish the messages transmitted will be processed through methods including disassembly reassembly and changes within the bit packets and sometimes through encryption That data processing of a more complex magnitude will be performed at various points in the system in no way renders the packeting process any the less data processing Thus it is conceptually more accurate to characterize the network as itself a unitary fully integrated computer with communication facilities employed 54 internally to link one function to another 53 Mathison Walker Regulatory and Economic Issues in Computer Communications 60 Proceedings of the IEEE 1254 1259 1972 54 Characterization of network operations as hybrid data processing services rather than as hybrid communications services may possess for network firms specific and important economic consequences distinct from those flowing from the Commission's general regulatory activities Under the authorized user provisions of their tariffs existing common carriers are constrained not to lease circuits to customers whose use would involve third party A-41 These important technical differences aside there are important policy reasons for withholding regulation from the divested network Focusing on the question whether a new service's relation- ship to existing common carriers warrants regulation of the service the Commission has overlooked a larger threshold issue whether the new service possesses those natural monopoly contours that historically have justified imposition of common carrier status communications essentially replicating the type of service provided by the carrieis themselves The effects upon network firms of a carrier dti ia1 ui der its authorized user provisions are different and certainly more immediate than Commission c jracterization and reguli '- ion of the firm as a conrnon carrier entry is chilled from the outset at existing carriers' private initiative The history of the Bunker-Ramo Corporation's efforts to lease Bell and Western Union lines for its Telequote IV service is instructive and is recounted in Irwin The Computer utility Competition or Regulation 76 Yale L J 1299 1306-1308 1967 D Smith The Interdependence of Computer and Communications Services and Facilities A Question of Federal Regulation 117 U Pa L Rev 829 848-849 1969 and more recently in Comment Federal Communications Commission Regulation of Domestic Computer Communications A Competitive Reformation 22 Buffalo L Rev 947 961 1973 55 It is at best risky to speculate on the role that nomenclature plays in regulatory decisions but it is entirely possible that the chance academic characterization of early networks cognate to ARPANET as computer utilities played some part in the Commission's decision to view them as indicating regulation as public utilities The term has for better or worse fallen into common parlance See e g D Parkhill The Challenge of the Computer Utility 1966 Irwin he Computer Utility Competition or Regulation 76 Yale L J 1299 1967 Paul Baran's early observation that in essence computer 'utilities' are not utilities deserves more attention than apparently it has been given The computer 'utility' user is not restricted to doing business with any one company If you are not satisfied with your service or are concerned about price you can always 'go' elsewhere Similarly any single computer installation is not forced to serve all potential customers on an equal basis The big customer may expect preferential treatment either in terms of price charged or speed of service P Baran The Coming Computer Utility--Laissez-Faire Licensing or Regulation 1967 A-42 It is this overarching question--whether the market to be occupied by ARPANET will itself possess natural monopoly contours--and not the subordinate one--whether existing comuvon carrier service will somehow be prejudiced by the network's operation--that properly forms the starting point for inquiry Having the tail wag the dog is not an unpardonable act in all circumstances but when it results in foreclosed consideration of central economic questions it deserves to be discouraged c A-43 REGULATION BY THE MARKET The preceding discussion suggests that interests in the efficient supply of a wir e range of computer communications services will probably best be served by vi ARPANET situated in a vigorously competitive environment free from significant government control either through regulation as a common carrier or through ownership as a public firm This means lirst that government should at some early point shed the bulk of ARPANET'S components into the private domain retaining if any only those few components--the basal R D function is one--perceived as critical to the network's continued functioning in the public interest It means too that divestiture should be to several firms rather than to one To be sure even if a network firm were structured along monopoly lines there would be some element of competition in the sense that bidding for the monopoly franchise would be competitive and--if the franchise were durationally limited--continual From the available evidence though it seems unlikely that competition in this form would stimulate a high enough level of continued and properly focused investment in innovation and a sufficient degree of diversity and economy in services to justify taking this route over the more thoroughly competitive one Third barriers to entry should probably be kept low and more important kept uniform ARPA must be vigilant to assure that all prospective entrants are given equal access to existing technical knowhcw whether developed within ARPA or by its contractors Equality of access should effectively deprive any single firm of the sort ot technological headstart that would likely lead to a monopoly position under present conditions and A-44 should help to generate the needed degree of competition and diversity both nationally and regionally Assuring entry by firms of the appropriate size and number- -the one justification for giving a firm or small number of firms the competitive advantage special access would entail--appears from the facts available to be attainable without this artificial inducement Although competitive conditions can be expected generally to discipline a network's firm's operational decisions--decisions respecting price quality marketing material labor and investment--toward achieving the larger objectives established for the network the degree of success achieved by market forces cannot be expected to be uniform for all decisions There may be some concern for example that the competitive firm's decisions affecting the quality of its services will not always comport with larger needs Specifically there are two ways in which decisions on quality may be perceived to depart from desired norms First it may be feared that competitors will shave the quality of their services to a point beneath the standard of reliability essential to the network's integrity The concern in this respect is that absent regulation breakdowns in quality might with distressing frequency go undetected until after their harm is done When the harm is to highly sensitive interests and threatens to be on a massive and unsettling scale the need for prospective quality maintenance by an institution other than the private firm may be found compelling I is perceived needs of this sort that sus- tain the federal Food and Drug Administration and that may be seen as calling similarly for government supervision of network performance quality Second because any element in the network should be able to interconnect with any other element in the network there is a need for compatibility and consequently for standardization of interconnection formats the obvious comparison being to the early need for uniform gauge railroad track throughout the nation if rail transport was to enjoy a proper degree of efficiency A-45 Indecision and wrangling among network members as to the appropriate interconnection standard to be employed could lead to serious dysfunctions at the network's outset In a related area home-use electronic video recording conflict over standards has plagued the nascent industry for years and may be the single most prominent reason for its failure so far to achieve viability Even though there will be certain natural economies to standardization in the case of the ARPANET so that the various systems can be expected to shake down over time to a single standard the critical questions are how much time this will take and whether the losses to be sustained in the interim are sufficiently outweighed by the generalized advantages of strict reliance on the market Standardization and the maintenance of prescribed levels of service could be achieved under the techniques of retention or regulation discussed in earlier sections of this paper Govern- ment could for example retain responsibility for developing interface and service standards and could establish ai agency to enforce ' ne rules adopted At least two other techniques rooted in market rather than regulatory functions are available for quality control within the evolvfd ARPANET Under the first technique government control would be accomplished through government purchase of services from the network with the appropriate performance standards presumably initiated by the government but tempered by negotiation written into the service contract this technique may be called regulation by purchase Under the second technique which may be called regulation by cooperation quality control would hi administered by a cooperative board or trade association consisting of members elected by network partici pants RF ULATION BY PURCHASE The federal government will be a major--and in the near term the predominant--purchaser of network services A-46 Indeed it may be said with some accuracy that the single most valuable asset to be received by successful ARPANET bidders is an implicit government agreement to purchase their services As a purchaser government may be expected to exert some special influence on network operation Contracts might for example call for a government priority on the network's facilities and in case of vital public need for preemption of network time Special government needs may also require the creation of bypasses in heavily traveled areas to assure the unimpeded flow of government and other user messages when peak periods coincide with emergency conditions It is the more general formative effects of governr jint purchase that are of interest here The government particularly if it acts through a single broker rather than through a number of independent departments will be in a unique position to affect the way in which network decisions on quality are made Just as government could through retention or regulation prescribe interface standards and levels of performance in terms of errors permitted reliability and data rate so it could by specifying ts criteria as to any or all of these in its purchase contracts stimulate network participants to meet these standards voluntarily The extent to which performance requirements in government contracts will have a spillover effect establishing network performance at the same level for all users private and public fill depend upon economies of network operation not yet fully discernible If the hardware and software built to government quality specifications might efficiently be deployed to meet the needs of private consumers as well and if economies of scale counsel against the construction of redundant facilities for the commercial sector then the spillover will be complete and the regulatory consequences most effective The economies involved in standardi- zation suggest that governme-t initiative in this respect will prove decisive Whether government leadership on other quality Decisive because so long as no network operator has an investment in and consequently commitment to any interface standard each will be better off in adopting a standard that it knows will be adopted by all or most A-47 1 i issues will be equally influential is open to question Examina- tion and comparison of the effects on private suppliers' decisions of compliance with government purchase requirements as for example Department of Defense purchasing specifications for pharmaceutical and medical supplies may prove instructive in this respect The advantages of regulation by purchase when it possesses extensive if not complete spillover effects stem largely from its flexibility and specificity To begin with a market situation suffering the inflexibilities of the regulatory process would obtain only if government were a monopsonist a position which in connection with its purchase of network services it would not occupy Also while some of the elements of a Turkish bazaar do creep into administrative hearings on proposed common carrier tariffs and while i_ regulated network managers might be expected to have some say informally and through the hearing process in the promulgation of rules governing their firm's activities the relevent interests would on balance probably be advanced and accommodated more effectively around the bargaining table at opposite sides of a proposed contract Also regulation's tar baby effect already 57 alluded to under which regulation of one aspect of an industry may quickly lead to the need for regulation of another and still another until a situation approaching complete government ownership results can be avoided through the inclusion in purchase contracts of those specific parameters on which performance is desired REGULATION BY COOPERATION General responsibility for prescribing and supervising the protocols and quality of network service might alternatively be vested in a cooperative organization or trade association consisting of network participants with executive responsibility delegated a governing board in which representation would be based on among other factors regional situation and user orientation Membership in the association would presumably be open to all S7 See supra p A-8 A-48 network operators the form of governance would be democratic The problem with the cooperative approach lies in the phenomenon observed by Adam Smith two hundred years ago that people of the same trade seldom meet together even for merriment and diversion but the conversation ends in a conspiracy 58 against the public or in seme contrivance to raise prices If trade associations are themselves generally unassailable on antitrust grounds their decisions nonetheless invite government scrutiny for anticompetitive effects And while the courts the Justice Department and the Federal Trade Commission have all shown some tolerance for cooperative endeavors respecting research exchange of technical information advertising standards 59 and safety programs even some of these decisions may be proscribed if untoward effects on price and quality competition are demonstrated Thus for example cooperative programs to standardize products and services while frequently sustained necessarily produce some anticompetitive effects the firm serving consumers who desire substandard services may be prejudiced and the standard itself may by restricting supply tend to rigidity price structures Also though one commentator has counselled that attempts to promote standardization should for this reason probably be limited to such noncontroversial matters as the safety and possibly the durability and efficiency of the product even safety programs can be faulted on policy grout the consumer who prefers a lower price to more safety may find CO A Smith Wealth of Nations Book I ch 10 pt II 1776 59 See Monroe Practica Antitrust Considerations for Trade Associations 1969 Utah L Rev 622-623 1969 See generally Borowitz Joint Business Actions by Competitors Are Any Permissible 32 Ohio State L J 683 689-698 1971 Levin The Limits of Self-Regulation 67 Colum L Rev 603 633-635 1967 Monroe Practical Antitrust Considerations for Trade Associations 1969 Utah L Rev 622 625 1969 A-49 his range of choice constricted While questions of antitrust liability pervade the cooperative technique a question larger than restraint of trade stands at the threshold whether decisions which on their face seem best placed under cooperative control should in fact be centralized One question whether a cooperative mechanism should be employed to allocate among HOSTS differentially rewards corresponding to the utility of their programs may prove particularly nettlesome Absent some system of property rights the HOST who invests heavily in the development of a new and useful program will the program's high utility notwithstanding be unable to recoup his investment a competitor who has not similarly invested will simply cadge the innovator's technique and market it at a price equivalent to his marginal cost--a price that given the low costs of replicating information can be expected to be well below the innovator's average cost Unable to recapture his investment through the price mechanism the prospective innovator will be disinclined to innovate altogether and one of the network's objectives a high 62 degree of program innovation would be defeated One remedy for this might be for a cooperative to tax all transactions and allocate the revenues among HOSTs in sums proportional to their contribution to the system Cumbersome at best a reward system of this sort would be largely unnecessary 63 See Turner Consumer Protection by Private Joint Action 1967 N Y State Bar Ass'n Antitrust L Symposium 36 4C This position assumes of course that social welfare will best be served by the availability of the widest possible variety of goods and services and necessarily ignores the persuasive argument that as a function of overinvestment some variety increases may be undesirable from the welfare standpoint Cf Markovits Fixed Input Investment Competition and the Variability of Fixed Inputs Investment Their Nature Determinants and Significance 24 Stan L Rev 507 1972 See generally Arrow Economic Welfare and the Allocation of Resources for Invention in Nat 1 Bureau of Economic Research The Rats and Direction of Inventive Activity Economic and Social Factors 609 1962 A-50 if present systems of monopoly subsidy-patent copyright and trade secret are the applicable candidates--could be counted on to give to programs the kind and level of protection that would 63 enable recovery of research and development costs ' Yet as presently framed these three bodies of law offer sparse incentive for investment in software innovation the Supreme Court has recently rendered a decision casting considerable doubt over 64 the patentability of computer programs and though programs are presently accepted for registration by the Copyright Office the level of protection accorded seems hardly worth the registration fee Protection of programs as trade secrets a technique widely employed in the software industry today may be the answer for the future network as well particularly if the network's high degree of security can be relied on to guard against the unauthorized disclosure of proprietary data Yet there too the Supreme Court has raised troublesome questions intimating that the trade secret monopoly may improperly conflict with federal competitive principles and for that reason be invalid And perhaps more important reliance on a trade secret system may since secrecy is its essence undesirably inhibit the exchange of technical information and the development of new techniques from the teachings of preexisting knowledge--one of the signal values of the patent system If trade secret protection endures or f present systems of monopoly subsidy are augmented by a new system for the 63 Even with an adequately functioning system of monopoly subsidy some program of rewards for major achievements or prizes for attaining desired performance criteria might provide a needed auxiliary spur to innovation 64 Gottschalk v Benson 409 U S 63 1972 Copyright Office Circular 31 D May 1964 See generally Cary Copyright Registration and Computer Programs 11 Bull Copyr Soc 362 1964 66 Lear Inc v Adkins 395 U S 653 674-675 676-677 1969 A-51 protection of computer programs 67 it would seem that reliance on the property mechanism created because consistent with the proposed generally competitive structure of the network may be CO superior to a centralized system of rewards The problem of transaction costs--measured in dollars and delay--usually associated with the marketing of patents and copyrights need not obtain in the ARPANET for he system's technology is uniquely situated to administer the bargaining and billing functions with a speed and efficiency not available in other industries See e g Galbi Proposal for New Legislation to Protect Computer Programming 17 Bull Copyr Soc 280 1970 68 Compare Baxter Legal Restrictions on Exploitation of the Patent Monopoly An Economic Analysis 76 Yale L J 267 273-274 1966 69 See Goldstein Information Systems arid the Role of Law Some Prospects 25 Stan L Rev 449 454 n 15 1973 A-52 CONCIJSION The central premise is that properly analyzed the public firm represents the last logical step in the regulatory process and that the determination to divest a public firm commands the kind and at least the level of consideration given the determination to regulate private ones As applied to the ARPANET this premise leads to the specific conclusion that while the risks associated with the network's initial experimental stages justified formation as a public firm the goals established for a fully distributed commercially operative evolved ARPANET seem most likely to be fulfilled through the network's divestiture to a number of independent firms situated in a vigorously competitive relatively unregulated environment This means not only that as a general matter the free market solution should be preferred to the regulatory one--that for example the divested network should be placed beyond the grasp of the FCC--but also that the market to which the network is divested should be insulated from avoidable anticompetitive clogs Care must for this reason be taken in the disposition of present network assets--knowhow particularly--that unjustified competitive advantages not be afforded and entry barriers indiscriminately erected It should be underscored that the reason our economy generally relies on private markets rather than on courts legislatures and regulatory agencies to shape firm decisions respecting price and quality is that the market seems the most fluent mediator between the profit motive and consumer demand Because consumer choice is the economy's touchstone the A-53 question facing the economic planner is in a very real sense not whether decisional power should be lodged in government or in the firm but rather whether it should be lodged in government or the consumer Because government is both a prospective regulator and prospective consumer of the services to be provided by the divested ARPANET it will be doubly postured to influence firm derisions Because as applied to the ARPANET the regulatory process will suffer formative deficiencies not shared by the purchasing process it is the latter that recommends itself as a means for channelling network decisions in the desired directions particularly since unlike exercises of the regulatory power exercises of government purchasing power will naturally diminish in effect precisely at the times when diminished effects will be most appropriate as the network moves in the longer term toward a larger more varied and commercially oriented customer base A-54 APPENDICES ON INSTITUTIONAL ISSUES-PREFACE Appendices B through F This section of the report consists of five separate appendices all relating to institutional alternatives are numbered as Appendices B through F They They were all prepared primarily by Marc U Porat The first of these separate appendices B reviews alternative industrial structures possible describes the present trajectory of development towards one of these possibilities as being most likely unless active reconsideration is taken and it describes what the writer believes to be the most desirable course of action and the reasons for his position In support of his arguments useful background information is presented reviewing some of the most recent changes taking place The second appendix in this series of five Appendix C is a detailed description of the operation of a possible consortium or industry association of packet switching entities including suppliers and users This appendix provides much fine grain detail discussion of the day-to-day procedural operation of an imaginary consortium and provides a flavor of how such an organization might work A fine level of detail is included in this report since the concept of a consortium has not been considered before in this application It was felt that filling in some of these details would help improve the usefulness of discussion about possible organizational arrangements The third appendix in this series Appendix D is a Delphi exercise prepared early in this study In this the staff con- sidered a spectTvim of alternative options narrowed them down to four and then expressed their subjective judgements Considered were differences in the characteristics and expected operetional behavior This appendix here suggests the broad range of alter- native institutional arrangements initially considered and some of the reasons why certain arrangements were narrowed for further investigation ASf Appendix E is a simulation also performed early in the project It examines the expected behavior patterns of competing organizations in the hypothetical situation where such organizations owned different segments of a single network and where strict rules of behavior specified in advance were followed This appendix addresses the question as to whether actions beneficial tc the entire network would result if each separate owner made decisions solely in its own best interest The appendix shows how one might go about programming this behavior to predict performance in advance of a real world situation Lastly in this set of appendices is Appendix F which is the users manual for the simulation model described - II Appendix 3 ti f A 1119111131th ALTERNATIVES i I by MAM U PREFACE The emergence of a packet switching network industry introduces unanswered questions and potential benefits to the producers the consumers and the government This appendix considers one viewpoint of the present trajectory embarked upon by the new industry and suggests reasons for an alternative consideration by the major interested parties These arguments are tentatively planned to be explored in further detail and a Working Paper may issue in 1974 B-i SUMMARY Since packet switching is an immature industry the example of the specialized common carriers SCC and the value added networks VAN will be used as a parallel for discussion The birth of these common carrier-like services was accompanied by aggressive lejal and economic maneuvering and counter-action on the part of both the existing regulated utilities and the wouldbe entrants The past issues and the specific positions taken by the various actors will be discussed because of their present implications Factors to be considered are the regulatory boundary interconnection third-party resale tariff offerings VAN's and Section 214 and credibility Using this background as a basis for prediction the writer concludes that the packet switching business shows a clear tendency towards an eventual oligopolistic industry structure An oligopoly here operating under regulated competition appears to contain several generally undesirable economic fallouts a tendency toward resource sharing inefficiencies price discrimination and imperfect market performance These could result in depressed joint industry profits suppressed industry growth higher prices to the consumers than might otherwise be necessary and continued regulatory disputes Such sub-optimal outcomes are said to result from the nature of the structure of r the industry which prevents the separate entities from cooperating among one another to joint and mutual benefit This appendix considers the alternative of a more coopera- r f tive structure While there are a variety of alternative arrangements possible ranging from laissez faire industry to B-l full horizontal monopoly the realistic spectrum is much more limited ' There is no recent history of laissea faire working in similar cases for the regulated common carriers while horizontal monopoly begs regulation The narrowed spectrum of allowable arrangements is further restricted when antitrust issues are considered What appears to be needed is a cooperative industry arrangement in which government is a member This arrangement could be cooperative such as a farm cooperative or as an industry clearinghouse such as ASCAP or as a consortium such as EDUCOM The exact structure is secondary to purpose For the purposes of discussion it is called the Packet Switching Consortium or consortium for short The three critical functions of this consortium or consortiumlike organization are 1 to facilitate entry 2 to establish universal interconnection between member networks and 3 to serve as a payment clearinghouse to administer shared costs The economics of this consortium suggest as will be shown that a sizable measure of resource sharing and economies of scale is possible By guaranteeing free entry into the consortium and requiring universal interconnection entry and operating costs will be substantially reduced and some antitrust problems will be avoided as well Such an arrangement could also accelerate the propagation of the new services into the least profitable markets and could increase the variety of offerings brought to the marketplace The existence of a consortium-like arrangement could create conditions to aid innovation and R D investments by offering its members a higher degree of market security and a competitive environment in the areas where competition aids the consumer By sharing facilities it is suggested that the market could operate at a higher degree of efficiency to the benefit of both the consumer and the producer Lastly it is held that the FCC could be relieved of many regulatory functions since the consortium is to a major degree self-policing It is always vulnerable to antitrust action should it exceed its legal perogatives B-2 TYPOLOGY OF INDUSTRY STRUCTURES In Figure B-l the spectrum of possible structures of the data communications networks industry is organized into a 2 x 3 matrix The rows separate the structures into regulated versus non-regulated domains illustrative Of course the dividing lines are merely But for purposes of discussion they allow a useful differentiation The columns divide the entire market characteristic spectrum into three distinct groupings These are labeled monopoly oligopolistic competition and pure competition A purist would argue that there is no such thing as a perfect monopoly as there is no pure competition These are merely useful categories which permit us to discuss the almost infinite range of possible alternative arrangements and describe them as locations in a simplified matrix In the matrix the boxes are labeled Type 1 through Type 6 to facilitate discussion Now let us consider which of these boxes represent more reasonable and realistic alternatives Later we shall consider the question of desirability in more detail B-3 TYPE MONOPOLY OLIGOPOLISTIC COMPETITION PURE COMPETITION AT T common carriers The present trajectory Non-existent by definition REGULATED MCI DATRAN PCI TELENET et al 2 1 UNREGULATED Non-existent by definition 4 Figure B-l Television networks 3 Taarest workable approximation consortium 5 Typology of Structures for Computer Communication Industry B-4 6 These six possible outcomes of Figure B-l can be narrowed down to two most realisitc outcomes by the following rationale In the Type 1 case the regulated monopoly the FCC has determined that as specialized common carriers are legitimate enterprises in the public interest it must follow that economies of scale are not sufficient to preclude entry into the field Entry costs for specialized carriers ranging from $50 million to $250 million have been experienced well below the barrier to entry which seems to define a natural monopoly And the entry price for packet switched carriers is even less -- in the $20 million range Each entrant to the new industry has a fair degree of price control suggesting that a unique monopoly price does not exist and that monopoly power can be challenged successfully Not convinced by the FCC viewpoint in this matter AT T has recently argued that the whole specialized carrier industry should revert to a regulated monopoly or at least that there should be a moratorium on new approvals Type 2 the natural regulated oligopoly appears to be a more likely present trajectory It will be argued below that this trajectory will lead to sub-optimal industry growth Type 3 the regulated pure competition is non-existent by definition as is Type 4 the unregulated pure monopolist Type 5 the unregulated oligopoly is a viable alternative but has for historical reasons been supplanted by Type 2 difficulty is achieving it as a stable state of nature The The long-run equilibrium of a perfectly competitive industry is characterized by 1 Price equals marginal cost In a perfectly competitive economy a firm's output decisions do not affect industry prices the absence of monopoly power Hence profits are maximized when price is set equal to marginal cost 2 Supra-normal profits are avoided and consumer surplus is maximized B-5 3 Each firm produces output at the low point of its average cost curve firms which over- or under-produce are eventually eliminated from the market Thus effi- cient resource allocation is assured However the underlying communications media employed by the new packet switching industry cable microwave satellite have been traditionally subject to regulation The introduction of the computer as a coTnmunication switch has confounded the analysis sufficiently that i too is almost by default being included under the regulatoiy umbrella The point will be developed later that an oligopolistic industry whether regulated or not is a less desirable alternative Thus the exclusion by default of Type 5 is not a source of major concern Type 6 the unregulated pure competition is generally held by most economists in the U S as being theoretically the most desirable of outcomes and the case for competition has been advanced in numerous economic tracts In summary the feasible outcomes of industry structures are represented by Type 2 and Type 6 The present trajectory appears to suggest a Type 2 industry while the author believes that the normative position of a Type 6 structure is more desirable if achievable However the Type 6 outcome is unattainable in its pure form owing to factors that will be discussed in detail below A second-best approximation is proposed in the form of a selfregulating and competitive consortium-like arrangement of computer networks See F M Scherer Industrial Market Structure and Economic Performance Chicago Rand-McNally Co 1970 Chap 2 B-6 RELEVANT BACKGROUND Let us now consider the debates now going on in the regulatory agencies and the courts as new specialized carriers are born as they may apply to packet switching Packet switching faces a somewhat different set of problems than do the specialized carriers buu whatever happens to the specialized carriers is to a degree relevant to packet switching whether it be in rates charged for trunks or degree of freedom of action The MCI and DATRAN Specialized Carrier Applications In 1968 Microwave Communications Inc MCI filed an application with the FCC to operate a microwave link between St Louis and Chicago MCI proposed to offer data services to customers at a considerably lower price than was then available through AT T MCI planned to use local telephone lines to distribute connecting with telephone company facilities on MCI customers' premises and using its own microwave relays for inter-city traffic AT T res- ponded vigorously alleging that MCI was proposing to re-sell communication service to a thrid party and that such practice violated AT T's tariff The AT T response sparked an intense legal battle B-7 MCI won its first round with AT T when the FCC granted the company's application in FCC Docket #18921 as a Sectior 214 Specialized Common Carrier not to delay I The FCC further ordered AT T service to MCI and to immediately effect inter- connection of telephone company equipment with the specialized carrier's equipment Shortly following the MCI grant the FCC approved a similar application for the Data Transmission Corporation DATRAN Existing Carrier Response AT T has responded to MCI and DATRAN in two closely linked countermoves a by seeking to delay interconnection and b by offering a competing service These developments are examined to provide a flavor of the difficulties encountered in regulated oligopolistic competition A major change in a long term prohibition against interconnection occurred with the historic Carterfone case wherein telephone companies wore compelled by law to provide interconnecting service to non-telephone-company-provided facilities The argument raised by the telephone company then as now was that serious degradation of service would result with eventually increased costs pajised on to the subscriber by attachment of equipment potentially harmful to the telephcne system MCI claims that it has been ex- periencing difficulties inducing AT T to offer local interconnection with MCI trunks despite the Carterfone ruling and despite the FCC order granting MCI's application under Section 214 specifying that AT T was not to delay service to MCI and to effect interconnection As an accommodation to Bell's fears regarding system degradation by foreign equipments connected to telephone lines the FCC requested that data access arrangements DAA and a certification program be established between Bell and potential entrants This middle ground position is proving to be unpalatable to AT T as its Chairman Mr John de Butts in a recent speech before the National Association of Regulated Utility Companies NARUC stated that we cannot live with the deterioration of network B-l i performance that would be the inevitable consequence of 'certification' and the proliferation of customer-provided terminals that would ensue from it Chairman de ButU also called for a mora- torium on further experiments in economics aimed at increasing competition in the telecommunications industry In a departure from past AT1T policy de Butts called for more regulation from the FCC to halt the growth of regulated competition and to restore AT T its natural monopoly In a thirty-nine page petition to the FCC on October 4 1973 Bell asked the Commission to defer from granting further applications for facilities by the new competing carrier entrants pending outcome of evidentiary hearings Calling dismal the record of regulated competition AT T cited the railroad industry as a prime example AT T's position here appears to be to delay interconnection while trying to show the FCC the folly of its recent liberalization of entry rules and encourage it to return the telecommunication industry to a fully regulated monopoly Type 1 It is late in the game for changes as present estimates for the specialized common carrier industry suggest that it will grow from $1 45 billion in 1972 to an estimated $7 6 billion by 1980 This is too great a vested interest constituency to go away quietly AT T has shifted the battleground away from the national level and into the state level The reason in part is that the losers in the game of non-averaging will be the rural areas who would lose their invisible subsidy Citing Section 2 b of the Communication Act of 1934 AT T argues that Congress clearly intended to leave intra-state communication regulation to the states The focal point of state action on interconnection problems is presently in states' rights conscious North Carolina which had as recently Telecommunications Reports #38 9 24 73 Telecommunications Reports #40 10 9 73 Market study by Frost Sullivan The Specialized Communication Market in Telecommunications Reports #36 9 10 73 Telecommunications Reports #44 11 5 73 B-9 as mid-1973 threatened to flatly deny any and all interconnection The arena is notably much larger with skirmishes also being fought in Alabama Arizona California Colorado Illinois Iowa Maryland Mississippi New Jersey Oregon Pennsylvania Nebraska Tennessee Texas Virginia and Wisconsin Early decisions in North Carolina and Nebraska are temporarily stalled with commissions gathering evidence to legitimize a ban on interconnection MCI has filed an injunction in Philadelphia to force AT T to interconnect and has also initiated a letter-writing campaign to its customers including one hundred firms in Fortune's 500 urging the FCC to take a more active -ole in the proceedings MCI's Chairman McGowan stated that if we're unable to get interconnection we'll go out of business tomorrow Financial Impact The adverse financial impact is being felt throughout the rpecialized carrier industry DATRAN stated that its line of credit has dried up domestically and they have sought and received capital from abroad Even USTS the ITT subsidiary which has access to its major capital generating parent and which is faced with a relatively low cost of- entry of about $25 million is claiming hardship In a recent statement USTS claimed that financing problems of the specialized common carriers were brought about in part by the generally unfavorable attitude of the financial markets toward new ventures This unfavorable atti- tude was further complicated and prolonged by the continuous rumors of the future competitive actions of established carriers Announcements concerning present and future competitive rates services and facilities by the established carriers have seriously stifled the financing plans of tn' emerging carriers and have seriously impaired their financial and therefore their competitive viability Telecommunications Reports #46 11 19 73 Telecommunications Reports #28 7 16 73 B-10 Antitrust Suits Another somewhat related development has been a series of antitrust suits around the country by equipment manufacturers seeking totally unfettered interconnection to Bell A recent case seen by some as another Carterfone is the MACOM Products Corporation vs Bell MACOM produces the Name Caller automatic dialer a small device that attaches to Bell lines and dials frequently used numbers automatically Bell has refused inter- connection citing harm and system degradation This suit is said to form a direct test of the applicability of the Sherman Act and seeks to settle whether or not the Act prohibits telephone companies from restricting interconnection The court decided that the Communication Act does not immunize the defendant from antitrust regulation and has requested th FCC to aid the court in gathering harm data Independent Telephone Companies The independent telephone companies generally have concurred with AT T's position but with some exceptions On the one hand Hugh P Wilbourn Jr President of Allied Telephone Company Little Rock Arkansas aivises that like Winston Churchill's wartime England the telephone industry shoulu fight in every arena -- state commissions courtrooms and the FCC to defend the fully integrated network it has built over the past ozie hundred years against a small band of willful men in governuent and academic circles who are determined to foster competition in the regulated franchised telephone industry The horror stories on service problems are beginning to come in from around the natior On the other hand Paul Henson of United Telecommunications Incorporated feels that interconnection is here to stay and it Telecommunications Reports #14 4 9 73 Telecommunications Reports #24 6 18 73 Telecommunications Reports #37 10 17 73 B-ll is better to bend with the wind than break He adds the world's greatest telecommunications system s threatened more by other forces than it ever will be by a limited amount of regulated competition Congressional Review The antitrust implications raised by AT T's position are not lost in Congressional circles In the summer of 1973 the Senate Committee's Antitrust and Monopoly Subconar ittee under Chairman Philip Hart D Mich chose to investigate communications as the first of seven industries and held hearings from July 30 to August 2 1973 AT T defended its anti-interconnection position by arguing that MCI is cream-skimming the density routes and effectively stealing traffic from Bell since MCI is not faced with costaveraging over all markets -- including the not so profitable rural areas AT T DPS AT T has developed a new technology called Data Under Voice DUV which it is proposing to market under the ATT DDS Digital Data Services offering MCI and DATRAN oppose the new offering seeking to convince the FCC that DDS is anti-competitive and will destroy the specialized common carrier industry ATT DDS will in all likelihood seriously undercut MCI's and DATRAN's prices DAT- RAN has recently filed r motion that DDS development be subjected to a five-year moratorium but the FCC denied the motion The DUV offering is expected to enjoy enormous 'price advantages by using underutilised bandwidth in AT T's long-lines By incurring a small capital expenditure when compared to the cost of building a network from scratch the unused bandwidth cc n be used for data Telecommunications Reports #43 10 29 73 telecommunications Reports #44 11 5 73 B-12 Hence DUV can be charged at close to the marginal cost rather chan the average cost It is to AT T's advantage to demonstrate to the FCC that its capital outlays were small since by the fair rate of return rule its tariff would be concommitantly small -- and undercutting of MCI and DATRAN rates could be achieved while playing the game according to the rule of if there is going to be competition allowed we can compete also The specialized common carriers understand this very well and have charged that AT T is being unfair in its cost allocation At the same time that it is raising monopoly telephone rates elsewhere AT T is making elaborate plans to cut its prices selectively in those relatively insignificant areas where it is beginning to face competition This hi-lo plan pur- ports to be a departure from nationwide averaging but it is actually only a two-tier scheme of averaging low average rates will prevail where competition is threatened and high average rates will be charged where AT T has no competition The spec- ialized carriers'-arguments were to no avail and following a favorable U S Court of Appeals ruling in New York AT T filed the hi-lo tarrif with the FCC to be effective January 14 1974 In partial response to the anticipated 'hi-lo defeat MCI filed its tariff as a three-tier national averaging scheme The issue of pricing is by no means dead since it is unlikely that the specialized carriers have reached a price equilibrium Another view is that a classical price war is being fought but is less noticeable than a gas war under the slightly astigm tized and distorted environment of regulated competition In partial response to price tariff maneuverings the FCC now requires that any carrier who offers new classes or sub-classes of service shall give sixty days notice at the time of filing new or revised tariff schedules Telecommunications Reports #20 5 21 73 # Telecommunications Reports #46 11 19 73 Computer Decisions July 1973 B-13 A case in point that bears major implications to the development of the commercial packet switching industry trajectory derives from Packet Communications Inc application before the FCC PCI applied to the FCC under Section 214 of the Communication Act to gain quick entry into the industry PCI is a member of the vague class of value added networks VAN's which are under under the value-added concept permitted to re-sell communication services leased from AT T to private users provided that more than simple reselling takes place As PCI is a packet switched system it uses computers to route perform error-checking and provide billing Hence it can be argued o fall within the yet not completely defined class of hybrid carriers There is equally strong ground for PCI to force a rulemaking procedure by the FCC and take the position that VAN's are not really carriers and hence not subject to FCC regulation since they are really unregulated competitive enterprises In the writer's opinion it would be to PCI's long-term advantage to be freed from the constrictions of the regulatory environment but in the short run it would have meant the start of an interminable set of proceedings before the FCC Such a battle would be very costly since PCI a new small company might face the combined opposition of AT T Western Union MCI and DATRAN -- both the common carriers and the specialized carriers PCI weighed the long-run costs of pursuing the 214 application against the short-run costs of being barred from entry altogether and chose the former Telenet which also has an application outstanding noted with pleasure the Commission's policy of liberalizing the authorization of other VAN applications The outcome of these battles at the fuzzy portion of the regulated unregulated spectrum has broad implications for a potentially much larger matter of concern to ARPA-IPT's interests in aiding the development of computer resource sharing in this country Computer resource sharing requires communications in combination with remote processing or access to remote data Telecommunications Reports #4 1 29 73 B-14 bases Yet the carrier licensed to use data technology believer in a right unto perpetuity to provide such communications Thus Western Union consistently holds that the information utility services and resource sharing activities are merely telegrap type services that they alone are franchised to provide Among the shared data base systems that Western Union has regarded as violators of their historic charter include Graphic Scanning Corporation facsimile Graphic Sciences Inc facsimile Titan Industries the Hilton Hotels subsidiary NCS Computing Comdata Network offering the trucking industry a money transmission facsimile Transceiver Corporation interstate message permit and money transactions Transport Data Communication message switching and Xerofax Incorporated interstate message permit and money transmission It is difficult to estimate the number cf new information services that have a substantial potential market and are being delayed because the would-be entrepreneurs are awaiting final -judgment on the interconnection issue In the interim the social costs of delaying the implementation of these information services are borne by the public It is also difficult to present the full flavor of the debate going on at the FCC on the issue of packet switching Cabledata Associates has been compiling and organizaing this material and has sent copies to Range Measurements Laboratory with actions up to November 1973 However one new major decision should be noted FCC has approved the PCI Section 214 application that the One mr y argue that the issue is now finally determined -- that this business will now and always fall under the regulatory jurisdiction of the FCC But in this business nothing is fixed it is just made more difficult For example in PCI's approval the FCC said In the event th- t PCI in the future should seek to modify its basic service offering xn a manner that will alter its status as common carrier PCI v 11 be obliged to obtain prior authorization for such change Potentially PCI could ask for de-classification B-15 as a Section 214 Common Carrier and possibly total removal of regulatory power over the firm's operation but it is now clearly within acknowledged FCC jurisdiction It is still always possible for the whole industry to police itself in a manner that protects the public against the abuses that motivated the creation of regulatory law in the first place but such reversal of the industry structure would require a remarkable degree of unity on the part of the new firms And if recent events are valid indicators of future actions the firms are eminently prone to in-fighting and to seizing immediate short-run opportunities so that a cooperative industry action does not seem to be feasible in the short term without sufficient additional motivation Let us now consider the nature of the regulated oligopolistic structure further B-16 ANALYSIS OF DATA CARRIERS The following section presents an economic analysis of an oligopolistic industry the computer communications industry Although it is a regulated industry the following remarks will be seen to be relatively insensitive to whether the industry is regulated or is unregulated The analysis is considered from the firm's point of view assuming initially no regulation Superimposing regulation on the structure would merely possibly reduce viability and profitability for the whole industry and not affect the discussion Again in the following we disucss specialized carriers as well as packet switching carriers because of the greater measured experience There are sufficient parallels to allow drawing of conclusions from the history of one sector to another rts or November 1973 eleven companies are either in or have formally announced their intention of entering the computer communication business This list includes existing carriers specisslizec common carriers and packet switching carriers All ha e issued tentative development plans including technical specifications costs and site installations About five of the firms are operational four more are expected to be fully operational by 1974 and two others plan to go into full service in 1975 In addition a twelfth firm has annou ced that it is signing contracts for domestic satellite data transmission between seven m vtropolitan centers All twelve firms are partially operational today with some in the early shakedown stage Figure B-2 shows an over-lay map of these twelve firms' announced topologies Since nach of the larger firms will have more than one route between the cities shown below Table B-l and Table B-2 show only city-to-city pairs not number of trunk circuits available B-n II Inna-a nun- 5- run alga nu app-35 0 ii no aural i 0 n9 Ilia cl ails-Bangla- ni In It noa-o hills urn - nl inn 1 5 n nil-3aan Alicia- 03 an IlilII-ull l ISD ing 0 95 ml Hotlink-Iii uqost-uws anon main - nun if n manwm n gHm m n58 ngmEHmmHOZ mmaH mm n apical 2333i Gnu Mf y hi rid- Hid- u-uu Air-Ila 33 5 0 in 10-92 lad lain- 2 in 51 no 00 ISInulla wxa TABLE B-l CITIES WITH SERVICE PROVIDED BY ALTERNATIVE CARRIERS ASC ATT CPI DATRAN MCI NCCC PCI SPCC UVI WTCI wu T Seattle Portland Sacramento San Francisco Los Angeles San Diego Salt Lake City Denver Phoenix Minneapolis Milwaukee Omaha Chicago Cincinatti St Louis Kansas City Tulsa Oklahoma City Fort Worth Amarillo Dallas Austin San Antonio El Paso Houston New Orleans Memphis Birmi qham o Miami St Petersburg Jacksonville Richmond Washington D C Philadelphia New York Baltimore Pittsburg Hartford Boston Cleveland South Bend Detroit Grand Rapids Atlanta Hobbs Corpus Christi Louisville Indianapolis Las Vegas Buffalo Rochester Columbus Augusta o o o o o o o o o o o o o o o o o 0 o o o o o o o o o o o o o o o o o o o o o I ftp o o o o m i o o o o o -- o m o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o e o o o o o B-19 3 4 5 4 4 4 8 2 5 4 1 o o o o o o o o o o o o m o o 4 3 2 6 8 o o 0 4 3 1 8 2 4 3 8 3 3 1 1 1 1 3 8 4 7 2 3 1 5 3 1 4 1 5 1 1 1 1 1 1 1 1 1 170 N 52 TABLE B-o2 IND2PENDENT PAIRS T rn nt-fco-Point Portland Seattle San Francisco Los Angeles San Diego Los Angeles Phoenix Los Angeles Omaha L nver Milwaukee Minneapolis Chicago Milwaukee St Louis 'hicago Memphis St Louis Oklahoma City Kansas City Oklahoma City Dallas Houston Dallas Houston San Antonio New Orleans Houston St Petersburg Miami Richmond Atlanta Washington Richmond Philadelphia Washington New York Philadelphia Boston New York Cleveland Pittsburg Chicago Cleveland Number of Number of ln lependent Pairs Redundancies 2 5 3 2 3 2 2 2 3 2 5 6 4 2 2 2 2 6 7 7 2 3 B- 0 1 4 2 1 2 1 1 1 2 1 4 5 3 1 1 i 1 5 6 6 1 2 52 o from the giant AT T to mid-size These twelve firms rangxng Telecommunic 4 a western unxon to tne w j souther Pucifrc and ester onUnental tion all Pi - serve a 1 prese ted i It- states e atr o J _ an intetesting Table B-2 When Table B P autriwtlon for pattern emerges see Figure B-3 o conforming to y 20 K e X v number of cities served where I Zber of competitors xn cxty k constant multxple e error term A short summary of the graph reveals that Number of sites 131 Number of cities 52 Mean sites city 2 6 Mode sites city 1 0 Median sites city 3 0 2 15 Standard deviation 2 1 The present trajectory reveals a clear pattern of the emergence of a clap'-'--al oligopolistic industry The average number of companies ferving a city being 2 6 and the dominant mode at a monopolistic one firm per city Each firm presented its pros- pectus with some if not full regard for its competitors' intentions and whereas the largest metropolitan areas are amply covered by four five six and even seven firc the smaller ones are essentially selected to minir' ze competition A very primitive form of geographical differentiation appears to be occurring The pressure from oligopoly to monopoly are increased by the presently weak venture capital market The smaller firms all face very rocky shores in their attempt to establish a market beachhead in part because the legal questions of regulation B-21 K o H O _ En O 05 W 0 H P OX N ii in en vD H 53 H D s Pi CO I n saiuo JO HaawnN i i i i ST B-22 Jb--L OT g 8 are st 11 unresolved supra The existing carriers are opposed to entry and have raised the cost of entry to all except the rich or the v y stubborn As the venture capital dries up and as the lines of credit are exhausted the smaller companies are becoming less not more competitively viable These are not healthy signs for those that pretcr competition B-23 DUPLICATION BY DATA CARRIERS A closer examination of Figure B-2 reveals the exten of the duplication Twenty-two big city pairs are represented by more than one transmission line Eleven city pairs are covered by two lines four pairs are linked by three lines oie pair is covered by four lines and six other pairs are covered by five six and even seven parallel lines see Table B-3 In all the overall system shows fifty-two redundant city pairs -- transmission channels that add nothing to the overall network except cost See Figure B-4 Of course two parallel identical trunks have twice the bandwidth of a single trunk But most of the cost is in new routes the incremental costs for added bandwidth is small Thus these fifty-two lines represent millions of dollars in potentially wasted resources When faced with such evidence of inefficiency the salient question hinges on the prospects of network interconnection Will the firms realize that the most expensive part of their business that is establishing new communication routes is subject to dramatic cost savings if they would interconnect Is there something innate in the free market system that will foster this sort of efficiency and will the invisible hand tend to push aside the firms that resist the market forces Unfortunately a fool-proof star gazing service is not yet in existence and the writer is forced to draw on past observation f industry behavior if any light is to be shed on the question Students of industrial marks'- structure have repeatedly observed a phenomenon which for lack of a better term became known as the Prisoner's Dilemma Game Scherer pp 142 - 145 B-24 H 1z w 1z Pi H W CO o vo A H W H ro I Nj i m H CM CO H Vss saixiD 50 aaaimN B-25 es The Prisoi' r's Dilemma For the reader who is unacquainted with the original game a description of its dynamics are in order Baker are charged with committing a burglary Mr Able and Mr The District Attorney is unable to prove his case unless he can obtain a signed confession however he is certain that he can obtain conviction on a lesser charge possession of illegal wiretap equipment The two suspects are interrogated in separate rooms and are given a choice of sentences Able is told that if he con- fesses and spills the beans on Baker he Able will get off scot free while Baker will get a ten year sentence If they both confess Able is not kept informed of all developments of course both will serve six years behind bars If neither confesses the District Attorney can gu antee that Able will spend at least one year in jail on the lesser charge same opportunities Naturally Baker is given the These options very neatly form the elements of a minimax game and one which has both a solution and a stable equilibrium To demonstrate this point the prospects of Able and Baker are presented in game matrix format Baker Player Don't Confess MOV3 Confess Baker 1-Yr Able Don't Confess Baker O-Yrs Optimal strategy Able 1-Yr Able 10-Yrs Baker 6-Yrs Baker 10-Yrs Dominant strategy Able Confess Able 10-Yrs Able 6-Yrs Figure B-4 The Prisoner's Dilemma B-26 From Abie's point of view if Baker does not confess then Able can do one of two things confess and go free or not con- fess and with Baker get the one year penalty confess then Able gets one of two outcomes If Baker does if Able also con- fesses they both get six years and if he does net nonfess he gets ten years in prison Since Able is minimizing vhe number of years spent behind bars his dominant strategy in bt -th cases is to confess Baker's choices are symmetrical and he wxll also choose to confess So the dominant stable outcome will net both men six years in prison whereas we know that a non-dominant strateqy -- namely not to confess -- will have yielded both men only one year in jaj1 t Application The Prisoner's Dilemma is a generic form of a game that appears constantly in industries dominated by oligopolies The outcome of the game is that the dominant non-optimal strategy is chosen by players who are trying to second-guess each other in favor of an optimal but non-dorainant position that requires trust in the other player In terms of the main problem at hand consider the following game Firm II Player Move Interconnect Don't Interconnect 14 10 Firm I Interconnect to consortium of networks Optimal strategy 10 Firm I Don't Interconnect to consortium of networks 6 6 9 Dominant non-optimal 14 9 Figure 3-5 Interconnection Strategy Cell entry Profits 7T B-27 From Firm I's point of view he faces two outcomes If Firm II decides to interconnect with a nationwide consortium of networks he can either join the consortium profits 6 or hold out and fight for a market share against the entire consortium profits 14 If Firm II chooses not to interconnect then it would be a high mortality struggle if Firm I also chooses not to interconnect profits 9 or Firm I could choose to enter the relatively calm world of the consortium profits - 6 From Firm I's perspective not interconnecting with the consortium is a dominant strategy as it is also from Firm II's perspective The game then has a solution and an equilibrium the firms will not interconnect hence achieving profits of 9 units apiece A non-dominant but optimal strategy was available -- to join the consortium -- but it was not taken because of distrust of the other fellow's motives A critical assumption built into the Prisoner's Dilemma game is that cell 1 non-dominant interconnect yields higher industry profits than cell 4 dominant no interconnect it deserves further discussion The present trajectory forecasts the development of an oligopolistic industry supra The pric- ing pattern in such industries is characterized by extreme sensitivity by each firm of the others' pricing decisions especially with reference to the pricing leader B-28 The Missing Institutional Arrangement The underlying argument for encouraging interconnection is that it eliminates inefficiencies and permits the industry to grow more rapidly and provide increasing services to the public The bigger the pie the bigger the piece that each entity could theoretically own The question is whether it is possible to make it worthwhile for the individual entities to concentrate moving together as an industry iri their common interest or whether their resources are better spent in duplicating one another's facilities in the larger cities and holding monopolies in the smaller ones This question may already be answered in the overall digital communications industry But the oppor- tunity for the packet switching industry to avoid some of these problems is clear And as it is in the interest of ARPA to provide itself with access to the largest effective digital co jnunications network with the properties of a packet switching network it is in the interest of ARPA to encourage the packet switching industry to cooperate with itself in having the industry evolve in a direction most desirable in ARPA's and the country's interest L B-29 Choice of Name The challenge that we address below is finding a way to encourage the data Communications industry or at least the packet switching component of the industry to work together to achieve the maximum industry economy of scale and to do so in as fully competitive a manner as possible We shall describe a mechanism consider its operation and then consider its results We do not know the ideal name to attach to the proposed new institution It could be called an industry trade association it could be called a non-profit coorporation it could be a farm-cooperative arrangement or it could be any one of a large set of other institutional possibilities Fc the sake of convenience we shall nail the required insitution a consortium But the reader should regard the word more as an adjective than a noun is secondary to its functions three separate functions The exact form of the organization And it must perform exactly free entry universal interconnection and a payment clearinghouse to administer shared costs The discussion of the consortium is divided into two parts the mechanics and the economics of the consortium Below we briefly describe how the consortium works discuss the economics Then we And lastly we consider the detail mechanics in a separate Appendix section Entry Guaranteed free entry into the organization by potential competitors is a fundamental attribute If any artificial barriers to entry can be erected either at the inception of the consortium or at some later point in its development then the initial purposes will have been undermined Any packet switched network wishing interconnection into the consortium should be allowed entry subject only to a minimum set of conditions that protect the established members from harm For example if an applicant is financially unstable and cannot be expected to fulfill its obligation under the payment B-30 clearinghouse arrangement then the application should be legitimately challenged A reasonable compromise in such an event could be the posting of a covering insurance bond universal Interconnection Th ability to interconnect is critical to the formation of a combined network operated cooperatively by independent packet switching networks The technical feasibility of interconnection seems assured but will require the mutual agreement and cooperation of the networks involved to build suitable gateways through which packets may pass from one network to another Strategies which have thus far been explored seem to require either that protocols at the packet switch level be fully compatible or that a basic host to host protocol be mutually agreed upon which can use the various different packet switching systems simply as transmission facilities Since the former require- ment cannot be met at least among current international networks the latter requirement appears to be the only other immediate solution The area still needs research and experi- mentation and effort in this direction will be spurred on by the formation of a consortium for interconnection Payments Clearinghouse In the cases where a user traffic is routed exclusively through one network no inter-networh payment scheme is needed The customer is billed in a straightforward manner according to the firm's normal pricing structure However in cases where a customer's point- o-point requirements cause traffic to flow through a gateway and into a neighboring network or two or more a revenue sharing scheme is needed to reconcile the usage A payment clearinghouse is such a device A method of managing a payment clearinghouse is to keep an inventory of packets flowing from a member network and computing a credit-debit sheet for each member packet count is not needed A packet-by- Telephone companies for example B-31 use statistical sampling for determining separations as it is sometimes called A network which drew heavily on other members would be in a continual deficit position In the structure of independently- owned subnets that we are considering we find economic signals occurring here to cause that network to rearrange its topology with the objective of minimizing payouts characterized by several parameters Every network can be peak and average delay peak average reliability and peak average traffic flows In a companion Appendix E Independent Nodes Economics Simulation Model we have considered the contribution each component of a network makes to the effectiveness of the entire network The computation of such a measure was achieved by eliminating one node from the network in an iterative fashion and examining the impact on the system parameters listed above Nodes that are efficient cannot be removed without a serious degradation of the overall system performance whereas inefficient nodes from a topological point of view can be removed with either no impact on the overall system or a net improvement in the overall system behavior Therefore a payment clearinghouse arrangement could serve a dual purpose- first it could facilitate the formation of a resource sharing consortium by allowing networks to utilize each others' facilities at a fair price and second it would be a constant incentive and source of information for every member of the consortium to experiment with their topology and find the most efficient configuration B-32 Appendix C GN FORMATION OF A COMMON INTEREST CONSORTIUM OF PACKET SWITCHING ENTITIES fay MARC U PORAT PREFACE This appendix is written to encourage a more detailed discussion of the operation of a hypothetical packet switching consortium how it might function how it might go about resolving differences how it would provide the clearinghouse function how it would assure open entry and hew it would create and enforce standards The intent here was to create an existence proof is one way of doing the job And it seems to work Here Therefore there are no reasons to believe that a workable structure cannot be built In the following the language has a certain guardhouse lawyer quality about it This does not mean that this is a carefully prepared legal document Tt is not Rather it is just an attempt to try to list the major contingencies in a formal manner nothing more is intended or should be inferred C-i TABLE OF CONTENTS 1 0 THE CONSORTIUM C-l 2 0 THE BOARD OF DIRECTORS C-5 3 0 OFFICERS OF THE CONSORTIUM C-8 4 0 RULES OF ENTRY INTO THE CONSORTIUM C-10 5 0 MEMBERSHIP IN THE CONSORTIUM C-16 6 0 THE MEMBERS COMMITTEE C-19 7 0 MINIMUM PERFORMANCE STANDARDS C-23 8 0 STANDARDS FOR INTERCONNECTION C-30 9 0 COSTS AND REVENUE SHARING C-32 10 0 TERMINATION OF MEMBERSHIP C-36 C-iii 1 0 1 1 THE CONSORTIUM NAME The name of this organization shall be the Packet Network Consortium or Consortium in this report 1 2 ORGANIZATION The Consortium shall be a non-stock not-for-profit corporation incorporated in the State of Delaware 1 3 PURPOSE The purpose of this Consortium is to encourage the development of the digital packet-switching capabilities in the public interest of the U S facilitating inter-network exchange of data and services and encouraging individually owned and operated packet-switched systems to be interconnected freely 1 4 SCOPE The Consortium shall have the right to enter into contracts invest its funds in short term securities and engage in all other actions normal and appropriate to a not-for-profit corporation It shall not have the right to incur debts beyond its current assets In the event of dissolution all assets will revert either to a not-for-profit organization chosen by the Board or to the Federal Treasury C-l 1 5 LIMITS OF SCOPE While the Consortium is concerned with the general well being of the industry its scope shall be restricted solely to matters affecting two or more separately owned packet-switching networks The Consortium shall have the right to set and enforce internetwork standards and agreements between Members and to protect the rights of the consumer with respect tc services offered by any Member that can negatively affect the well being of 1 6 tne industry MANAGEMENT The Chief Executiv Officer is the Chairman of the Board who is a Member of the Board of Directors and is elected by the Directors See Figure 1 The board is responsible for resolving major disputes and questions affecting membership status and matters affecting the public interest The Board is selected by the Members voting in proportion to their dues paid or contracts issued in the field of packet networks The dues payable by each meicber will be in proportion to the annual gross revenue derived from packet services The Members Committee is a forum for expressing the interests of the Members to the President in which each Member has a single vote The President serves as chairman of the Members Committee and serves as Executive Director for the Consortium He is respon- sible for day-to-day overall management serving as Chief Operating Officer He is appointed by the Board C-2 3 s Q Uou 2 Js A 8 ou g s 5 8 ia 8u 8 6 H I w H N M 8o U c w 04 H S3 ' X S3 g 1 m N 10 oH b to t-f BH 0 O 1 4J tQ R g b H O 8 F- u B 6 H K w as CO H 1 0 C-3 H W Sg Much of the work of the Consortium is carried on in Subcommittees organized by the Members Committee C-4 -Ar 2 0 2 1 BOARD OF DIRECTORS MEETING OF THE BOARD The Board of Directors shall meet annually during the first week of February at the offices of the Consortium or at such place within the United States as the Board may determine or may meet by conference telephone call Additional meetings may be held at any time by majority choice of the Board but all Directors shall be informed at least three weeks in advance by telephone or by SNDMSG mail of the schedule for such meetings 2 2 DUTIES OF THE BOARD The Board of Directors shall have power to change the bylaws of the Corporation to appoint the President Vice President Secretary and Treasurer and to set the salaries and terms and conditions of all officers The Board shall serve as a tinal arbiter for all applications for membership to the Consortium and all other matters not resolved by the Membership Committee 2 3 COMPOSITION OF THE BOARD The Board shall be composed of nine members initially representing the public sector and in no case shall there be more than four who may be from companies supplying packet communications services and no more than two from government service C-5 A 2 4 SELECTION AND TERM j The initial Board of Directors shall be One year appointments 1 to be specified later 2 3 ' i Two year appointments 4 5 Three year appointments 7 1 9 The term of office for other than the initial Board shall be three years No Board member may serve more than two consecutive tenr 3 without being off the Board for at least one year The Chairman of the Board shall be a member of the Board and be chosen by thsi Board at its annual meeting He shall serve as Chairman for a period of one year 2 5 VACANCIES Any vacancy in the Board of Directors o' 'I be filled for the balance of the present term by a majt e of th i remaining Board at the next scheduled Boarc -o -- 3 Board members may resign at any time by written notification to the Chairman C-6 2 6 REMUNERATION The Board shall set tha salary and other considerations for the officers of the Cjnsort ua No salary shall be paid to the directors except in remuneration for direct expenses incurred in service and token honorarium lor attendance at meetings 2 7 FISCAL YEAR The fiscal year shall coi ide with the calendar year C-7 3 0 3 1 OFFICERS OF THE CONSORTIUM SELECTION AND TERM The officers shall be selected by the Board of Directors and serve at the pleasure of the Board 3 2 PRESIDENT The President shall be the Chief Executive Officer of the Consortium and shall normally serve as chairman at the meetings of the Membership Committee He shall conduct all other duties as defined in the By-Laws 3 3 VICE PRESIDENT The Vice President shall serve as Acting President in the absence cf the President and shall be responsible for all other duties assigned by the Board or the President He shall also serve as Chairman of the Management Services Subcommittee 3 4 TREASURER The election of the Treasurer shall take place on the first meeting of the Fiscal Year The Treasurer shall serve as chair- man of the Finance Subcommittee and shall be responsible for performance of that Subcommittee C-8 3 5 SECRETARY The Secretary shall affix the seal of the Corporation on all documents and contracts as required by law and shall serve in any other capacity as defined by the Board or the President 3 6 RESIGNATION Officers iaay resign only upon due written notice to the Chairman of the Board in conformity with the terms of their employment agreements t C-9 L i 4 0 4 1 RULES OF ENTRY INTO THE CONSORTIUM GRADES OF MEMBERSHIP There are several grades of membership in the Consortium Full Membership is normally restricted to owners of packetswitched networks that interconnect with others Provisional Membership is the grade of membership for representation from organizations that meet essentially all of the requirements for Full Membership but are deficient in some specific major regard Normally the Provisional Membership status lasts 90 days in which it is anticipated that the deficiency will be corrected and Full Membership status restored A 30 Day Provisional Membership state is used for the case of an organization expected to terminate its connection to the Combined Network Associate Membership is used for observers and other individuals not normally affiliated with an organization owning a connected network Subnet Member includes representation of organizations that in and of themselves do not constitute a full network requiring enroute packet-switching for others Examples would include an organization owning only a Host computer or an organization owning only a TIP sprving a single geographical area Table 1 describes the rights and responsibilities of the C-10 TABLE 1 j RIGHTS AND RESPONSIBILITIES OF VARIOUS GRADES OF MEMBERSHIP u S u 0 c 9 Associate Member Observer u 0 0 ss 0 0 0 1 c 0 0 0 m 1 V 1 z 0 s H id c 0 oH i 0 -H f-H tu 0 c en 0 Serve on Subcommittees to U 1 10 0 1 10 1 Right to Attend Meeting 10 1 0 Full Voting Right V in 4 1 z 4 1 O ss w V 1 V z U s1 Q c H u X Non-obligatory right to interconnect Right to interconnect - general H to Access to Management Services A 0 Participate in revenue-sharing agreement 1 z V z 0 c 1 ss ss 1 1 0 1 VI 0 1 Meet all conditions for Full Member not including performance standards z 0 VI 01 51 Meet full performance standards 0 0 ss V O tn w 0 Z c 2 1 0 0 11 Contract with Hosts for services 1 2 10 0 1 Right to change add or delete components that affect the Combined Network Own full independent network w H 1 a 1 w 0 c z 0 ss z 0 ss 0 1 Own partial network that cannot work without interconnection Meet reliability and full performance standards EH H a H ffl H W 8 1 3 in 0 a 0 ss 0 ss 0 ss VI Pay dues of 1% of gross revenue 1 o 0 ss Pay $200 year o V 0 i Attend 75% of meetings of Members Committee 1 ss 0 ss ss C-ll s various grades of membership These rights and responsibilities are described in more detail in the following sections 4 2 APPLICATION FOR MEMBERSHIP Each prospective Full Member of the Consortium must formally apply for membership and fulfill the following requirements to gain entry See Figures 2A and 2B for hypothetical application forms which suggest the form of the information sought 4 3 CREDENTIALS Demonstrc te financial technical and managerial capability Each prospective Full Member shall show through a documented presentation that it is capable of providing better than the minimum allowable quality of service Evidence should include an audited recent financial statement presented to the Financial Subcommittee a detailed statement of prior or present activity in a related technical field or acquisition of a technical staff to assure technical competence presented to the Performance Standards Subcommittee presentation of an organization chart delineating major areas of responsibility with respect to Consortium membership and obligations presented to the Membership Subcommittee 4 4 CONTRACTUAL CONSENT All members are bound to perform pursuant to the rules and regulations of the Consortium except in such cases as the Board elects to grant special exemptions or privileges to a member upon the discovery and proof of extraordinary circumstances Failure to abide b Consortium rules and regulations shall cause Full Member status to be reduced to 90 Day Provisional Member- C-12 mm Page i of 2 pages PACKET NETWORK CONSORTIUM APPLICATION FOR FULL MEMBERSHIP 1 Full operating name 2 H Q Address 3 Name Address of all affiliates branch offices etc 4 Parent organization Tel # Name Address Telephone 5 Type of organization 6 Statement of purpose 7 President 8 v P Operations 9 V P Marketing 10 Technical manager 11 Appendices ' Corporate Charter By-laws Most Recent Financial Statement and Annual Report r Interconnection Standards Facilities Certification a FIGURE 2A APPLICATION FOR FULL MEMBERSHIP page 1 C-13 Page 2 of 2 pages As duly authorized representatives of we petition the Packet Network Consorium for J Full J_J Provisional Membership and in recognition of the rights Associate I Sub-net and benefits derived therein agree to fully abide by the rules and regulations of the Packet Network Consortium as defined in their Charter and By-laws as effective on this data as may be modified We certify that the gross annual revenue of our organization related to packet communications was $ and taxes before expenses Enclosed herein is the greater of one percent of that amount or $300 Amount attached $ President Date Secretary Date FIGURE 2B APPLICATION FOR FULL MEMBERSHIP page 2 C-14 oMM ship or disqualified from membership 4 5 OBJECTIONS TO ENTRY Upon submission of all necessary credentials by a prospective Consortium Member any Full Member may raise an objection grievance or request to bar entry on the following show cause grounds the prospective member has not demonstrated good faith in previous dealings with a member of the Consortium and or the prospective member has failed to demonstrate financial technical or managerial capability and or the prospective member declines to abide by the rules and regulations of the Consortium and or the prospective member's location in the Combined Network threatens to seriously degrade the overall network performance 4 6 MOTION TO BAR ENTRY A prospective application for Full Membership may be barred or reduced to 90 Day Provisional Membership by a two-thirds 2 3 vote of the Members Committee A prospective applicant who is thus barred from entry into the Consortium as either a Full or 90 Day Provisional Member may either re-file the application after a ninety-day 90 period without prejudice or appeal the decision to the Board of Directors to be heard at their next scheduled meeting The decision of the Board will be binding and if entry is barred no further application from the applicant will be accepted for one year However the prospective member is always free to take court action if he feels that the decision is unfair In the event that a Motion to Bar Entry is raised by a Consortium member the Members Committee may by a majority vote elect to stay the application for a thiity-day 30 period During this period the substantive issues raised by the Motion to Bar Entry will be studied by the Membership Subcommittee and a recommendation of action provided to the Members Committe for their review and vote C-15 5 0 5 1 MEMBERSHIP IN THE CONSORTIUM FULL MEMBER RIGHTS AND DUTIES Upon entry into the Consortium as a Full Member status the Member is granted the following rigrus full voting membership in the Members Committee meetings - a Full Member may hold an office in the Packet Network Consortium right to join subcommittees of the Coordinating Committee right to subscribe to all management services supported by the Consortium and to join in all activities experiments and conferences sponsored by the Consortium right to interconnect to any or all nodes in Full Members' network subject to technical standardization considerations right and obligation to partake in revenue separation agreements right-to enter into contract with any Host for services right to add or delete IMPS or TIPS or communications links The duties and obligations of Full Members shall include attendance at not less than seventy-five percent 75% of al Members Committee meetings during the twelve-month period following entry into the Consortium and each subsequent twelve-month period agreement to support the Consortium by prompt payment of membership dues stallments Dues shall be payable in pro-rata monthly in- Being in arrears in excess of sixty 60 days shall cause the Full Member status to be downgraded to Provisional Membership C-16 5 2 PROVISIONAL MEMBER RIGHTS AND DUTIES The Provisional Member status is granted a limited set of rights consisting of right to attend all meetings of the Members Committee open to Full Members right to join subcommittees of the Coordinating Committee right to subscribe to management services supported by the Consortium and to join in activities experiments and conferences sponsored by the Consortium right to interconnect to any node in the Consortium subject to a nonobligatory agreement by a Consortium member to interconnect A Provisional Member is bound by the same duties and obligations outlined with respect to attendance and payment of dues Provisional Member status is granted to a member by a majority vote of the Members Committee in the event of the fcxlowing a Full Member requests such status and or a Full Member fails to meet the minimum performance standards and or Tails to meet the fee payment schedule fails to abide by the Interconnection standards defaults on a cost-sharing contract with another Full Member defaults on the inter-network revenue sharing arrangement The Provisional Member status expires after a ninety-day 90 period at which time the Members Committee reviews the conditions of the Provisional Member Should the conditions listed in the aforementioned sections not be rectified the Members Committee may elect by a simple majority to- terminate membership in the Consortium or re-issue a 90 Day Provisional Member status Such a re-issuance may be obtained a maximum of three times at which point mandatory termination of membership occurs C-17 k 5 3 ASSOCIATE MEMBER Any member of the Members Committee may nominate an individual affiliated with government university media industry or any other concerned group or individual interested in being an observer to receive Associate Member status into the Members Committee meeting of the Packet Network Consortium A motion to seat such a member shall be carried by a simple majority vote 5 3 1 Rights and Duties An Associate Member may attend all meetings of the Packet Network Consortium open to the Full Member participate in all events and functions of the Consortium and join a Subcommittee of the Consortium Such a member may not vote in the Members Committee except to cast an Opinion vote The membership fee for Associate members is $20C a year non-refundable C-18 6 0 6 1 THE MEMBERS COMMITTEE REGULAR MEETINGS 6 1 1 Date Regular meetings of the Members Committee shall be held on the first Tuesday of each month Except where the first Tuesday constitutes a National state or religious holiday then the regular meeting shall be re-scheduled to the next available day and a notice of such change shall be made by the Chairman of the Members Committee the President 6 1 2 Participants All Tull Members Provisional Members Associate Members and Subnet Members hold seats in the Members Committee Only Full Members may cast Action votes in the Committee but all members may cast Opinion votes 6 1 3 Minutes Minutes of the Members Committee meet-' ngs and of all subcommittees of the Members Committee shall be supplied at reproduction plus handling cost to all grades of membership As a matter of policy all business conducted in the name of the Consortium other than that relating to personal data shall not be held from any other member of the Consortium or from the public Violation of this article shall jeopardize the Full Member status C-19 and constitute grounds for termination of membership 6 1 4 Procedure All meetings of the Members Commitcee and Subcommittees shall use the parlimentary procedures under Robert's Rules of Order 6 2 DUTIES OF THE SUBCOMMITTEES The Members Committee shall form such subcommittees as defined in the various sections and articles of the by-laws or as the need arises or as requested by the Board of Directors 6 2 1 Member ship Subcommittee To oversee the entry procedures and to issue a recommendation on each case 6 2 2 Performance Standards Subcommittee To oversee the reliability of the Combined Network and to certify eacy Member's Down Time report and to offer advice to the Members Committee on any technical issues which may arise from technical changes in the member networks 6 2 3 Interconnection Standards Subcommittee To oversee the protocols for interconnection and to offer advice to the Members Committee on issuer arising from interconnection standards 6 2 4 Finance Subcommittee To certify each tntry applicant's financial statement to C-20 i assist Consortium Members in any cost and revenue sharing to recommend changes in the membership fees to collect dues to oversee the financial operations of the Consortium to review the annual balance sheet 6 2 5 Arbitration Subcommittee To arbitrate disputes between Members when both parties agree to arbitration A Member will be disqualified from service by the Arbitration Subcommittee if any doubt should arise regarding conflict of interest or prejudice with respect to the disputing parties or the issue at stake 6 2 6 Management Services To oversee or negotiate the operations cf all management services offered by the Consortium to its members and to act as liaison between the Members Committee and the management services operations 6 3 MANAGEMENT SERVICES The Members' Committee shall provide certain management services open to all members of the Consortium as defined below or as subsequently defined by the Members Committee The Members Committee shall be empowered to either staff such services or to contract such services out pursuant to a recommendation by the Management Services Subcommittee 6 3 1 Network Information Center To post rind publish all minutes motions reports changes and other information relating to tht activities of the Consortium as defined in the by-laws or as subsequently defined by C-21 i i-- -- 3fa the Members Committee To maintain preferably in machine- retrievable form any documents submitted by the Committee relating to the operations of the Consortium or services offered by its Members 6 3 2 Network Monitoring and Measurement Center To monitor the operation of the Combined Network and detect malfunction in any component to alert the members of the network in the event of a malfunction or disruption in service to aid in repairing such a malfunction to keep logs on traffic flows in the Combined Network including performance data on reliability and delay time to conduct experiments as requested by the Members Committee and to make available all such data as a public record upon request 6 4 QUORUM A legal quorum of the Members Committee shall be defined as a majority of all Full Members 6 5 SPECIAL MEETINGS A special meeting of the Members Committee may be called by the Chairman of the Board the Presdent or by at least one-third 1 3 of the members of the Members Committee All Members except Associate Members shall be notified at least ten 10 days prior to the actual date of a Special Meeting C-22 7 0 MINIMUM PERFORMANCE STAND RDS All members except Associate Members of the Consortium shall maintain performance records as described below The Consortium as a whole shall support a Management Service to evaluate the technical performance of each node Maintaining a Full Member status in the Packet Network Consortium is contingent upon maintaining at least a minimum quality of service throughcut all parts of the connected network that can affect the performance of the Combined Network defined below Regular reviews of each node's performance will be provided at each Members Committee meeting by the Consortium Management Service for network analysis The target minimum quality of service is that the Combined Network composed of all member networks shall be operational twenty-four hours per day every day of the year The target reliability standard for the Combined Network is to be defined Each member shall guarantee that its nodes fulfill all agreed to standards 7 1 MAINTENANCE Any individual Gateway IMP or equivalent or TIP or equivalent may be taken out of service as required by routine or emergency maintenance provided that the fractional amount of time involved in such maintenance shall not exceed five percent 5% of all time computed as follows To he defined 7 1 1 Routine Routine maintenance and program changes which interrupt C-23 Service shall take place only during published Scheduled Maintenance Period the Performance St irds Subcommittee shall collect and publish such information during the regular meeting The Scheduled Maintenance Period shall extend between 0100 and 0600 Eastern Standard Time 7 1 2 Emergency Any failure which interrupts on-going computation from any TIP or IMP occurring outside a published Scheduled Ma ntenance Period as defined in 7 1 1 shall be counted as Emergency Down Time Emergency Down Time shall be measured from the first detection of failure until the failing unit is restored to full service The duration of Emergency Down Time shall be multiplied by ten 10 when computing the time involved in maintenance discussed in 7 1 See Table 2 7 1 3 Transient Any transient failure which interrupts on-going computation from any TIP or IMP for three 3 minutes or less shall be counted as a thirty 30 minute Down Time failure Any transient failure longer than three 3 minutes shall be considered as Emergency Down Time as described in 7 1 2 7 1 4 Modem Failure of a single TIP input modem from the user shall count as a failure of one-tenth 1 10 its time duration in computing TIP statistics 7 2 7 2 1 PERFORMANCE REPORT Failure Statistics Failure Statistics for each TIP and each IMP shall be prepared each month and certified as being correct C-24 Jk Signed copies TABLE 2 PACKET NETWORK CONSORTIUM UP-TIME AND MAINTENANCE-TIME OUTAGE EQUIVALENTS Outage During Equals Up-Time Outage During Scheduled Maintenance Period 1 Minute 10 Minutes 6 Minutes 1 Hour 1 Hour 10 Hours C-25 4te4 4ib i m mm of the performance reliability report shall be sent to each Host installation attached to any part of the Combined Network and to the Members Committee The Member Committee shall publish such reports and make them available to all Consortium members 7 2 2 Customer Complaints All customer complaints received from any user concerning any component of the Combined Network its agents or representatives shall be compiled by the Performance Standards Subcommittee and entered as part of the operation record Full text of such complaint records shall be made available and shall be maintained by the Performance Standards Subcommittee and made available to all Consortium members 7 3 EXCESSIVE OUTACE Excessive outage beyond the limits set forth in Table 3 may result in the downgrading of a Full Member to a Provisional Member status 7 4 HOST CONNECTION Any new Host wishing a new connection with a Full Member of he Consortium and have its service available to all users on the network shall be provided with access to an IMP within s ixty days 60 from the time of issuance of a formal contract between the Consortium member and the intended Host Upon the receipt of an Intent to Connect statement the Members Committee shall post and publish a description of the Host See Figure 3 Upon publication of the Intent to Connect any objections modifications or grievances arising from such an intent may be brought to the Members Committee or to an Emergency Forum by any C-26 J B a MM MMMiHMlHi PACKET NETWORK CONSORTIUM APPLICATION FOR ADDING NEW HOST COMPUTER FACILITIES TO THE NETWORK a Full Operating Name b H Q Address c Name Address of parent organization d Name Business address oft Tel i President ii Vice President iii Technical Manager iv Liaison with Computer Network Consortium Member e Log-on procedure at HOST site f Full list and description of services offered by HOST Attach g Price list of each service listed above Attach h Name of Consortium member s serving as a connection to the Combined Network FIGURE 3 HOST APPLICATION C-27 i TABLE 3 PACKET NETWORK CONSORTIUM ALLOWABLE OUTAGE TIMES FOR ANY NETWORK FORMING PART OF THE COMBINED NETWORK Hours 100% Up-Time 24 hours x 365 25 days 8766 00 98% of Maximum Up-Time 8590 00 Maximum allowable outage 175 32 Allowable Down-Time 1 hour per 365 25 TOTAL ANNUAL ALLOWABLE OUTAGE 540 57 Mean Daily Allowable Outage 1 48 C-28 am Consortium member within fifteen 15 days after the publication date The Members Committee or Emergency Forum may by majority vote elect to stay an application for a thirty-day 30 period on the following show cause grounds a The intended Host has demonstrated bad faith in previous dealings with a member of the Consortium b The intended Host has demonstrated technical incompetence in previous dealings with a member of the Consortium c The intended Host's connection will degrade the technical performance of the Combined Network Upon the termination or the thirty-day period starting on the day of receipt of the Intent to Connect statement the intended Host shall be connected Failure to so comply barring extraordinary technical failures may jeopardise Full Member status in the Consortium C-29 m 8 0 STANDARDS FOR INTERCONNECTION The Packet Network Consortium shall create and support a subcommittee for Interconnection Standards The functions of the subcommittee are a To insure interconnection standards between all members of the Consortium via at least one Gateway each b To agree upon standards for inter-network protocols specifically at the Host-Host and basic process-process level c To join in protocol experiments leading to more efficient use of the Combined Network d To certify all prospective Full Members' facilities with respect to system compatability such that overall network performance will not suffer degradation e l HOST-HOST PROTOCOL To be specified 8 2 PROCESS-PROCESS PROTOCOL To be specified C-30 4 4 8 3 CERTIFICATION OF FACILITIES To be specified 9 0 COSTS AND Y NUE SHARING All Members except Associate Members of the Consortium are eligible for cost sharing arrangements Such arrangements are completely optional and are negotiated on an individaul case-bycase basis by the Consortium members involved 9 1 LEASED LINE COSTS Exclusively intra-network leased lines are the responsibility of the individual Consortium member However whenever a leased line connects two or more gateways the parties in question may split the cost of that line The cost-splitting formvla may be negotiated in any way that is deemed satisfactory to the Consortium members in question e g pro-rata traffic flows prorata projected traffic flows etc 9 2 JOINT VENTURES Any member of the Consortium is free to engage in a joint venture with other members of the Consortium The joint venture may include hardware acquisition e g satellite link or service and marketing e g acquisition of Host or TIP facilities 9 3 MANAGEMENT SERVICES All Consortjum members may utilize the available management services the cost of such services approximately in proportion to membership fees paid C-32 9 4 INTER-NETWORK TRAFFIC I Revenue-sharing arrangements occur in the case of internetwork traffic flows i e use of facilities belonging to one Consortium member by another Consortium member Each member's gateway node performs an accounting function to keep track of packet-origin At the clos of the monthly billing period a i statement is issued by each member See Figure 4 The monthly statements are processed by the Consortium and tiny accounting errors or anomalies are resolved The Consortium processes th statements and issues a bill to each member See Figure 5 failure to reconcile all debts to other Packet Net- work Consortium members is defined as forty-five 45 day arrears and may result in downgrading membership to Provisional status C-33 Jl l PACKET NETWORK CONSORTIUM INTER-NETWORK ACCOUNTING FORM BILLING PERIOD MEMBER NUMBER PACKETS TO MEMBER NO PACKETS FROM MEMBER NUMBER 1 2 3 n 1 2 3 o o - FIGURE 4 INTERNETWORK ACCOUNTING FORM C-34 PACKET NETWORK CONSORTIUM MONTHLY RECONCILIATION STATEMENT BILLING PERIOD MEMBER NUMBER Adjustment To From Credit debit Member No 1 2 3 o n FIGURE 5 MONTHLY RECONCILIATION STATEMENT C-35 Al i a--Sfc BP 10 0 10 1 TERMINATION OF MEMBERSHIP ESTABLISHING A NEED FOR TERMINATION Need for termination is established by several methods Each has a time limit associated with it to facilitate any transitional difficulties that may result from termination 10 1 1 Member-Requested Any member may upon receipt by the Members Committee terminate Consortium membership after a thirty-day 30 period The Members Committee upon receipt of such an Intent to Terminate notice shall post and publish the details of the termination 10 1 2 Network Consortium-Initiated Termination Any Consortium member may initiate termination proceedings against another member if a That member has held a Provisional Member status for a period not less than ninety days 90 b That member has been convicted in a court of law in a proceeding involving the operation of the Packet Network Consortium or adversely affecting the operation of a Consortium member Termination under the above circumstances shall be effected by a simple majority vote of the Members Committee C-36 10 2 30 DAY PROVISIONAL STATUS A 30 Day Provisional Member status is assigned to a Consortium member and after thirty days 30 all connection with the Provisional Member is severed 1C 3 TRANSFER OF OWNERSHIP Membership in the Consortium is not transferable by sale lease or grant In the event of a transfer of ownership the new owner shall petition the Consortium for entry in the normal manner C-37 Appendix D A DELPHI EXERCISE EXAMINING FOUR ALTERNATIVE COMMUNICATIONS POLICY OPTIONS by MARC U PORAT Hi PREFACE I This appendix is a work product generated during a period of discussions about completely different ways that ARPA might go about solving its problems of pressures for growth of the ARPANET This particular piece was a Delphi interrogation of the staff during its discussions It is included partly for histor- ic reasons and partly because it shows that alternative industry structures were considered before acquiring a fixation on the single final suggestion that is examined in detail in this larger report Since this Delphi discussion took place almost six months ago much discussion has taken place and what is called the HCCN is now generally referred to as the combined network D-i i INTRODUCTION During two weeks in August 1973 the Cabledata Associates' research staff engaged in an informal Delphi on the question of ARPANET'S future The staff explored several possible directions for divestiture each of which might lead to a markedly different future for the computer-communications industry as a whole Following a preliminary discussion to set the scope of the study the group participated in the first iteration of the Delphi The original five scenarios were collapsed and redefined into four scenarios as presented in the report The group acknowledges that the description accompanying ne scenarios is necessarily vague by virtue of the broad scope of the context However the aim of the Delphi was more to interchange our own thinking on the subject in a systematic manner rather than any attempt to forecast a future It was successful in stimulating debate and compelling the group to develop a common language METHOD The four scenarios were discussed by the group until agreement was reached on the definitions of the evaluative terms The writer attempted to capture this image in the form of a brief descriptive summary Each of the scenarios was then evaluated by th3 participants with reference to fifteen criteria vtpms An p xplication of the evaluation criteria was included in the Delphi package and is shown in the next section of this paper followed by the definition D-l of the rating metric The group then discussed the outcomes with the aim of reconciling or at least airing the major differences DEFINITION OF CRITERIA ITEMS The computer-communications industry is assumed to develop under the stated scenario such that by 1978 the criteria item will be rated supply rating 1 5 Reliability Service will not be unduly interrupted a customer can expect error-free delay-free service on call High capital investment good technology good management responsive to changing demands Equality Neither the purveyors nor the customers suffer discrimination or both purveyors and customers are equally on the short end of the stick Innovativeness New hardware software organizational ideas are diffused and implemented quickly Change and experimentation are encouraged not hindered Stability When change occurs it does not result in disruption of the operating system long-range planning is possible from the customer point of view Next month will seem like this month D-2 - tk Rate of growth Of traffic information services etc compared to the other scenarios Ability to raise capital The industry looks profitable venture capital is attain- able and stock offerings will be well-recieved Speed of raising capital The P E ratio starts high stays high Implementability Organizational the management structure is sufficiently familiar or disguisable that an organization can be efficiently formed and launched Implementability Technical the technical issues can be resolved -- not waiting for state-of-the-art breaktiirough Inipl emen tab i 1 i ty Legal the scheme won't encourage intra-industry lawsuits anti-trust actions FCC heel-dragging Congressional eyebrow raising Implementability Political toe-stepping can be kept to a minimum special interests won't be fatally offended powerful enemies will not be D-3 TM i incurred The political process will encourage the scenario I Impleroentability Social the scenario will receive no press or good press consumer groups will be in favor of it the public reaction will be favorable Good PR Social welfare The scenario will pull society in a desirable direction -- good national policy high utility in the long run Privacy System hard to tamper with reasonable standards of privacy can be assured Probability of occurrence Given everything what are the real chances that the scenario will oc ar RATING METRIC 1 Terrible worst of all possible outcomes Almost nothing good to say about it 2 A real problem worse than today's conditions Major over- haul required in this area 3 About the same as today acceptable with complaints Lots of room for improvement but functioning 4 Quite good no major flaws or complaints Definitely not a D-4 --_ ____ __ _ _ a problem area Attention and resources could be turned to other matters 5 Excellent best of all possible outcomes to say about it D-5 -A- Almost nothing bad FOUR SCENARIOS SCENARIO I BUSINESS AS USUAL The game is played by the old rules in the usual manner Computer communications networks are classified so as to fall partially under government common-carrier regulation The in- dustry splinters along regulatory edict lines suc h as a Virtually unregulated hybrid data processing with message switching incidental to data processing Example Tymnet b FCC regulated hybrid communications with data processing incidental to message switching Example PCI The industry structure aligns to primarily fit the regulatory constraints rather than the market-place All technical issues and tariffs in the regulated case are argued individually The OTP NBS GSA become heavily involved in setting the government policies with respect to the new industry especially in government purchasing SCENARJO II POST OFFICE COMSAT A Public Information Utility Corporation is set up by Congress after Executive Department request It is funded partially by the government by common carriers and by stock offerings Any user may have access to the net if he pays the tariff may be a private individual a private company a branch of D-6 A user government an information utility service or one of the hybrid computer communications companies have access via the PIUC network Private nets attach t and The PIUC raises its own capital and invests it as it sees fit e g creating satellite links leasing ATT lines buying IMPS etc Competition between firms using the PIUCNET is possible and is encouraged Legislation is enacted to ensure privacy SCENARIO III NATIONAL COMPUTER COMMUNICATIONS NETWORK NCCN The National Computer Communications Network NCCN is established as a non-profit association It is composed of com- peting entities each of whom offers an information service These components of NCCN take on disparate forms such as a HOST centers with powerful general computing facilities b Small HOST centers offering a specialized information utility c TIP operate s who perform the local marketing and local customer hand-holding service functions d franchi se sub-nets which lease IMPS TIPS and or HOSTS to individual entrepreneurs e Support companies which help the components in technical managerial marketing problems All members in NCCN agree to abide by a charter and have an active role in changing the charter When a new membership is approved the member agrees to abide by NCCN rules serves in three functions 1 An entry exit mechanism for its components and an information center 2 Guarantee interface standards between com- ponents or sub-nets D-7 The NCCN 3 Provide a revenue-sharing function according to the contribution of each -omponent to the whole network Line-leasing marketing management accounting etc are generally left to the member components The NCCN operates on a minimum budget raised by levying a fee on each of its members SCENARIO IV ADAM SMITH PLUS BIG STICK The government's regulating of the industry is minimal primarily via enforcing anti-trust statutes A free market develops and flourishes with small and large companies coexisting and flourishing The government centralizes its purchasing under one roof and becomes a significant customer on the market-place It wields considerable leverage in forcing interconnections between networks using such powers as boycott subsidy and anti-trust threats Monopolistic contours do not take form in the industry leaders free entry to the industry remains possible cross-subsidization and price-fixing do not occur 0-8 TABLE 1 THE FOUR SCENARIOS Group Measures N 5 Evaluation Criteria I II Business s Usual Post OfficeCOMSAT III NCCN IV Adam Smith Big Stick 1 Reliability 3 2 3 4 3 4 3 2 Equality 2 8 3 6 4 2 3 3 Innovativeness 1 8 2 2 4 2 3 8 4 Stability 2 4 3 3 2 6 5 Rate of growth 2 4 2 8 4 3 2 6 Ability to raise capital 2 8 3 2 3 2 3 2 7 Speed of raising capital 2 4 2 8 3 2 8 Impl ementabi 1 i ty 8 organizational 3 8 2 4 2 4 3 2 9 -technical 3 4 3 3 4 3 10 legal 3 4 1 6 2 4 2 6 3 1 2 2 6 2 8 3 2 2 3 6 2 6 2 2 8 4 2 2 8 2 6 3 4 4 2 6 2 8 4 3 22 24 11 political 12 social 13 Social welfare 14 Privacy 15 Desirability overall 2 1'ROBABILITY OF OCCURENCE 44 ZN 15 p 2N 15 1 2 746 2 688 3 600 2 946 1 208 268 792 707 D-9 GROUP JUDGMENT While the rankings by the group are perhaps overly more indicative o r the internal value systems of the individual respondents the do provide some insights into the existence of consensus IN SUMMARY The Most Desirable Outcome The Most Likely Outcome NCcN x 3 6 Business as Usual p 44 The Least Desirable Outcome The Least Likely Outcome Post Office COMSAT x 2 7 Post Office COMSAT p 1 The group's preference for Scenario III NCCN was clearly xpressec The NCCN Scenario received highest ranking or tied for highest ranking in twelve out of fifteen criteria items The following chart outlines its areas of strength and weakness Scenario III National Computer Communications Network Strong Areas Equality Innovativeness Rate of growth Social implementability Social welfare Privacy Overall desirability Weak Areas Organizational Implementability Legal Implementability Political Implementability The NCCN received a probability of occurrence of p 22 exactly half that given to the Most Likely Scenario Business as Usual Thus the group felt that the most desirable course is not the most likely present trajectory D-10 Appendix INDEPENDENT NODES ECONOMICS SIMULATION MODEL MARC U PORAT TABLE OF CONTENTS INTRODUCTION E-l IMPLICIT ASSUMPTIONS E-2 EXPLICIT ASSUMPTIONS E-3 IMPLICATIONS E-6 FIGURE 1 f lulation and Gaming Model Appendix A - MODEL LISTING E-8 E-9 Ar 2ndix B - INIC DICTIONARY E-18 Appendix C - SAMPLE INEC RUN E-19 E-l INTRODUCTION The general purpose of the Independent Nodes Economics INEC model is to lend insight into the ARPANET divestiture issue and into subsequent network behavior patterns sented is in its 1973 August 21 stage The INEC model as pre- Real data have not yet been loaded into the model e g snapshots of the ARPANET topology and statistics The model was built on a set of implicit ai d explicit assumptions These assumptions are stated here to allow substitution of alternate assumptions which would be more closely representative of the nature of ownership being examined INEC is a highly gen- eralized model suitable for various interpretations In it we assume that each node or a collection of nodes a sub-net can be owned by competing entities Therefore each gamer on the INEC model represents a sub-net owner We assume that all sub-nets share a protocol and are interconnected E-l ASSUMPTIONS IMPLICIT ASSUMPTIONS Assumption 1 The future of the ARPANET is not yet frozen and we are still in a policy R D stage decisionmakers involved in the divestiture question will be operating with some mental model of the network's r'uture Assumption 2 Network behavior is sufficiently rational and quantifiable to be modeled and the resulting model will correspond sufficiently to reality as to be useful Assumption 3 All players in the game will be seeking to maximize one or more goals e g profits traffic security social welfare etc and all behavior in the model will be guided by these motives Assumption 4 An optimal network can develop in a situation wherein each component is optimizing and that such behavior can occur only under a set of process rules Assumption 5 Such rules or algorithms can be developed in a modular or parallel fashion and with repeated experimentation an optimal set of rules can be devised that achieve Assumption 3 i e system optimization without component sub-optimization E-2 These five implicit assumptions can be reassembled in terms of explicit assumptions which in turn can be axiomatized and programmed into the model These axioms can be altered independently EXPLICIT ASSUMPTIONS Assumption 6 Packet network traffic demand has a measurable growth rate and can be modeled as follows 1 Demand f computer costs GNP previous demand population The growth rate is sensitive to five factors cost of computation communication GNP growth rate previous demand population and a consumer taste variable Assumption 7 A network can be described in terms of a finite number of parameters as follows Given Compute Network performance A network topology A demand matrix parameters The performance parameters currently used are 1 Mean path length 2 Average capacity 3 Peak capacity 4 Average delay 5 Peak delay 6 Global reliability 7 Delta global reliability -EsX Assumption 8 Different topologies can be tested until an optimal network is designed The fastest convergence on such a design is produced by human machine interactions as follows goal system state parameter calculation issumption 9 Each configuration can be analyzed on a by-component basis example by IMP in the following manner Analyze Given Case and a traffic matrix ir PS in analvsis 1 2 3 4 2 1 3 4 3 1 2 4 4 1 2 3 5 All This iterative method yields each IMP's contraction to the network depending on its location and linkages 2 Cor lr-jyjtion f topology performance Assunptior 3 0 Each sub-net's share of the traffic and revenue pie is computed ac-cordiivj to that sub-net's contribution to the network i e E-4 a suD-net or IMP contributing very little in terms of capacity or reliability or occasioning delays will not receive much traffic throughput 3 Share of the pie - f contribution Assumption 11 All sub-nets or IMPs charge the same price to all customers regardless of distance volume or other function Therefore a sub-net's or IMP's revenue is proportional to its share of the pie 4 Revenue - f share of pie traffic fixed charges Assumption 12 A sub-net's or IMP's attractiveness to owners and investors is its profitability which in turn is based on that sub-net's and the other sub-nets' shares 5 Profit f maximum share actual share Assumption 13 Sub-net or IMP owners or potential owners who operate under a narrow profit-maximizing criterion will be motivated to change their IMP's location or linkages until profit is being maximized The model is therefore subject to dynamic growth and change E-5 IMPLICATIONS One mode of implementing such a model has been previously described in Paul Goldstein The Proposed ARPANET Divestiture Legal Questions and Economic Issues CAWP #101 1973 July 27 and in Marc Porat A Decision Tree Addendum to CAWP #101 CAWP #102 1973 August 8 They have been alternately referred to as the trade association cooperative or consortium methods Under the aegis of a consortium of sub-net or IMP owners a set of orderly procedures are devised These rules guide both the individual investor and the industry as a whole to achieve opt Lr Ckl growth in the shortest time possible One way of avoiding dead end or potentially dan aging network alterations is to devise an entry algorithm with clear rules applicable to all members Assumption 14 The consortium uses the following two entry criteria for each discrete change proposed by one of the members of potential members 1 The proposed network will generate more traffic or revenue than the existing network 2 The proposed network performance will not degrade the old performance standards 6 Entry f network revenue traffic performance Assumption 3 5 The consortium can change its o i rules of entry according to a charter E-G Assumption 16 The ARPANET divestiture might be useful as leverage in developing the consortium charter Note sharing mode Inc the model is to be run in BASIC Interactive timeThe model resides in the Interactive Application system and can be accessed by typing GET-INEC The INEC model is written for a maximum 10-nodc network It 2 occupies almost 100 000 words in BASIC and can therefore not be expanded to a 20- or 40-node network If such a network is desired the model can readily be translated into another language to run on a larger system Cabledata Associates Inc maintains an account at Interactive Applications Inc Contact Marc Forat for arrangements 2 Maximum sir' e 100 000 words E-7 CJ 2 O o o 1o wi z D O H o S W S a rf o H s o 2 o u w 9 o o O % 9ft oP A r- w H H to H O 0 oH 0 0 S o u o ss Al o -o n ro o8 -a co -P a a H W n orl tu E-8 I NEC 1 PEM # o ooo o o oo o o ' ' 10 PEN INDEPENDENT f' nrE tCCfuTr-'iC 'IHfCI'KIFTE PEN IHFLEDHT 'o'EP'TCM U PEM ' 0 DIM CM lU'J 105 PRINT -I'D YOU r ieD H UTE 'I MM o' JHL''W MO 120 IM-UT c IP r i -f fi y-EN 130 IF Ci VF - -Hen I 05 IW 150 165 18 0 195 210 Ch'Hlti INTPC pc-iriT DC vci r rr M DICTirNSPV J an PEN INPUT C't IF Cf H3 Tr N 24 0 CH IN DICT P EM 270 30 0 315 33 'J 45 36 0 375 405 420 4 35 45 0 4 65 43 0 4 Qp er j 10 5 o5 C 40 crcrcr i r i 5 70 c t- cr 600 6 15 h 3 0 A 45 660 r -t cr 6 90 05 o 20 i - c 50 65 30 95 o7' 10 z 25 z 4i i SS i i i - - EM PEM CCNTPIEUTIMN PIE ravrriTM --w r- PEM PEM PEM PEM PEM C EM 'EM 1 IIMEN 'ID -'I r EXISTING o'Z-T i ooQLD ' N y MMT C HM -o n C E PEPFC-MPMCG PP-TIC MC'THLY E TH FPTES PEVENUE3 TC I' PS NETt JD L IMAGES riCI'EL EVPLUPTICM IM 3 INVSNTCPV F '3 H L j P PEM EP'IC TPpFFIC r-M MD 3 T I M M PEM PEM PEM PEM PEM PEM PEM PEM PEM PEM EM PPHPIT HMD LC33 TC P1PS Li INVESTMENT TH-EErfJLD CGHVCpiiHTICf'PL I ' PUT Et Cv D-I CHPPhCTEP STPING BflTfl ' CONSTANTS INDS Ml HE DAT Pi INPUT P-B D T FRIF T TDTPL NETUHPf- TpfiFFIC Tl T TOTHL NETWORK REVENUE PI '2 DIMENSION DECI PFPTION DIM TIM DIM DIM niM DIM DIM DIM DIM DIM DIM DIM EC 103 EC 103 DC 103 EC ICO-HC 101 K 10 3 CC 103 PC 103 UC 10 3 VC103 we 11 r 3 DC 7 3 FC 73 PC 7 3 LC10-10 3-MT 10 103 SC 10 i0 3 TC 10 103 PC 10-10 3 GC120 23 air 19 3-EtC 13 PIC503 PEM PEM INITIALIZING THE EDEL is THE IMULPTIL N CLOCK KEPT FY KCNTWS TD CHPNGE TIME - INPUT T TD THE CUMMPND MODE K PEM Cl l FOP l- Cl TG 120 IF DI YES THEM 370 C1-1 GCJSUE 3175 GOTO 960 E-9 ooro o o o-- ' 0 0 15 0 945 'S 0 75 rp n 1005 1020 1035 1050 1 065 1 fi o 0 1 095 111 0 1155 114 0 1155 117 0 1185 1800 1215 1230 1245 1260 1275 1290 1305 1320 1335 1350 1365 1330 1395 1410 1425 1440 1455 1470 1435 1500 1515 1530 1545 1560 1575 1590 1605 1620 1635 165 0 1665 1630 1635 1710 1724 1725 1740 6n U0 7390 KF'tfjf i niTi H nn o v r 'i rp INPUT c IF r-n THEN 82'55 IF C-l THEN 3 35 GnTU 825 PEN P -- P U EP 2 ' oF F MTTE' PRINT cnM NP ' irpUT M IF ' THEN c'ZZf IF $ o p TH EN 1135 IP ET- I THEM 1830 IF BI L THEN 13 05 IF BI-- ' THEN 44 0 IF Et THEN 714 H IF T1- M THEN 11 TO IF F- T THEM 1575 PRINT CCMMFND MOT FECObNIIEIi DTO 960 CHRIN INTRO PRINT PEFGPT - INPUT R GUTU 1330 PRINT IMP HOP-TELETE INPUT Fi PPINT IMP 5 INPUT R bTJTrj 1665 c pirrr -LINK RPD--DELETE INPUT Ef PRINT LINK JJ'S' J INC'UT R B FDTD 1950 PEM PEPCPT RCCESS BY hUN ER IF R l THEN 5190 IF R 2 THEN 5325 IF R 3 THEN 5445 IF R 4 THEN 5550 IF ft - THEN 5655 IF R 6 THEN 58 05 IF H 7 THEN 5955 IF R 8 THEN 6235 IF R 9 THEN 6495 IF R 10 THEN 6675 PRINT REPORT fl HDT FOUND GCTO 960 PPINT INPUT TIKE WRX 1978 1 '' INPUT Ci C1 C1-19EURS 10 K R GD2UE 8175 GOTO 960 PEM DELETE IMP IF B f R THEN 1845 IF F D THEM 1140 IF RMO THEN 1200 IF PC fi 3 0 THEM 1800 N N2-1 E-10 1755 177 0 17C5 1800 1X15 15 DO ' '45 1360 1375 17 9 M 13''3 1 0 19 5 1950 1963 lr-30 1995 01 0 E0E5 040 04 7044 055 070 510 0 115 130 145 160 Z'i 7 i_ i i J 19 0 05 0 50 i T i 34 65 0 c 95 310 uic'5 34 0 -itr tr -- I ' ' 370 4 0 0 415 43 0 445 460 475 90 5 05 5 0 - cr -i tr C -' - J 550 565 530 FCfil 0 PPIMT -IMP e W DELETED 3D IG J6 0 PRINT IMP - r- ' C7 TnEc TO' I'ELETE GOTO r 60 F i'1 ADD IMPS IF FTR H THEN 19c0 ME N - -l Ff R -H -iriT IMP r RPl ED EGTQ 530 FPIfiT I'-P H RLCEFDV THERE C-GTD 960 PEM LINKS PEM DELETE Ll'v- IF R o P THEN f 5 PPIMT EP C -' PLERSE CETVPE FuTD 96 0 IF E'I H THEN 50 IF El oo THE ' 114 IF H' 10 THEN 035 IF EMO THEN 13 0 IF M FI -B I- I THEN 10 IF RCH 0 THEN 160 PRINT o iMP R r 0Ei MCT EXIST IM-D SIDLE TC GuTC 96 0 IF PCD1-0 THEM 160 PRIMT IMP BEES MGT EXIST EGTD 960 I FOE DELETE lit IDLE TC DELETE UW NCR FI-O MCE P --0 PPIMT LINKS R B DELETED GDTG 530 PPINT LINKS R B NOT JW PPE TD DELETE E-GTD 96 0 PEM LINK RDDITIGM IF R 10 THEN 30 IF EMO THEN 3 5 IF PCR O 0 T EM 310 PPINT IMP Ri DDES MGT EXIST LINK RPD TMFGE3IFLE 3GTG 96 0 IF RL E 0 THEN 355 PPIMT IMP VE DEES MGT EXIST LiriX ROD IMPCSSIELE GuTD 96 0 IF MCR-E 1 THEM 43 0 MCft B l MIB R 1 PRINT o LIMKS H F RrirED EGTC 53 0 PPIMT LINKS''SRJBr'PLFERDY THEPE GGTD 96 0 PEM IMP PPGP CP1JSIMG LINK DPGPS ''RMTGHRTICV' FDP 1 1 TD 10 IF I P THEM 565 IF MI I H 0 THEM 565 MC I -R 0 MC R I 0 PPIMT LirtK i R DPrPFED' NEXT I PPIMT NETHCR EV'RLUHTICM ' E-ll INPUT CT '610 c 4V c' ' 364 0 2655 - 6 3 5 iron 745 27 0 ii TO THEM THEN n i IF C P M f Hf- TNG R K LF-FILG PP'TNT RFERRPE TC r-H' R FILE IF3N -- Fr f i'E E' PPIHT 'o 1 ' P' ' -r- cr RNMfH - rvriTE PERT' PRINT 2FCG23 'H-pF 13' Tt tTC FRINT -3 ' r-U'E FCE 15 IFXCrij PEN PPOGFPM H FBU-E ENTER 15-H't G o FOP i--i re 10 u i 5t H i U $ K i G C'C i o NEXT I PRINT CC 1 o CCc ISGJiCC IJOt -IT 1 FPiNT MI -GI NC G FI GX- TI G I K FDP 1 1 TC 10 FOP J l TO 10 IF _ I vJ -0 THEN 2395 IF LCI fJ3 t rH N 925 PRINT 0 GOTO 94 0 PRINT '1 IF I J-50 THEN 2925 IF 1 10 THEN 2235 PRINT NEXT J NEXT I PRINT FOR 1 1 TC 10 FDR J l TO 10 IF MC I J -0 THEN 3090 IF MCI J ---1 THEN 3120 PRINT 0 GOTO 3135 PRINT 1 IP I J 50 THEN 3130 IF 1 10 THEN 3ISO PRINT NEXT J NEXT I PRINT PEM PRC3PRMMEC PfiUSE FOP SWITCHING TC THE NETWCPK EVALUATION ENTER ISO fi B TAPE IF 9 99 THEN 726 0 PRINT PREPARE TC MAKE THE NETL'DPK EVALUATION EfiTfi PRINT MO o TURN ON PUNCH 315 SEC PAUSE ENTER I5 fi B FDR 1 1 TO 10' FOP ' 1 TC 10 IF MCI J 0 THEN 3375 IF Mr I J 3 1 THEN 34 05 PRINT 0 GDTD 3 20 PRINT l IF I 50 THEN 3465 IF 1 10 THEN 3465 PRINT NEXT J NEXT I E-12 PRINT ' ' ' 30 5 320 339 0 29 0 2925 5940 29 5 5 297 0 985 3000 3015 303 0 3045 3060 3075 3 09 0 3105 3130 3135 3150 3165 3130 3195 331 0 3240 3255 3260 3270 32S5 3300 3315 3330 3345 3360 -o -o i J 3390 3405 3420 3435 3450 3465 34S0 3495 3510 - rr - cr - _' C _' i c-i er er _ i 3570 - -ZT ST i _i _l 3-f iJO 3615 3645 37 05 3730 - -- - r 'o ' -o J 3750 3765 3730 3795 3310 334 0 -o c- c- 3370 - i-t -i C7 3300 5915 3330 3945 3 9 u 3975 3990 4005 4 02 0 1 fj - cr t t_l J 4050 4 065 4 03 0 4 035 4110 4135 414 0 4155 4170 4 135 43 00 4315 4330 4245 436 0 4375 433 0 4 3 05 4 33 0 4 3 35 43 'JO 4 365 PPINT PGP 1 1 TO 10 FPINT K 1 IJ Gf STCc I 3I iC3'I 5 - T 4 I o_ o 7T5 1 1 PPINT TCt I 3 Ct TC r I j 3I-TC3 -71 iTC9 i T 10-1 1 NEXT I PPINT LIH-3' ENTIP 6 0'H p PPINT ENB OF 3E33IEN EVE rap PPM pcf i3 M rz '3 T THli o H f MT i ILE P OFET-EP CFPNGF 3 IN THE PEN NETUCEJ-' TQPCLC3V EC TFFCUGH p N 3-TY-'E E'v'PL i-'PTIEN PEN RLL PEL EVENT T3T9 13 3TEr'E3 EN P ThrE CR'ETTE F -EP T FE PEN C- 3Eh E ETHEP CCNVENTZNT NETTLN o NE - fiQT ICPTICP o PEN o o o o oo PEN PEPFCrM NCE rjUTPUT- -PEN NrVI IN N''ll 7 FEN rNI l NEPN PPTH LENGTH PEN N- I 3 RV'EPP E CPPPCITY PEN N'T 3 EPr CHPPCITY PEN N I 4 MVEPPGE TELHY PEN r i EPH TELHY PEK N I 6 '-HVE-PC FELIPPILITV PEN N o I 7 -MINI NUN PEL T ET- M I TV PEN 1 1 TQ 11 EACH r G'iT'flifii- rHE - VPPpV EF TPTP POP TV PEN ITH INF'S CCNTPIFUTICN TG THE NET EPff SEE CCNT'TTUT lEC 'EM ECUHTIEM TELO PEN CDNTirfjHTICN PON I NT I PL I ZING PPINT LC9L GLl' CPEETTE C FfiPEP TPFE PEP CCNTTNU3 TIZN prjNPEP 1 1 TE 10 INPUT KI3 PL I H I Oi I 'CT I 3 -UC I 3 NEXT I INPUT CC 1 3 -DC 3 3-EC 3 3 DC 4 3 EC 5 3 EC 6 3 Ef 7 3 IN UT NlJH P1 Tl -K INPUT P'J 3 0 PEP 1 1 TG 10 IP I 6 THEN 4 065 IfF'UT Pf FOP J l TO 10 3 3 1 IF P3CS S3 0 THEN 4140 LC I J3 1 GETD 4155 LC I 3 3 0 NEXT J NEXT I INPUT P3 3 0 FDP 1 1 TO 1 CiIF I 6 THEN 45 0 INPUT c 3 FDP ' 1 TO 10 s - i IF Pf 3 33 0 THEN 43 35 MC l J 3 1 OTTD 4 35 0 MC I J 3 0 NEXT J NEXT I E-13 4 69''' 13 ' ' ' o o44 iju 141 i i 4 i25 44-i n 4465 4500 4St5 4530 43 45 4 6 0 4575 45 0 46 05 -65 0 4 63 5 465 0 4665 466 0 4 6' 5 4710 4725 474 0 4755 477 0 4735 4 300 4S15 433 0 4645 4'3 6 0 1 - -7 er ' o r J 469 0 4905 496' 0 4935 495 0 4965 4960 4995 5010 502 5 5040 5055 507 0 5035 510 0 5115 5130 5145 5160 5175 5190 5205 522 0 5235 PRINT P -KFP' -S TC J' rl'r r 6 f-L ENTER t'j MTF FC - i TQ 11 INPUT fiC I 1 NC I 2 r-r I 3 3 NEXT I PRINT fFC INPUT o f FEN CCNT JFUTION ECU HON PEN 1 1 HMF K'3 F P fREI'F K2 3 4 K 7 4 ' o -1 L'MT li'H o i 6 3'f'C i r I 5 1 o rc t- n 1 I T IH 1 'oo2 3 M3- 7 FOP 1 1 TG 10 Hl i l a--t'C I i 1 o c l i K' h I t23 ' 'T4 nt I 4 ' ' o b 6 4L I '6 I 1 H3 U3 -K3 NC I o 3 o' r5 -C I 5 J - j 'K r riL I CCI Hl ft3 fl3 NEXT I PEM INITlHLlZINE THE CONSTRNTS P2 T2 Z 0 pcM ECONOMIC PEFO-T COMPUTATIONS PEM -- PIE DIVISION E -- FGP 1 1 TO 10 Z Z C I 1 HEXT I FOP 1 1 TO 10 EC I CCI Z r E T i PEM -- TOTAL TPPFFIC T2 -- SOSUB 3175 FOP 1 1 TO 10 FDP J l TO 10 IF ML I JJ 0 THEN 4920 T2 T TC I J HEXT J HEXT I PEM -- IMP REVENUE I J -- FOP J l TO 10 It J EC J T2 T NEXT J PEM -- TOTPL PEVEHUE R2 -- FOP J l TO 10 P2 P ICJ HEXT J PEM -- IMP PPOFIT 1 LOSS V I -- FOP 1 1 TO 10 VCI '-EC I -l'-H - T2 T HEXT I GOTO 6915 PEM REPORT GENERATION PEM PEPCPT 1 FPIHT LIN 2 PRINT TFPCFIC lEMPiNr 1963 1 PRINT P'HCrET THPOUEHPIJT IN ' 0006 PRINT 12 3 4 5 6 7 PRINT E-14 IF n-3 Tu-M 5 1 r f NAT FP INT IJ ING 5330 ' 9c 10' 5r' 5395 5 10 54 00 PRINT _IN G2TTJ 960 FEM PEPCPT 3 PRINT LIiK3 PRINT TPP'FIr EEfJC 1 634-'II GuTC 5305 PPINT LIN-3 MAT -PINT u iM' 5330 T 54 1 5 PPInT LIN'3 ' 5430 5- 90 5505 55c-0 GETD 960 PEM -- FEFCP7 3 -- PPINT LIN--3' PFIN7 GLD ET' CPPPINT LirK3 NAT PRINT L PPINT LIN- 3 c- c -i er - _o Z -' GDTD 960 555 0 PEM -- PEPCPT 4 -- PPTNT LINv3 PPINT MEM NET'C-K TCFOLCGY PPINT LIN' 3' ' MAT PPINT M tz - - er ' o C ' 534 0 5 55 5 335 Ci IC _ t I ' r- - - ' o o ' '-- -- cr er er _' -' _' _' 5SI'I 10 -- 64 0 er er 670 c - er 2 i c- 700 715 73 0 er 745 er 760 c i 77 790 5 SO 5 c Rpfi j tr 1 o - '- - ii 395 rr 910 er Or -e er 94 0 er i er ' S er 970 5 h - 0 0 0 r 015 t 030 045 6 06 0 075 090 h 1 05 i 130 6 135 er i i- o 10 T33GLC9Y PPINT LIN''3 GCTU 9t0 FEH -- PEPCPT 5 -- PPINT LIN 3 PPINT OLE INPS INVENTCFY PPINT LIM 2 FCP 1 1 TD 10 IF C'C I 3 0 THEN 5760 PPINT C I 3 NEXT I PPINT LIN'- 3 GCTO 96 0 PEN -- PEPCf-'T 6 -- PPINT LIfK PPINT NEW IfIPi INVENTORY PPINT LIMCS FDP 1 1 TO 10 IF PC I 3 0 THEN 5910 PRINT FT 13 NEXT I PRINT LICKS I3DTE5 960 PEN -REPORT 7 -- GD3UB 7365 PE3TCPE 6075 PPINT LIN PPINT NETWORK PERFORMANCE nUTPUTT PPINT LIfK2 PPINT OLD N1 PPINT DATA MEAN PATH LENGTH DATA AVERAGE CAPAC I TV DATA 'PERK CAPACITY DATA AVERAGE TELAY DATA PEAK DELAY E-15 K-1T- 0 I'RTR RVFFRGlV PrLIi'iFILITV CRlM MIHJ JM FLIHFTLITV - i - n - 1 1 TO 7 1 ' o- o FRf Hi o o- ' 1 U Fc i ttT o o c4 0 IMR -E IJ-IH6 6225JHT OC I cc I 3-FC I j-CC I F I r I 19R o5X o ' 3D 3D 2X MEXT I LIM- 2 GCTO 360 EM -FEPCPT 8 -P IMT LIH''c PPIMT 01 S ECCM'JJTC PEPDPT PPIMT Lir '2 FPT -JT IM t THFV 1 iV CHMT-IP'JTIGM TRB ' 5 v PPIMT HRP1 ET HHPF TRB 41 PE'v'EMlJE T7S''53 PCFIT o LDT3 ' PPT MT FPIMT - - - cr _ C C -' 6 3 l_i 0 o- 315 63 0 6360 6390 64 0 5 64 c 0 o 435 645 0 646 5 643 0 6495 651 0 6525 654 I'I 6555 657 0 6535 66 ft n 6615 663 0 6645 666 0 6675 6690 6705 6720 6735 6750 6765 6 80 6795 6310 6825 684 0 6855 6870 6385 6900 6915 6930 6945 6960 6975 699 0 7005 020 7035 IP fi TJ-iEH 6585 FCP 1 1 TO 10 IP or I 3 0 THEM 6450 PPIMf I TRE'- 12 EC I TiTRB 6' DC I J' TPB' '40 HC I 11 TFiF''52 J'JC I 1 NEXT I PPT MT LIM 2 GOTO 960 PEM -- PEPDPT 9 GC3UB 7365 PPTMT LIN62 PPIMT MEW ECCNGi-TC PEPGPT PPIMT L IM- 2 GOTO 63 45 FC1P J l TG 10 IP PC J 3 0 THEM 6630 PPIMT J TRS 12 CC J 3 TREE'S -EC J 3 TRB 40' IE J 3 'TPl- 52 JVC 13 ME XT J - PIMT LIH 2 GOTO 960 PET1 -- PEPDPT 10 -- GD3UB 7365 PPIMT LIM 2 PPIMT METMGFK SUMMRP'V PPIMT LIM' 2 PPIMT TRB 12 TOTRl TRB' 26 ' o RVEPRGE PC'IMT TRB'- 33 TDTRL JTRBf 50 RVEPRGE PPIMT TftPaOj-NETWGPK TRp 26 5 IMP o PPIMT TBB 37 NETWD iTRB' 53 5 IMP PPIMT TRBC11 ' TC'RCFIC ' TftB 6 J-TPR-PIC PPIMT TfiB 37 PEVEr iJE TRB '30 ' PE''EMi E PPIMT GLD TRB 11 T1 TflB' c6 Tl-Ml -TRB' '37V' P1 oTRP- 50 - P1- M P-TMT NEW TRB' 11 T2 THB'-26 T2 --MC TRB 7 C TRF- 50 P2 -r-'2 PPIMT PRTIG '-TRE 11 5 T2 'T1 PPIMT LIM GGTD 960 PEM iM'JLnTIDM PESET z o PEM pPDPD'fiL RCCEPTRNCE FCP 1 1 TO 7 PC I 3 HC 11 13 NEXT I FOP 1 1 TC 7 FC I3 PCI 3 DC I 3 Z Z PC I 3 E-16 7 50 7065 roio 7095 71 10 71 5 i IF Z 'IM 2 THEN 71 10 PRINT PErcGi-MEr-T REJECT CHRMGE z z Z-DTD 930 PPIHT FECCCMMEND RCCEPT CH GE 7 'Z GDTD 9-50 7140 155 170 7185 7200 EM PFGFC ED CHRNGE GQUJE 7365 MRT C ZER MAT E ZER MRT I ZEP 7215 MRT M Z P 733i 7245 72 0 7275 7290 7305 7320 7335 r-'HT P-ZEP MPT P ZEP MRT T ZEP MRT V-ZEP 2 0 2 0 T2 0 PRINT PCCDD ED CHRMGE RECPTED NEXT HFTI-TED 7350 GOTO 7650 73 5 PEN ZUEPDUTlriES CHECKS IF EVPLURTICN 13 CCMFLETEfi 7330 IF C l TUSH 7425 7395 7410 7425 744 0 455 7470 74S5 7500 7515 7530 7545 7560 7575 7590 7605 7620 7635 7650 7665 7630 7695 77i0 7725 7755 7770 7735 7300 315 733 0 7345 7360 7375 7390 7905 7920 7921 7935 PRINT 2GFPV YUUP PFCFTiSETj CF 'REZ HAVE NOT YET IFEN EVRiURTFU 5TJTD 9 0 RETURN PEN P-'DFDZED CHRNGE IM LEMENT Ii G3SU 7365 MRT P C MRT D E MRT H I MRT L M MRT G P MRT Q P MRT 3 T MRT U V N1-N2 R1 P2 T1 T2 PRIMT PPQPDSED CHRNGE IMPLEMENTED PRINT PRINT CONTINUE INPUT D$ IF D I f G THEN 7770 IF Ii VES THEN 7755 GOTO 7665 NEXT K PRINT S'RVE FILE INPUT C3 IF C 'I ND THEN 7260 PRINT LIN 2 R-99 GDTD 264 0 PRINT END CF ZESSION EVE STOP PEM I NT IRLI ZING THE C'ESIEENT VR RELE2 PFM MRT G ' M TRFPIF T PEN MRT 3 ZTCFEZ THE ER IC TPRFFTC MRTF'IX -19 2 ' PE3TCFE 7 -'5 ERTR 3 2 Oc 96 14 53 - 2- 9 4 14 o - - 53 E-17 '365 -4'J L' -iT i 35 4 1 i 33 o -j 6 44 c o 16 ij -t rl 01 0 T o L J CT r c- o3 t o 0 o5 10 1 1 o31 o9 6 L'AI H ' ' o c' ' J' 4 oi_' o ' o L 4 ot- jc 1 o 4 j DATA 93 6 355 3 43 J 0 35 7 7 3 DATA t 7 o5 43 6 3 13 7 43 o 3 o 3 MAT PE9D SC 10 10 J PEM G 13 THE MGMTh'LY Si-'CWTH PATE GF I'P Tfl TRAFFIC PEMANP G l 01 I - - I H H j40 070 0 1 01j 1 15 1 ' j 45 160 1 75 1 90 3 05 o- 3 0 l-_ - SO L 65 30 95 3 10 1 1 1-1 iH IT - - C i - 4 0 o370 3335 34 00 3415 343 0 3445 3460 3475 3490 3505 3530 tr -r 'o_ 1 J 3550 3565 3530 3595 3610 86 5 364 0 3655 3670 3685 PEM rPPPTF 3TDPEB IN T T 30 PETU N PEM SUPPnuTINE'i BFTNGIMG MAT T--S FCP H-l TG i i MAT f'HE MODEL L'F TO IFEEP T- G T NEXT H PPIMT 3IMULHTION TIME 1963-KCT '1 0 PPIHT PERSYtt PETUPH PEM INITIALIZING A MEM 0 PUN PEM -- B D H' J LUQ a -- FDP 1 1 TG 10 Et I 3 100 DC I 1 HU I 1 1 IT I 3 1 Q I 3 I HEXT I PEM INIT 0 5 PEr-CPMRMCE DATA FGP 1 1 TQ 7 or I 3 1 NEXT I PEM IM IT MAT ' LINKAGES MAT L IBN PEM IM IT CONSTANTS A B D 0 F l Pl Tl t Ml M iO PEM IMITIALIZIMG THE BLACK BOX MAT N L MAT P 0 C1 G03UB 3175 GOTO 960 END E-18 IfiEC DICTIONARY THE USER WILL FPEOUEMTLV BE PPCHPTED PGP INPUT TO CONTROL THE MODEL OPERATION WHEN THE MODEL IS IN THE CCMMHHr MODE ANY CF THE FOLLOWING INPUTS NAY BE USED 'TYPING THE FIPST LETTER ONLY IS SUFFICIENT ' fi A DD TO ADD AN IMP OP A LINK D IKELETE TO DELETE HN IMP OR fl LINK E E VALUATE TO SEND THE FPOPUSEB CHANGES TG THE NETWORK EVALUATION MODEL 6 GC0 TD APPROVE RND IMPLEMENT A PROPOSED CHANGE FOLLOWING R MODEL EVALUATION o I KNP TO INITIATE RN IMP DFEPRTICN SEQUENCE L L IMO TO INITIATE A LINK OPERATION SEQUENCE N N OGG TD REJECT AND ABORT A PROPOSED CHANGE FOLLOWING A MODEL EVALURTION P R EPDRT TD CALL UP ONE OF THE TEN REPORTS REPORTS TRRCFIC DEMAND NETWORK TOPOLOGY IMPS INVENTORY NETWCPK PEFFCPMAhCE ECONOMIC PEPOPTS NETWORK SUMMARY T TUME OLD MEW DOTH 10 TO DRING THE TRAFFIC DEMAND MATRIX IjP TD AMY USER SPECIFIED TIME THE USER WILL PLSO BE PROMPTED POP ''YES' AMD 'NO' ANSWERS AND FOP A '0' OR '1'' RNIVEP IF AN INPUT ERROR IS MADE THE USER WILL USUALLY IE PUT BACK IMTO THE COMMAND MODE FOLLOW ING AN EPP-DP MESSAGE E-19 SAMPLE INEC RUT The following is the output of an INEC session shown is for demonstration purposes only The machine prompts the user with a statement followed hy a ' ' GET INEC The data The user responds with the appropriate command RUM ' INEC DO YOU NEED A USER'S MANUAL YES NO N0 DO YOU NEED A DICTIONARY N0 IS THIS A NEW CO' OR CONTINUATION M' RUN 0 SIMULATION TIME 1968 2 READY C0MMAND7R REPORT --_ TRAFFIC DEMAND 1968 1 PACKET THROUGHPUT IN 000S 12345678 33 205 85 10 33 250 437 3 92 17 202 19 54 1194 29 1 1 2 1 5 96 63 170 24 24 34 81 39 255 48 14 1054 22 5 I 9 9 2 8 1 52 34 30 1 7 66 16 8 43 2 249 1 23 9 59 475 1 64 0 12 414 I 68 5 7 1 101 6 35 I 3 2 44 2 11 81 9 321 7 42 9 10 89 1 238 8 39 0 35 4 53 12 16 2 60 5 2 5 2 2 7 3 C0MMAND7R REPORT # 2 I 2 34 209 87 o0 34 255 446 3 94 17 206 19 55 1218 30 1 1 2 I 5 DEMAND 1968 a PACKET THR3UG HPUT IN 000S 7 6 4 5 3 98 64 173 24 24 35 83 40 260 49 14 1075 22 5 1 9 9 2 8 1 53 35 3 1 7 67 16 8 44 2 E-20 254 1 23 9 60 485 1 65 0 12 422 l 69 5 7 1 103 6 36 1 8 9 10 3 2 45 2 11 83 9 327 7 43 91 1 243 8 40 0 36 4 7 3 54 12 16 2 61 5 2 5 2 2 COMMAND R REPORT # 3 A T represents a link a '0' represents no link The topology was initialized as the identity matrix i e no links OLD NETWORK TOPOLOGY 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0' 1 0 0 0 c 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 I COMMAND K REPORT #77 SORRY YOUR PR0P0 2D CHANGES HAVE NOT YEY BEEN EVALUATED COMMAND R User asked for net REPORT 0 8 work performance outputs which arc as yet unavailable OLD ECONOMIC REPORT IMP # 1 2 3 CONTRIBUTION MARKET SHARE 100 100 100 4 too 5 6 7 8 9 10 100 100 100 100 100 100 o21 REVENUE PROFIT P CL0SS 1 1 1 1 1 1 i 1 1 1 1 1 I 1 1 S 1 1 1 1 COMMAND K REPORT 019 SORRY YOUR PROPOSED CHANGES HAVE NUT YET BEEN EVALUATED COMMAND R REPORT # 10 SORRY YOUR PROPOSED CHANGES HAVE NOT YET BEUR N EVALUATED COMMAND IMP ADD DELET A ItMP 12 IMP if 2 ALREADY THERE COMMAND LINK AUD DELETE7A LINK # S E G 2 4 l 2 LINKS I 2 ADDED NETWORK EVALUATION N0 COMMAND L LINK ADD DELETE A LINK 0 S E G 2 4 1 2 LINKS 1 2 - ALREADY THERE COMMAND L LINK AUD DELETE A LINK S E G 2 4 2 3 LINKS 2 3 ADDED NETWORK EVALUATION NO COMMAND L LINK ADD DELETE A LINK # S E G 2 473 4 LINKS 3 4 ADDED NETWORK EVALUATION N0 C0MMAND L LINK ADD DELETE7A LINK # S E G 2 4 4 5 LINKS 4 5 ADDED NETWORK EVALUATION N0 C0MMAND L LINK ADD DELETE A LINK 4'S E-G 2 41S f LINKS 5 6 ADDEJ NETWORK EVALUATION N0 C0MMAND7L LINK ADD DELETE A LINK #'S E G 2 4 6 7 LINKS 6 7 ADDED NETWORK EVALUATION N0 C0MMAND7L LINK ADD DELETE7A LINK #'S E G 2 4 7 8 LINKS 7 8 ADDED E-22 Tha user Is adding links to the network INEC accepts new links for evaluation C0MMAND7L LINK ADD DELETE7A LINK 'S E G 2#4 9 10 LINKS 9 10 ADDED NETWORK EVALUATION N0 C0MMAND L LINK ADD DELETEVA LINK t'Sf E G 2 A l S LINKS 1 5 ADDED NETWORK EVALUATION NG C0MMAND7L LINK ADD DELETE7A LINK 0'S E G 2#4 i 10 LINKS 5 10 ADDED NETWORK EVALUATION N0 C0MMAND7L LINK ADD DELETE A LINK 7 S E G 2#4 W10 LINKS 1 10 ADDED NETWORK EVALUATION N0 COMMAND R REPORT # 4 and this is the resulting topology NEW NETWORK TOPOLOGY 1 1 0 0 1 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0 1 1 I 0 0 0 0 0 0 0 0 1 I 1 0 0 0 0 0 i 0 0 1 1 1 0 0 0 1 0 0 0 0 I 1 1 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 1 I 1 1 0 0 0 1 0 0 0 1 1 E-23 At this point the user is ready for an evaluation All relevant data are dumped A more recent version of INEC accepts output on a To a5 Instxuiacuts Model 733 tape cassette C0MMAND7E NETWORK EVALUATION Y S PREPARE TO MAKE THE HALF FILE VHE FIRST OUTPUT WILL BE USED TO RELOAD THE MODEL 0N A CONTINUATION RUN N0W TURN 0N PUNCH IS SEC PAUSE 100 1 1 2 3 4 5 6 7 8 9 10 000 0001 100 100 100 100 100 100 100 100 100 o0001 0001 10 0001 0001 10 1 2 3 4 5 6 7 8 9 10 0001 0001 I 1 0000000000100000000001000000000010OO0UO0000100000 00000100000000001000000000010000000000100000000001 1100100001 11100000000 1100000000111000001001110001 000011100000000111000000001llOOOOOuOOll11000100011 PREPARE T0 MAKE THE NETWORK EVALUATION DATA NOW TURN UN PUNCH 15 SEC PAUSE 11001000011110000000011100000000111000001001110001 00001110000000011100000000111000000001111000100011 33 6633 33 6633 209 12 86 7085 10 201 255 025 445 784 17 3417 3 0603 o 93 8492 206 06 55 0854 29 5829 19 3819 1218 51005 5 10OS 61206 1 0201 2 0402 173 417 t 97 9296 24 4824 24 4824 64 2663 34 6834 48 9648 39 7839 260 125 82 6281 14 2814 1 0201 1075 19 22 4422 5 1005 2 0402 9 1809 9 1809 8 1608 61206 34 6834 7 1407 53 0452 71407 30 603 43 8643 3 67 3266 16 3216 8 1608 2 0402 60 1859 254 005 1 0201 23 4623 9 1809 484 547 71407 65 2864 12 o 2412 0 7 1407 422 321 1 0201 5 1005 69 3668 71407 71407 6 1206 103 03 35 7035 11 2211 44 8844 2 0402 3 0603 2 0402 9 1809 327 452 7 1407 42 8442 82 6231 39 7839 242 784 90 7889 1 0201 t 8 1608 P 7 1407 0 35 7035 4 0804 3 0603 12 2412 2 0402 54 0653 16 3216 61 206 0 5 1005 5 1005 2 0402 2 0402 2 0402 END 0F SESSION BYE D0NE The session is over until the network evaluation parameters are returned E-24 GST-INEC RUM INEC DO YOU NEED A USER'S MANUAL YES NO N0 DO YOU NEED A DICTIONARY The user is ready for a continuation run The old system state statistics are fed in and the model is initialized N0 IS THIS A MEW CO' OR CONTINUATION C ' RUN t LOAD OLD CASETTE OR PAPER TAPE FOR CONTINUATION RUN 100 1 t 1 1 100 1 1 2 2 100 1 1 #3 3 100 1 1 A A 100 1 1 #5 #5 100 1 1 6 6 100 1 1 7 #7 100 # 1 1 8 8 100 1 1 9 9 100 1 1 10 10 0001 0001 0001 0001 0001 BAD INPUT RETYPE FROM ITEM 6 0001 0001 10 10 I 1 1 10000000000100000000001000000000010000000000100000 00000100000000001 000000000010000000000100000000001 9 11001000011110000000011100000000111000001001110001 000011100000G00111C00000001110000000011 I 1000100011 INITIALIZING INPUT 0 K PREPARE TO INPUT THE NETWORK EVALUATION DATA 2 3 15 1 5 2 At 4 8 999 865 2 6 18 2 0 2 0 4 0 998 840 Then the network 7 evaluation para2 8 ll#1 4 1 9 3 9 999 865 meters are loaded 1 9 09 1 7 2 7#S 7# 998 872 3 0 04 1 1 3 5 4 8 801 720 3 1 25 t 5 2 6 5 0 993 884 2 4 20 1 8 2 5 5 2 998 868 2 6 17 1 2 2 5 5 1 996 825 E-25 NAG INPUT 0 K SIMULATION TIMES 1968 HEADY KCCCOMMEND ACCEPT CHANGE Z COMMAND7R REPORT 077 I NEC analyzes the net- v ork analyoiii and iramediately recotraac oda to either accept jr rejecc Mie change The user then accesses several reports to aid in the decision 137390 NETWORK PERFORMANCE OUTPUTS OLD NEU 0001 0001 0001 0001 0001 0001 0001 f EA 'i PATH LENGTH AVERAGE CAPACITY PEAK CAPACITY AVERAGE DELAY PEAK DELAY AVERAGE RELIABILITY MINIMUM RELIABILITY 2 5 17 1 7 2 5 5 999 o 87 OIFF RATIO 2 4999 1699 1 6999 2 4999 25000 1700 17000 25000 50000 o 4 9999 9989 3699 9990 8700 COMMAND R REPORT # 9 NEW ECONOMIC REPORT MP CONTRIBUTION 1 2 3 4 5 6 7 8 9 10 MARKET SHARE 115943 110557 4 19293E-02 128404 1 69S18E-02 16435 201113 9 78669E-02 4 23895E-02 8 04654E-02 52 353t 49 9216 18 9331 57 9806 7 i68U 74 2119 90 8122 44 1916 19 1409 36 334 REVENUE PROFIT 8692 51 8288 68 3143 54 9626 75 1273 17 12321 7 15077 9 7337 31 3178 04 6032 67 1 195 28 791 451 -4353 69 2129 o 52 -6224 06 4824 47 7580 68 -159 922 -4319 19 -1464 55 COMMAND R REPORT # 10 NETWORK SUMMARY OLD NEW RATIO TOTAL NETWORK TRAFFIC 1 3748 61 3748 61 AVERAGE IMP TRAFFIC 1 374 861 E-26 TOTAL NETWORK REVENUE 74972 3 74972 3 AVERAGE IMP REVENUE ol 7497 23 4 COMMAND 6 PROPOSED CHANGE IMPLEMENTED After reviewing the reports the user decides to implement the change E-27 Appendix F USER'S MANUAL TO INEC INDEPENDENT NODES ECONOMICS SIMULATION MODEL by MARC U PORAT o INDEPENDENT HCPE ECONOMICS o I NEC ' SIMULATION MODEL J- P-- MR NIJRL CABLEBATR R i'SDC IRTb' PALD HLT2' Cfil-IFCPN VERSION HUG 3 1973 MARC U POPPT TABLE 1 0 0 3 0 DF INTPODUCTICN F-2 LOADING THE MODEL F-3 2 1 R NEW PUN a P CONTINUATION FUN Rl TFPING THF NETWORK 3 1 IMP NODE RDDTTTONS S DELETIONS 3 3 LIN- riNTERNOrE ADDITIONS DELETION 4 0 GENEFATING R REPORT 5 0 SAVING A FILE e O CONTENTS F-4 F-5 F-6 5 1 SRV INS R HALF-FILL 5 3 SRVING P PULL FILE NETUOPK RNRLVSIS 3 EVALUATION F-6 7 0 IMPLEMENTING OP ABORTING A pp PQSED CHANGE F-7 3 0 I NEC DICTIONARY F 8 F-1 rn RCEE PORTIONS EC THE i rEP r v-i i-- i_ o -i ER E TVPE T HE SECTION NU -UEF E G E E DP o IN THE L AT E r 'H E YHU MILL RECEIVE fit L CONTENT E lECTTOH ' VGiJ IJP NT THE MPGLE 'o P-URL cLr YCE 'ALL P TY E BONE' 'HEM vcu KY-'E FINI ED t K I NO THE U --Y--M o WP pop PEnrv TCI RETURN TO THE f- RTH P DOERN USER RESPONSE N-LL 1 U IHTPQUUCTIDN THE I NEC MODEL WR DESIGNED INAPTLY AS A GAN IMG TDCL OP' NETWORK I IMULRTION THE FLAYERS IN THE GAME ppc ENTPEPENEUP OHNER G I-EE E- IM L'B-HETE AND UUNE-S GP HOST MACHINES THE Uzcrr OF THE GFVE I TO TH E AH EXISTING NETHOPK oMINIMAL ' AND INTPGB'-GE SUCCE - SrUL CHANGES IN THE NETWORK TOPOLOGY SUCH Ti-AT THE NETHGP Z FEFFEPNNACE AS A WHGLE I IMPROVED E C H LAYER FE-FESE'-TING A UF-NFT IN A COMBINED NETWORK HAS THE ELECTIVE EE M ' p'l ZING HI' -ALEE UP PROFIT THEPEFOPE- A 'GEEE' CHANGE EEC1 I NEC'S 'DU'T GP VIE' - IE CUE WHICH INCREASES BOTH THE MEMBER'S ECONOMIC POSITION END THE CDMBINEl1 NETWGFKS TOTAL TPPPFIC Cc' REVENUE r A FULL DISCUSSION C THE AC SUNF Tier- -- VHTPH UNEEPL IE THIS APPROACH MRV BE OBTAINED FROM CABLEDATA REEDE lATES UP 101 EDLIiETEIN THE PROPOSED RPFRNET DIVESTITURE LEGAL C'UECTIGNE ECONOMIC ISSUES ' IP 10S PDPflT P DECISION T EE ADIENDUN TG CAVR oi M UP 111 PGPRT R DELPHI EYERCIZE EXR INING FOUR RL TEFNRTIVE C COMMUNICATIONS GLICY OPTIONS UP LIE ERRRN PPELIMENPPV CONCEPT DRAFT C PGPRT HP ll'J PGPRT PROPOSAL ECP THE ARPANET GN FORMATIGN Oc R CC 'CN INTEREST CON SORT UM nr PACKET-SWITCHING ENTITIES TOWARD R DIVESTITURE PLAN FEP THE ARPANET HP 114 CEPF H RECHEST FOP THE FDLLQWIttS IS R THUMBNAIL SUr'hRPV GF THE R'SUEPTICNS 1 1 RLL NODES DP IMFS RC'E UNBEP INBEPENIENT CVNEPS I P l E R PLAYER NRY OwM MDPE THRN EHE BUT NGT RLL- I P 1 3 THE IMP D'-INEPS PPE IN CEM-'ETITIGN UTTH EPCH GTHEF FGP R SHRFE GF THE REVENUE 1 4 RN IMP'S 7-HRPE GF THE PIE IS CDNPLETEL ' BETEPMINET BY ITS CONTRIBUTION TG THE r-ETMCPK I E ITS LOCATION IN THE NET RND ITS LIHHPEES 1 5 RN ENTPEFENEUP MAY IMPROVE HIS IMP'S CG 'TPI BUT I CM BY PEflPPAf'GTNG ITS LOCATIGN LINKAGES 1 6 ALL THE IMP CP SUBNET GHNEPS APE FULL MEMBERS IN R PACKET CONSORTIUM THE CCNSCPTIUM SERVES IN TNG MAJOR FUNCTIONS A' AS A PAYMENTS CLERFINGHGUSE IN THE EVENT THAT TRAFFIC FLCHS THPGUEH EVEPRL INBEPENIENT NETWORS i- AND B' AS A CGGPDINRTINE MECHPr ISM TO INSURE THE SMGDTH OPERATION EE A CGMBINEB NETUCPK IN PARTICULAR- INTERFACE STANDARD AND MANAGERIAL SERVICE' F-2 LORD I HG THE MODEL U TO PUN THE MODEL TYPE 'GET- INEC- YDUP TERMINAL SHOULD BE EOUTPPED MITH A CRSETTE ME ' GPV CP FAPEP-THFE DEVICE TYPE ' 'IH- 'OU WILL BE PROMPTED WITH THE FOLLOWING C'UESTIOrts -IS THIS A NEV o CM' QP CGNTINUATICN U PIN- IF YOU DON'T KNOW TYPE 'o - IP YOU HAVE PLAYED THE GAME BEFORE PUT HAVE A PH-SICAL RECORD DF IT TYPE 1 1 INITIALIZING A NEU PUN Ier YOU TYPED --n-' THE SYSTEM LULL SE AUTOMATICALLY LOADED MTTH A STANDARD TUFTING TOPDLGGY CONSISTING CF fi TEN NODE NETWC K WITH NO LINKAGES YGU APE MOW rpPE TG ADD DP DELETE IMP PUT LINKS TO THE STARTING TCPCLOGY UNTIL MAXIMUM REVENUES PrE ACHIEVED INITIALIZING A CDNTINUATICN PUN THE MACHINE WILL PPCMPT YGU TG 'PREPARE TG INPUT YOUR OLD TAPE GP CA ETTE' THIS I THE PHYSICAL PECGPD OP A PREVIOUS PUN I E ' YGU A-'E PICKING UP WHERE YGU LEFT OFF LAST TIME AFTEP YDUP OLT DATA IS LCAI-ED INTO THE MCEEUJ IT MILL CRANK ITSELF Uc' TG ' PEED AND ADVISE YGU THAT IT IS -'READY' AT THIS POINT ALL THE ECONOMIC PEPDRTS APE PREPARED AND YGU NAY ACCESS THEM AT MILL SEE SECTION 4 CO F-3 2 0 RLTEC'ING 3 1 TW E I'TTI'TC IMP ADDITIONS AND DELETIDM- AMY L1FEPATIDN INVOLVING FN NCDEL BY IMPUTING RJ I TO T E MACHINE MILL THEN PECUEST fir' ''A' ADDITION OP DELETION ON THE IMP PECLOGiO IS THE IND HUMMER COMMAND I IMP ADD-DELETE IMP ' IMP I ANNOUNCED TO THE 'ofON F-ND' PFCNPT THCP A 'D ' TO SIGNIFY A IN' 'ENTGFY THE THIRD INSTRUCTION A IM THE ABO'-P EXAMPLE THE USER PEPUESTED TO PIT- IM TO THE INVENTORY IF THE CCE-AT'EM UP' IUCCE' SFUL Ju'r IvSTEM MILL SO TNFOFM THE USER AN OpEFATIGN MILL BE UNSUCCESSFUL IM TMO CASES R ATTEMPTING TO HDD RM ALFEAliV EXISTING IMP- AND o ATTEMFTTMA T 0 DELETE RM NONEXISTING IMP IF THE IJ3GF IT UNLLER FT Rr - ' POINT DUPING THE SIMULATION WHICH IMP'S EXIST- PEFOFTS - SH'TULD BE CALLED 'SEE SECTION 4 0 WHENEVER RM IMP IS I'PEpPED A J_ T E LINKS INVOLVING THAT Ih p MILL ALSO EE DROPPED AUTOMATICALLY THE MODEL MILL INFORM T-F USEP WHICH LINKS RPE BEING DROPPED 3 2 LINK RDDIT IONS RMD DELETIONS THE CPERRTIDNS INVOLVING LINKS r-t SI MILS TO THE IMP OPERATIONS IN PESPCMSE TO THE -COM 'AND PPOMc'T THE USE EMTEPS 'L' TD SIGMIPY LIN G ERATIGNS THE E T INI TRUGTIEN SHOULD BE R -R- GP B- TO SIGNIFY RDDTTIONS U DELETIONS rt 'HIFT ENTRY SHOULD BE RN IMP PRIP -- TMC MJMBEPS o SEPEPRTED BY CDMP COMNRND L LINK ADD-DELETE LINK 'S 3 3 D UNSUCCESSFUL RDD 0 DELETE OPERATIONS C-HN DC CUP IN FOUR EASES A RTTEMPTIMG TG RDD R LINK TG R NONEXISTING IMF f RTTEMPTIMG TG RDD RM ALREADY EXISTING LINK C ATTEMPTING TG DPCP R LINH PROM R NONEXITING IMP OP D ATTEMPTING TG DROP R NONEXISTIN' LINK THE SYSTEM MILL INFDFM THE USER IN THE2E CASES SHOULD THE USEP PECUIRE THE NETWORK TCPGLDGY PEPOPTS 3 RMD 4 RPE RVRILRBLE SEE SECTION 4 0'' F-4 4 0 REPORT GENEPnTIGN H USER MAY CALL HNY OF THE TEN CEFCPTC RVRILRPLE BY IMPUTING -P- 'REPORTS FOLLOWED BY R NUMIEP E G y - o p THE R- AILRILE REPDFTS RPE PEPDPT NEW OLD PFOFDSEB EXISTING 1 TPRFFIC DEMAND METWOPK TOPOLOGY IMP ' INVENTORY NETWORK REFFOFMAN IE SRLES HNRLYSIS NETWORK SUMMARY i2 4 - 5 I z Q 10 THE DISTINCTION BETWEEN -OLD-' RND -NEW SHOULD IE CLER 'LY UHTEPSTCGD BY THE USER THE SIMULATION BEGINS WITH Hfi -'OLD-' NETl-fDPK THE USER IS pPEE TO Q TEP THE NETWORK FY ADDING OF DELETING AN IMP DP A LINK RFTEP A 01 STETE CHANGE IS PROPOSED' IT IS EVALUATED BY THE SYSTEM '' SEE BEi CV - AND R REOCr KEM'iAT ION IS ISSUED TO IMPLEMENT DP ABORT THE CHANGE IF THE -'PDPOSED CHANGE IS IMPLEMENTED THE OLD SYSTEM IS SCRATCHED RND REPLACED BY THE NEW- SYSTEM ONCE THRT CPEPRTIOri HAS TAKEN PLACE A COPY IS MRDE CF THE UFTfiTED OLD SYSTEM RND IT IS CALLED THE -'MEW- SYSTEM FROM THE USER'S POINT OF VIEW OLD- REPORTS SHOULD BE C PL LET WHEN H HEED PRISES TO RSVIEi THE EXISTING SYSTEM I- NH NEW PEPDRTS SHOULD BE CRl LED WHEN THE USER WISHES TO REVIEW HI 'HEP GUN CHANGES BEFORE IMPLEMENTATION H MORE TERSE VERSION CF THE RBOVE PRPRGPRPH IS OFFERED CYCLE 1 CYCLE d DLD MEW OLD HEW 1 1 d CHANGES HEW 1 OPC'PO''RL OLD 'IHRIGES HEU S APPROVAL OLD ETC DMCE R CYCLE IS COMPLETED THE PREVIOUS 'OLD FILES RFE ERRS ED F-5 IRVING R FTLE fH'Z HhLr-FILE WHEN THE USER HAT PPUP'OTED R CHANGE IN THE NETJ 'CPK n GLLlGY THE MODEL ''ILL PROMPT -MODEL EVALUATION -' Ier ALL HE DESIRED CHANGES HAVE PEEfi ' Hr'E THE VIER 'oA ' ' E CE D YES' AT THIS POINT THE ' E_F ZHSuLD P GFCRE TO OUTPUT A FILE OF SYSTEM DQTA Cf-TC A PPCEP TfiPE 0 ' CRTETTE TA'-'E ALL THE OLD TECHNICAL RtsD ECONOMIC PEPGPTS MILL IE 'UNFED TUT IM PREPARATION PEP R LGNT TUIJRTICN UM LRTEP ON THI PHYSICAL RECORD MILL IE USED TC ELCRD R CONTINUATION FUN T p THE FIRST rILE IS MADE THE USER WILL IE- PROMOTED TO NA E R MODEL EVALUATION PILE THI I -ILE LULL IE ELT Tn A NETWORK EVALUATION MODEL WHICH MILL COMPUTE HEM TECHNICAL PARAMETERS FCR THE SYSTEM RPTPR THE FULL FILE RFTEP THE NETWORK EVALUATIONS RETURN AND THE CONTINUATION RUM EXECUTED THE USER M Y CHCCCE TO SAVE R -ULL EILE AS THE STARTING POINT FCR THE NEYT RUH THE FULL FILE CONTAIN' RLL TECHNICAL AMD ECONOMIC PEpCPTS o WITH T E PREVIOUS F rRO'ED CHANGES EVALUATED REPROVED RND IMPLEMENTED THE -HEMFILES APE ZEROED SD THRT THE FULL FILE IS APPROPRIATE rrp R CONTINUATION PELOPH U NETWORK EVALUATION THE PDPTIDN OF THE HALF-CILE 'SEE SECTION l SCr 'D ED OF THE PROPOSED NETWORK TOCOLOGY Rf- Il THE TRAFFIC DEMAND MRTPI iS SENT POP TECHNICAL EVALUATION THE ANALYSIS RETURNS OF PERFORMANCE DATA WHICH LATER' FORMS THE BPSIi FOP CSVU oUTIME EACH IMP'S CONTRIBUTION TO THE NETWORK UNDER THE GIVE TOPOLOGY THE RELATIVE IMP CONTPIEUTICNS THEN PRE USED IN COMPUTING EACH IMP'S REVENUE SHAPE THE PARAMETER' U ED IN THE NETWORK EVALUATION ARE r 3 4 C5 MEAN ATH LENGTH TRAFFIC WEIGHTED AVERAGE CAPACITY PERL CAPACITY AVERAGE DELAY PEAK DELAY 6 GLOBAL RELIABILITY DELTA GLDEAL RELIABILITY F-6 THE ANALYSIS IS CONDUCTED E CONSIDERING THE NETWORK IM AN ITERATIVE FASHION WITH ONE IMP DELETES FROM THE TOPOLOGY THI METHOD REVEALS EACH IMP'S IMPACT ON THE NETWORK' AS A iHDLE HENCE IT' CONTRIBUTION THE LAST ITERATION CONSIDERS THE ENTIRE NETWORK WITH ALL IMPS INCLUDED IMPLEMENTING DP REOc'TING A PROPOSED CHANGE 0 THIS IS THE LAST STAGE IN ONE CYCLE OF THE EMULATION THE USER HAS GONE THROUGH THE FOLLOWING STAGE o 2' ' 'C ' CD F INITIALIZED THE MODEL PROPOSED NETWORK CHANGES IMP AND LINE'S' PRODUCED A HALF-FILE ENDING A PORTION FOP EVALUATION AND RETAINING A PORT IOH FOR THE CONTINUATION PUN RECEIVED THE NETWORK EVALUATIONS RE-INITIALIZED THE MOTEL USING THE HALF-FILE AND THE NETWORK EVALUATION STATISTICS INSPECTED ALL THE NECESSARY TECHNICAL AND ECONOMIC AFTER THE HALF-FILE AND NETVCFf EVALUATIONS ARE LOADED THE MODEL WILL AUTOMATICALLY OFFER A EECCMMENBATIJN TO EITHER ACCEPT Cc' REJECT THE PPPOSED CHANGE THE ACCEPTANCE ALGORITHM USES TWO CRITERIA A DOES THE REVENUE AND OP TRAFFIC INCREASE AS A RESULT OR THE PROPOSED CHANGE'- - DDLS ANY DEGRADATION IN SYSTEM PERFORMANCE OCCUR' THE MODEL COMPUTE WEIGHTED RATIOS Cr THESE PARAMETERS AMD COMMUTES A ' Z SCORE THE USER MAY THEN CHOSE TO IMPLEMENT THE CHANGE BY TYPING ''GO'' OR ABORT THE CHANGE BY TYPING o'NOGO'' IN THE CURRENT VERSION r P THE MODEL - A USER IS FREE TO IGNORE THE MDDEL RECOMMENDATION ENTIRELY- AND IMPLEMENT OR REJECT SOLELY ON HIS OWN JUDGMENT A TYPICAL CONVERSATIOON MIGHT BE LOAD OLD CA3TTE OR PAC'EP TARE FOP CONTINUATION RUN USER LOADS OLD RECORD PNp ACTIVATES INPUTS LOAD NAC EVALUATIONS CRSETTE OR PAPER TAPE' USER LOADS HAS EVALUATION AND ACTIVATES INPUT RECOMMEND ACCEPT CHANGE Z l 34 TYRE GO- OR 'NOGO' GO PROPOSED CHANGE IMPLEMENTED AFTER A PPCPSED CHANGE HAS BEEN THUS IMPLEMENT ' 1'HE OLD PILE IS ALTERED TO REFLECT THE NEW CHANGES' AND THE CYCLE BEGIN' RPO 't THE BEGINNING F-- o- 0 INEC DICTIONARY THE USER WILL FREQUENTLY BE pPG'iPTED PGP INPUT IS CONTROL THE MODEL OPERATION HtN THE MODEL IS IN THE COMMAND MODE Rt-sY DF THE FOLLOWING INPUTS MAY IE USED oTYPING THE FIRST LETTER ONLY I SUFFICIENT ' fl FKIlD TO ADD AN IMP CP A LINK D D- ELETE TO DELETE AH IMP DP A LINK E E VALUATE TD SEND THE PROPOSED CHANGES TC T E NETWORK' EVALUATION MOTEL G Q' To APPROVE AND IMFLEMENT A PROPOSED CHANGE FCLLDTNG A MODEL EVALUATION I KMP TO INITIATE AN IMP OFEPATICN SEC'UENCE L L INK' TO INITIATE A LINK CPEPAT DN SEQUENCE N N OGO TO REJECT AND AEC-'T A PROPOSED CHANGE FOLLOWING A MODEL EVALUATION R R EPORT TD CALL P ONE OF THF TEN REPORTS REPORTS TRAFFIC DEMAND NETWORK TOPOLOGY IMPS INVENTCPY NETWORK PERFORMANCE ECONOMIC PEPOPTS NETWORK SUMMARY T T IME OLD NEW 1 3 5 Z 4 6 8 BOTH 7 10 TO DRING THE TRAFFIC DEMAND MATRIX 1 P TD ANY USER SPECIFIED TIME THE USER MILL ALSO EE PROMPTED FOP -YES AND 'NO ANSWERS AND FOP A iV DP 1 ' AiNSWER IF AN INPUT ERROR IS MALE THE USER WILL USUALLY BE PUT BACK INTO THE COMMAND MODE FOLLOWING AN ERROR MESSAGE F-8 1 APPENDICES ON FACILITIES ECONOMICS ISSUES -- PREFACE Appendices G through 3 The following four appendices were prepared by Ronald C Crane to describe a cost model structure for estimating the costs involved in the ARPANET They provide a do-it-yourself kit of tools and a data base to allow the user to consider any combination of elements that are in place at any point in time producing output analyses under a wide set of depreciation assumptions and costing bases F-1 u AppendixG ARPANET INVENTORY LISTING PROGRAM RONA C CRANE PREFACE This BASIC program is used to provide a listing of the facilities being considered in the financial analysis of the ARPANET The description of the forms of the output and the use of this program is contained in Appendix H following While written for the HP2C00F system this program can be modified to run on other systems on line in the ARPANET if required Or arrangements can be made to have access to these programs via the timesharing system used here G-i ARPANET INVENTORY LISTING PROGRAM FUNCTION This program lists all of t sites in the network and the equipment at each site General and development facilities are listed at the end of the printout All the information comes from the data base contained in DATA-1 and DATA-2 programs and is subsequently stored by FILMAK in the file FDATA TO USE PROGRAM RONA To use the program to get a listing log on to the timesharing system-and type GET-RONA carriage return type RUN carriage return The program will then run and produce about forty-five pages of output If it does not run or if you are not certain that the file FDATA is up to date use FILMAK to reload the file The printout has dotted lines where the paper should be cut for eleven-inch pages using friction feed The pages may not be exact for terminals Page 1 Table of Contents is printed at the end of the output SPECIAL NOTES Parts of the output format require that the terminal have a backspace capability What effect this has on terminals not equipped with backspace is not known at this time G-l VARIABLES USED IN PROGRAM RONA A Date equipment was added number of months after 1-1970 B Date equipment was removed number of months after i-1970 C Cost type l monthly 2 investment 3 sunk non-recurring Cl Class listing flag C2 Site flag D l Number of day in year internal variable D 2 Number of year internal variable D 3 Dummy variable for data routine E Equipment number F$ 35 Equipment description 35 characters maximum G Site number of equipment H$ 35 Site name 10 characters maximum I Connected site number used in leased lines only equals 0 otherwise 1 1 Month equipment installed 1 2 Year equipment installed 1 3 Month equipment removed 1 4 Year equipment removed 1 5 Month number of starting month 1 6 Year of starting month Hi High equipment class number of range being examined Ll Line count on page L2 Low equipment class number of range being examined M$ Month of present date P Page count number Ml Margin spacing left hand margin S Site number being examined FILES USED IN THE PROGRAM RONA File #1 FDATA - Semipermanent contains data base for ARPANET File #2 FILE2 - Temporary each sice File #3 FILE3 - Temporary used for accumulating the Table of Contents as program progresses through data base used for accumulating the data on G-2 POMfi 10 E'O 30 I- Tlr-i o o -4 IHT 'TIM' 3' '4 3 7 TM c fi THEM 40 4L 50 EUR0 70 30 90 1 u0 110 130 1 30 14 0 15 0 16 0 170 13 0 190 E00 10 '0 30 4 0 50 EE 0 7 0 30 3 i i C 31 0 3 0 34 0 s o l_i L 33 0 D f j TIM' 5 - 1300 ri j rif'' -' IF DC1J 31 THEM 30 Dill r CU D C3J Ier r C1J 59 Tw N 50 Ic I'Cl 9 0 THEN 3 0 IF Drn -o l u THEM 310 IF liCll IM THEM 34 0 IF DC1 131 ri-'EM 370 IF TU oo 1 THEM 400 IF LCI 43 THEM 430 IF UE13 73 THEM 4E0 1- Dili ' - 304 THEM 4c fi I DE 13 - 3 34 THEM 530 1 1 -DEI -334 M I'tCEMSEF ' 5GTC3 55 0 M JaNUSP Y' GGTQ 550 Lii L 11-31 f -- FEF' JI P 5G7G 55 0 It 11 D 13-53 r' r'V f cH 33TJ 55 0 DE 13-Ml 3-90 M i rf PlL ill UTG ID fl D 11-59 ri3 T'-MY 3DTO 55 0 Pm T-EI3-151 33 u __ _ p 400 41 0 4 0 430 44 0 450 4 t 0 4 70 4 0 4 - 0 c - - iun DtiJ-isi M- JULY 53 Tu 55 0 D i rtn- i fi' j'jiT' 35TE 57 0 C11 I' 1 - 4 3 M3 rPTti '7 s f ' 3'jTD 55 0 L lJ I Cn- 73 r 'i--- DcTatef 51 0 5 0 er -i 5 J- 0 G37G 55 0 L 13 M 11-304 Mr '''' -- 7HT pn M ti5J t t EMI' GF Tor v' r L- 1' r- L o V'' oo o T I TI F M 9 E IM G-3 ' ECTIUM - o o' PRINT PPT NT PRINT FT INT FPiriT PPI T FPI r PRINT 3 31j r ' 0 r 'r ' 700 no 73 0 U 4 ij ' j 0 3 U L IN-19' L1H o CriLELMTh H'-F c r ET Llr 5 Mi o PEM PEN I o 1 ' r DNTH IHrTflLtEI' I 3 MrjMTH REMOVED REM I 5 I'FiTE OF MCMTH ' ' 0 I 31 7 0 I l rt IC5 I 3 INT- I 1 13 I L6J ni I l -l INT' lCll l K3J E IC5 o 3 0 IC4 IHT-CI C3J oo1 ' IC6 4 ij IC3J-I C3 -i3 ar r- i -13 -15 0 '3 0 c70 73 0 3 0 300 31 0 330 3 0 IF I 11 0 THEN 37 0 SQTL 39 0 1090 1100 111 0 1 ISO 113 0 1140 Ir ' trironv i L 1 'o L' 31 FEH o ITEM I' TE SECTION 'SUE-POUTINE ' # 73 0 ' 0 0 310 1 03 0 107 0 1 03 0 INC PPIHT LIN'S - Ll 4 FEM # END OF TITLE PPGE FILE Fr'hTH FILE FILE3 P 3 DIM FIC35 H'f C353 GO TUB 3310 I 51 0 35 0 93 0 r1 0 3 3 i ' 3r0 1 000 1 Oil 103 0 113 0 104 0 1050 FH 3CC I hTE i_ Iri o 31 o ij r4j H 1 ' 3 - Bf P I '-4 VSA 1 I o 3' ' I N JTrtLLEP Rp MO'-'Efc ' EPS OF MONTH 1 im i3 I l '3-1 IF I C2J 0 GOTO 930 THEN 910 I 31 13 I 4 l 4 -1 PETUFN PEN ND CF ITEM DHTE SECTION F'EM MMP15IN SUEPOUTINE Ml 3 PRINT Pfl Ml RETUP CEM END CF MARSIN SUBROUTINE Pfius END SUBROUTINE PRINT LIN 63-L1 5 GQSUB 960 PRINT LAELEDftTP GAELSD TP flSSQClftTESi ASSOCIATES INC P-A'SE GOSUB 960 M M oo M$ DCI3 J BC 3 PRINT P P 1 PRINT LIM 1 J PRINT PRINT LIN 3 Ll 4 RETURN REN P Fflt E NUMIERJ1 15 SET TO 1 i lT THE BESINN 114 OF PPC3PPM REN END DF PftSE ENS SUEPQUTINE R EM LOAD uITE INVENTOPV INTD FILE 3 G-4 150 161j 17 0 ISO 190 I200 1210 1220 230 1240' 250 LCDO 127 0 I2i0 290 300 310 320 330 34 0 35 0 36 0 37 0 38 0 390 400 410 420 14 30 44 0 450 L460 147 0 143 0 490 5 CO 510 52 0 53 0 54 0 550 56 0 57 0 53 0 590 16 0 0 161 0 20 630 64 0 165 0 16 6 0 b7 0 80 169 0 700 71 0 20 P'EM rEM REM 'i- FFINT TOP D PAGE ALSO LaAI-S -ILE o- 3 THE TAFLE 3 SITE NUMIEP C2 SITE FLAG OF CONTENTS FIL FOP - 1 TD 93 C EM F E5ET FILES 1 2 REfiD 1 1 PERU 2 1 C2-0 IF TYP -I 3 THEN 1330 C 'ERD 1 H 6 C I' E P S SPH I IF G S THEN 1300 IF E 100 PUD E 150 THEN 1280 GOTO 1290 IF I S THEN 1300 5DTH 1230 PRINT t2 H E C ri E F jG Hi I END C l UDTQ 123 0 f Efl - ENI' OF THIS SUESECTIDN TD LCHII FILE 2 IF C2 0 THEN 150 0 PERD 2J1 REfiD 2'H E C D E Ft tG HJ I IF E 150 hND E 100 THEN 136U GOSUE 960 PRINT SITE Na - G HS PPIHT LIN 1 ' GD3UF 960 PPINT SITE INVENTORY HE I' C13 D C21 PRINT LIN'2 ' GCiUE 96 0 PRINT Er'JiFMEHT CLASS PRINT ThE-Nl l ' I L 1 11 PRINT 3 G-H' P END GESUE 224 0 NEXT 3 RETURN ' END 3F LOHE ANU RAGE TOP SECTION o '' ' CLASS LISTING iUEPDuTINE # L1 L1 1 C1 0 PEN cl FLAG INDICATING NO ITEMS IN SPDUP FDP THIS SUBPCUTINE REftD 2 1 IF T'iP'2 3 3 THEN 2 09 0 REfiD '-'C ' hi E C D E F'l- GJ Hi I IF E L 2 HND E H1 THEN 162 0 GOTO 156 0 GO SUE 960 GCi '3 0 IF E - 100 HND E 150 THEN 1660 GOTO i 0 IF I THEN 169 0 PPINT THE 6 M 1 LINE TD HI TRE'2S 11' I GDTC 172 0 PRINT TRE 6 fU LINE TD Ft THE 2S M1 G5 GDTO 172 0 PPIHT THF'6 m F J i PRINT LINING 1730 ThB' 34 M o - I Cll TRE C36 tfl I C2 G-5 J TRE '' 42 Nl 3D 'o- c h If B 999 THFN 177 0 r-cINT l 'l If i - 173O I ES THf ' 4J H1 o I T41 IMHGE 0 30 70 1730 1790 1' 0 0 i 3 1 i GC1TC1 173 0 PRINT o PPIflT Tf ipi 51 Kt ' ' 5 F'FInT IJIINI 1330 D 1 00 11 0 1 0 13 0 140 150 IF D 1000 THEN 1330 IF D'1 E 0G THEN 1350 PR IMT USING '330 INHGE o DrXDDD DDD GOTO 133 0 PRINT USING 1370 IMrtGF i o IMAGE ' ' IF C l THEN 1910 IP 3 3 THEN 194 0 IF 3 3 THF n 197 0 PRINT USING 19 0 IMAGE oMONTH GDTD 03 0 PPINT USING 1950 IMAGE INVEiTNENT GOTO 0 30 IF G 99 THEM 0 00 PRINT USING 0 0 GOTO 03 0 PPINT USING 010 IMAGE TQT L IMRGE NDN-REHUP L1 L1 1 31 1 IF LI 3 0 THEN 07 0 GDTD 03 0 GOSUB 1000 GDTD 1530 IF 31 0 THEN 110 GOTO 14 0 GDSUB 930 PRINT L1 L1 1 PRINT LINa L1 L1 1 16 0 17 0 19 0 190 00 10 0 30 PEN RESET FILE FOINTER READ 1 IF HI 1000 THEN 0 IF LI 55 THEN 10 GDTD 0 GDSUB 1000 RETURN REN END DF Ct_ ss 40 5 0 60 70 30 90 PEN '' C ss TITLE M CaNTFaiGDSUB c0 PRINT lOO LEASED LINES L 100 Hl 150 PRINT LIN 1 GDSUB 930 1 - o 'i 1 330 1 34 0 1350 133 0 l 370 1 - 5 0 1 3 0 1 j f 191 0 19 0 1 '-' ' - ' r 4 0 1950 1 93 0 197 0 198 0 199 0 000 E 01 0 303 0 0 30 04 0 050 06 0 07 0 090 090 3 0 0 -- NONE G-6 -- LISTING SUEPGUTINE SECTION 5310 5350 5 330 5 34 0 5350 5360 5370 5330 5390 54 00 5410 5450 PPIhT USING 5350 I' RSE LINE 11 CDNMECT 8D TE I II TE PR IMT US IMG 34 0 IMtiGE 5 COST TYPE GJS'JB 9c 0 PRIf T USING 37 0 INHGE 53K- 1 SITE Na BfiDPEC IPEMover 1 PRINT LirUDi Ll Ll 4 3DSUE 15 3 0 GOSUB 960 PRINT 15 0 MOPEKS 54 3 0 HI 130 544 0 545 0 L5 150 GSSLC 153C 546 0 5470 54 SO 5490 5500 5510 5550 5530 554 0 GC3UB 960 PRINT 180 - OBEN INTER-RCES H1-300 L5 130 6USUE 1530 6 5 SUE 960 5560 5570 555 0 5590 56 0 0 5610 365 0 5630 564 0 5650 2t A0 567 0 565 0 569 0 57 0 0 57 0 573 0 74 0 575 0 576 0 c i f U 573 0 379 0 58 0 0 5810 535 0 - o- - n 4U 50 ' 6 0 '0 j 5 0 Hi 300 L5 500 GUiUB 153 0 GUSuB 36 0 PRINT 300 TIP Ml 400 L5-300 G03UB 153 0 i r t 9 n PPIHT 4 00 LOCAL TELEPHONE LINES HI 45 0 L5 400 GOSuB 1530 i D' PI I PRINT 45 0 L DCBL TELEPHONE rfDUEMS HI 500 LS 45 0 30SUB 155 0 GOSUB 96 0 PRINT 500 TIP TERMINALS H -t'OO L8-50 0 L-O - UB V6 0 PRINT 600 H i r ij o HOST INTERFACE ' L5 60 0 00SUE 1550 SOSUB 96 0 PRINT 7 00 HOST MRCHIhE ' HI 50 0 i i I L5 iVi ip icr -i i G G IU B 9 6 0 PRINT 800 HI - 9 0 0 - - LDCRL FflClLMlE G-7 L8 30 0 G03UF 1530 GG 3 L'x 960 rF Hi f 900 H1 1 0 0 0 Lc 900 GOSUt 1530 G3 L t 1' '0 0 ETU ri LGCHL K- M EN ' r P CLAEE fWi 6 M - iT - TITLE AND CONTROL fECTIGN o ppf GENE- L FACILITIES LISTING SUS-ROUTINE o o ' I I I PEN 01 0 r 'EHl' THE -E PoCILITIE APE LISTED UNI'HP SITE NUHFEB 99 8 1 PEFiD -1 1 GG U3' 960 I u 4 0 3 US u 306 0 5 i 7 0 3 03 0 PRINT INRGE 8 - 31 0 0 31 1 0 31 3 0 3 4 0 31 0 31 6 0 31 r o 318 0 3190 3c 00 3c 10 C-ENEBf L FIND BEI ELCPMENT FACILITIES PRINT LIH- Ll S GOSUE 960 PRINT USING 3090 FAC ILIT-- 18 '0 PRINT USING 3 180 G0 3UE 96 0 IMHGE 5 COST PRINT USING 314 0 IMHGE 34x EftMseri PRINT LIN a j GOSUB 960 PRIMT 0 00 Ll 18 PRINT DI ITE U DATE 0 TYPE SREMOVED 1 DEVELOPMENT FACILITIES J3'0 GENERAL oo o DEI ELOFMENT FACILITIES J J 510 51' 0 53U 54 0 5 0 56 0 5 0 IF TYP 1 o THFH 386 0 READ 1 Rt f C B E Ft G H'i I IF 6 99 THEN 384 0 GOTO 3800 PR I NT 2 A B C B E F'l - G Hi - I END GOTO 3800 HI 50 L2 l GOSUE 1530 GOSUE 960 PRINT USING 3310 IMRGE 050 GENEPAL FACILITIES L2 50 Hl 100 GOSUE 153 0 GOSUE 1000 RETURN PEN END OF GENERAL FACILITIES 33 0 390 4 00 410 480 43 0 44 0 450 46 0 REN TABLE OF PEN FILE 3 is A TEMPORARY REHD C3 1 P l PRINT LIN 1 DIN B4E353 PRINT PRINT LIM 5 PRINT 4M 5 0 6 0 ' 7 0 80 9 0 o C 0 END LISTING ioiPuUTiNi CONTENTS SUIPOUTINE o FILE ALREADY LOADED WITH THE TAFLE IN I HRRHNET INVENTORY TAELE OF CONTENTS G-8 J47l 4 o i Ij 0 510 -30 3540 5550 3560 3570 353 0 3590 3600 3610 36c 0 5630 364 0 PRIMT LIM'3' PRINT ThB''15 SITE ND TAB 2S' 5 5ITE M VME PPIhT LIM 11 13 TftB'-53 ' RASE MC IF TYPO ' 3 THEM 3530 C'ERD 3 i t Bi C PPIMT THB 15 ft TRB 0 ' E3 TRB 60 C L1 L1 1 IF LK60 THEM 3510 303 UB 1 COO GOTO 5510 GDSUB 1000 EMI C3F Ti ELE DF CnWTHMTS SOSUE 1140 GOSUB 3990 GQSUB 333 0 END G--9 UTINE S'JBPQ' Appendix ARPANET INVENTORY bY EDNALD CRANE PREFACE This report is prepared using the program RONA described in Appendix G preceding This inventory listing shows the facilities that form the Communications part of the ARPANET The program is capable of producing snap-shots of the network at any point in time In some instances the data needed was not readily available so estimates were made However the writer believes that these listings provide a reasonably accurate statement of the components of the communications-related components of the netv ork adequate for economic analysis purposes In line with the limits of the present study this inventory is restricted to the computer-communications portions of the ARPANET and purpasely does not show or include the various host machines This program has been written to irclude such facilities at a later date should it be desirable to do so H-i - o i - - fiPPfiNET INI-ENTOWY TABLE OF CONTENTS SITE Mo 1 3 4 5 6 7 8 9 10 11 1 13 14 15 1 17 18 19 0 1 i_t_ 4 5 6 O 7 L 1 CO 9 30 31 Z C 33 34 o-7 o-'1 0 SITE PAGE MO NAME 4 6 7 8 9 10 11 1 13 14 15 16 17 13 0 1 3 4 UCLfl SRI UCSB UTAH EEM IMP 5 MIT RftND SBC HFlPl fiPI LINCOLN LFIES STRNFDRI IIIINIos CFISE CFIPNESIE MELLON UNIV PftDLIJPENN AMES TIP MITRE TIP RfiBC MBS ETfiT TINKER McCLELLflN USC TIP 6WC NDftT SBRT BELVOIR flPFfl TIP RBERBEEN BBH TIP CCRT XEROX FMWC TIP LftwPfNCE BEPK LFIB UCSB H UF II TIP RMLT GENEPFIL DEVELOPMENT CMlLHir-iTfi HrFPCIfiTE I MC S t F T F M rep t 197 3 t- PS 30 31 - -- oZ'C tj J 34 35 3A 40 41 4 43 F'fiC-E H-l 1 SITE HO ITE INIENTDPY UCLfi SEFTE-I EP 26 1973 EPUIFMENT Cl_ ASS LEASED L r ES o 100 0 CONNECT IIDATE 0 DATE 3 SITE NO BHDDED BREMOVED a LINE LIHE TD SDC LINE TG UCSD LINE TD UCSB 35 3 4-71 '-73 10-70 150 r MODEMS MODEM TO IJCSE 9-70 MODEM TO UCSD 9-70 MODEM TO SDC 9-70 COST I TYPE o29 -'MONTH 643 o'MONTH 459 -MONTH s $ 425 o'MONTH 425 -MONTH 425 MONTH $ s 180 MODEM INTERFACES - -- Mor-EM INTERFACE UCSB MODEM INTERFACE UCSD MODEM INTERFACE SDC 9-70 9-70 t $ $ 5 000 INVESTMENT 5 000 INVESTMENT 5 000 INVESTMENT IMP 9-70 $ 45 000 INVESTMENT 9-70 9-70 9 70 9-70 9-70 9-70 $ $ S S $ S 20 000 20 000 20 000 12 000 12 000 12-000 9 70 300 yips NONE 4 00 LDCAL TELEPHONE LINES NONE 450 LOCAL TELEPHONE MODEMS NONE 100 TIP TERMINALS NONE 00 HOST-IMP HOST-1MP HOST-IMP HOST-IMP HOST-IMP HOST-IMP 700 HDST INTERFACE INTERFACE SFTWR INT SFTWR 360 '91 INT SFTWR PDP-10 HDWR INT SIG 7 HDWR INT 360--91 HDWR INT PDP-10 HOST MACHINE NONE CAELEDATA ASSOCIATES INC 197 iEPTEMtEP 26 PAGE H-2 NDN-PECUP NDN-PECUP NDN-PECUP INVESTMENT INVESTMENT INVESTMENT 300 LOCAL FACILITIES LOCAL IMP M INT 9-70 420 'MONTH 900 LOCAL MANAGEMENT NONE CiiiLEi'flTh Aitsncirnt - SEPTEMBER o - '6 INC 1973 pi C-e H-3 SITE ND HE INVENTORY' SPI 26 SEPTEMHEP 1973 ECIUIFMENT CLRSS I i 1 On LEASED LIMES B CONNECT CDHTE t D TE H B SITE NO-CADDED EPEMOI ED B LINE COST TYP 12-72 10-72 8-72 $ $ % 229 -'MONTH 86 -'MONTH 3b 'MONTH 10-70 10-70 10-70 T t $ 425 -MONTH 425 -MONTH 425 'MONTH ISO MODEM INTERFACES - MODEM INTEPFP CE K-PfiRC 10-70 MODEM INTERFACE AMES 10-70 MODEM INTERFACE LBL 10-70 $ 'I 5 5000 INVESTMENT 5 000 INVESTMENT 5 000 INVESTMENT 10-70 S 45 000 INVESTMENT 600 HDST INTERFACE HOST-IMP INT SFTWR 10-70 HDST-IMP I NT SFTWR 10-70 oHOST-IMP HDWR INT 1 0 -70 HOST-IMP HDWR INT 10-70 S $ $ l 20 000 20 0 00 12 000 12 000 LINE TD LEL LIME TD KPDX LIME TD AMST 34 32 16 150 - MODEMS MODEM TO XEPOX-PPPC MODEM TO LEL MODEM TO AMES IMP 300 TIPS NONE 400 LDCAL TELEPHONE LINES NONE 450 LDCAL TELEPHONE MODEMS' NONE 500 TIP TERMINALS - NONE NON-PECUP NON-RECUP INVESTMENT INVESTMENT 700 HDST MACHINE NONE 800 LDCAL FACILITIES LDCftL IMP MftlNT 10-70 CftEt Eriis TFi SEPTEMBER A i iJL - ASSOCIATES 26 INC 1973 420 'MONTH i PI I' E H-4 900 LOCAL MfiMflGEWEHT -- NONE C i LEIERT fisSP'-IflTES I JC ScrTEMFER 5 S j 1 373 1 FR'SE I H-5 --l M SITE NO 3 SITE INI-ENTOPY UCSB SEFTEMEEP 6 197 EQUIPMENT CLRS LEASED LINES 10 0 B LINE CONNECT SDMTE 3 DATE 0 COST TYPE 9 SITE NO EflDDED BPEMDVEX' 5 10-70 11-72 I S 459 MONTH 1 147 -'MONTH 150 MOSEMS MODEM TO UCLfl 11-70 MODEM TO FNWC 11-70 I f 425 ''MONTH 425 --MONTH LINE TU UCLfl LINE TD FHWT I 00 190 MODEM INTERFACES 11-70 11-70 MODEM INTERFACE MODEM INTEPFP CE IMP 31 ' 0 3 T 5 000 INVESTMENT 5 000 INVESTMENT 45 000 INVESTMENT 'i-70 TIPS NONE 4 00 LOCAL TELEPHONE LINES NONE 450 LOCAL TELEPHONE MODEMS NONE 500 TIP TERMINALS NONE 600 HOST INTERFACE HOST-IMP HDNP INT 11-70 HOST SOFTWARE MDD 11-70 $ $ 12f000 INVESTMENT 201 000 NON-PECUP 700 HOST MACHINE NONE 800 LOCAL FACILITIES LDCPL IMP MRINT 11-70 900 420 MONTH LOCAL MANAGEMENT NONE CftBLEDPTfi AS-rOCIFlTES I NC SEPTEMEEP 6 ' 1973 - Pf E H-6 SITE UTAH 4 SITE ND INVENTOPY EURo SEPTEMBER 1973 EQUIPMENT CLASS 100 LEASED LINES I CONNECT EDATE LINE B DATE B CDST TVPE 8 SITE NO EAt'I'EI CPEMD 'Et P 12-71 12-72 f $ 6 597 ''MONTH 3 442 -'MONTH 150 MODEMS ' Mnr EM TD LEL 12-70 MOSEM TO ILL 12-70 I 425 VMONTH 425 MONTH ISO MODEM INTERFACES Mnr-EM INTERFACE 12-70 MJXJEM INTERFACE 12-70 $ $ 5 000 INVESTMENT 5 000 INVESTMENT J 45 000 INVESTMENT 12 34 LINE TG ILL LIME TO LEL IMPS 00 12-70 IMP TIPS 00 NONE 400 - LOCAL TELEPHONE LINES o NONE 450 LOCAL TELEPHONE MDDEMS NONE 500 TIP TERMINALS NDNE f 00 HOST INTERFACE PDP-1C Hr NR INTERFACE lc'-70 PDF-10 SFTWR MDI' lc'-70 12 000 INVESTMENT 20 000 NQN-RECUP 700 unST MACHINE NONE 80 0 LOCAL FACILITIES LOCAL IMF MRiNT 12-70 420 MONTH $ 900 LOCAL MANAGEMENT NONE CAri EIiATA A UCIflTES SEFTEMPEF o - I iJh 26 s IM 1973 FASE H-7 5 SITE MD SITE INI ENTOPV BUM IMP 5 SEPTEMEEP C6 1973 EQUIPMENT Cl_A S I 100 sasKSMs LEASED LIMES G CONNECT DDATE B DATE D 0 SITE NO BABBEI BREMDVED H LINE LINE TD HRRV LIME TD CCflT 9 31 6-71 150 MODEMS MODEM TD CCfl 4-71 MODEM TO HARVARD 4-71 COST I I TVPE oJ CT I I -oMONTH -oMONTH 4 5 -'MONTH 425 'MONTH 130 MODEM IMTERFRCES 4-71 4-71 oI 5 0 0 0 INVESTMENT 5 000 INVESTMENT 4-71 $ 45 000 INVESTMENT 600 HOST INTERFACE 4-71 HDWR INT 4 MACHINES SFTHR MOD 4 MACHINES 4-71 S oI 43 000 INVESTMENT 30 000 MON-RECUR MODEM INTERFACE MODEM INTERFACE c'00 IMPS IMP 30 0 TIPS NONE 400 LOCAL TELEPHONE LIMES NOME 450 LOCAL TELEPHONE MODEMS NONE 500 TIP TERMINALS HONE 00 HDST MACHINE NONE 300 LOCAL FACILITIES LDCAL IMP MAINTENANCE 4-71 430 -'MDf'TH 900 LOCAL MANAGEMENT NONE CADLEDATA ASSOCIATES INC SEPTEMBER 6 1973 PA E H-8 6 SITE MD SITE INVENTORY MIT 6 SEPTEMBER J 1973 EQUIPMENT CufiSS i 100 LEASED LIMES B S LINE CONNECT RDPTE 53 DATE SITE ND BHiii'Er' BREMDVED LIME TD CCflT LIME TD LL LIME TD ILL 3 1 O'- C 10 12 5-71 1 -71 150 MODEMS MDTJEM TD CCfl 6-71 MOSEM TO LL 6-71 MDDEM TO ILL 6-71 R D COST TVPE % $ $ 57 MONTH 114 -'MONTH 5 450 -'MONTH 1 425 'MONTH 4 5 -MONTH $ 4 5 y-MOMTH 180 u MODEM INTERFACES 5J000 INVESTMENT 10 5000 INVESTMENT 6-71 6-71 MODEM INTERFACE 2 MODEM INTERFACES IMPS 00 6-71 3 45 000 IN'-'ESTMENT 600 HOST INTERFACE 4 HDWR INTERFACES 6-71 4 SFTWP MOD 6-71 S $ SOJOOO IMP TIPS 300 HDNE 40 0 LOCAL TELEFHONE LIMES HDNE 450 LDCAL TELEPHONE MODEMS HDNE TIP TERMINALS 500 NONE 43 000 INVESTMENT NDf -RECUR 700 r HOST MACHINE NONE 800 LDCRL FACILITIES - LacftL IMP 42 0 -'MONTH 6-71 MAINT 90 0 LOCAL MANAGEMENT NONE CMILEEAYA SEFTEMITR ASSOCIATE'S 6 INC 1373 PAGE H-9 SITE Na 7 SITE INVENTORY PAND SEFTEMSEP 26 t 1973 EPUIFMENT CLASS I I 100 LEASED LINES 0 CONNECT EDATE 0 DATE H 8 SITE NO CADDED SREMQVED B LINE OS LINE TG ISI LINE TO UCSD 35 ISO MODEM MODEM TD UCSD 4-71 MODEM TO ISI 4-71 COST CC 7 'I $ 425 o-MONTH 425 -'MONTH 10 000 INVESTMENT IMPS IMP 0 0 - -'MONTH 860 -MONTH $ oI 130 MODEM INTERFACES MODEM INTERFACES 2 4-71 200 TYPE 45 000 INVESTMENT 4-71 -- TIPS NOME 4fiO LOCAL TELEFHONE LINES NONE 450 LOCAL TELEPHONE MODEMS NONE 00 TIP TERMINALS NONE 600 HOST INTERFACE HDHP INTERFACE 360 65 4-71 SFTNP MOD 360 '65 4-71 I S 2 000 INVESTMENT 20 000 NON-RECUP 700 HOST MACHINE NONE Q00 LOCAL FACILITIES LOCAL IMP MA INT 4-71 420 -'MONTH 900 LOCAL MANAGEMENT NONE CAJLEDATA ASSOCIATE INC SEPTEMBER 26 i973 FA'3E H--10 8 SITE HO SITE INVENTORY SBC 26 SEPTEMBER 1973 EQUIPMENT CLftSS i 100 LEASED LINES 1 B LINE CONNECT oI'STE 0 DATE SITE NO Hflr riEri REMOVED COST 3 B TYRE 4-72 4-71 % $ 'MONTH 9 -'MONTH 150 MODEMS Mor EM TD USC 4-71 MODEM TO UCLA 4-71 % 3 4 5 vMONTH 4 5 MONTH 130 MODEM INTERFACES MODEM INTERFACES 4-71 % 10 000 INVEST 200 IMPS - TM IMP 4-71 % 45 0 0 0 INVEST % $ 12 000 INVESTMENT 0 0 00 NON-F'ECUP LIME TO USCT LINE TO UCLA 3 1 300 TIPS NONE 400 LOCAL TELEPHONE LIMES NDNE 450 -- LOCAL TELEPHUt-F MDDEMS NONE 500 TIP TERMINALS NONE too HOST IMTEPFACE 1 HDNR INTERFACE 360 145 4-71 1 SFTNR MOD 360 145 4-71 700 -- HUST MACHINE HOME SOD LOCAL FACILITIES LOCAL IMP MRINT - -71 4 0 'MONTH 900 -- LOCAL MANAGEMENT - NONE CfiFUni'ftTfi A - OC lH f'ES SEPTEMBER 6 lN l' 73 PfvSE H-ll 9 SITE NO SITE INVENTORY H PI oPP SEPTEMBER 6 19 3 EQUIPMENT Ci_ASS i irifi LEASED LIMES D 9 LINE CONNECT EDATE B DATE SITE NO GRrir-Er EREMOVED COST TYPE oI 151 'MONTH 57 'MONTH 150 MODEMS MODEM TO BEN 5 6-71 MODEM TO HERD 6-71 I 4 5 oMONTH 425 'MONTH ISO MODEM INTERFACES MODEM INTERFACES 6-71 $ 9 5 LINE IU fiBRD LIHE TD BEN 10 0 00 INVESTMENT IMPS 00 4 5 000 IN VEST M ENT 6-71 IMP 300 6-71 TIPS NONE 400 LOCAL TELEPHONE LINES NONE 450 LOCAL TELEPHONE MODEMS NONE 500 TIP TERMINALS NONE 600 HOST INTERFACE 3 HDHR INTERFACES 6-71 3 SFTNR MOD 6-71 % $ 36 000 INVESTMENT 60 000 NON-PECUP 700 HOST MACHINE - NONE 300 LDCAL FACILITIES LOCAL IMP MAINT 6-71 $ 4 0 -MONTH 900 LOCAL MANAGEMENT NONE CAELEI'ATA ASSOCIATES INC SEPTEMBER 6 1973 fc PA' E H-12 SITE SITE I' 1 Ha INVENTDPY LINCOLN L BS 6 SEPTEMEE 1973 EQUIPMENT C'LfiSS I 100 LEASED LINES B B LINE CONNECT BDOTE B D TE SITE Na BRr-rtEi BREMDCEP LINE TD MIT 6 5-71 E C COST 3 150 MODEMS - 5-71 MODEM TD MIT 5-71 MODEM TD RflDT TYPE 114 -MONTH 4 5 'MONTH 425 MONTH 180 KODE INTERFACES 5-71 5-71 5 000 INVESTMENT 5i000 INVESTMENT 5-71 45 000 INVESTMENT 600 HOST INTERFACE 3 HDWP INTERFACES 5-71 3 SFTWP Mar 5-71 36 000 INVESTMENT 60 000 NON-RECUR MODEM INTFJ-TACE MODEM IMTL'- FACE 200 - - 1 - IMP IMPS TIPS 300 NONE 4 00 LOCAL TELEPHONE LINES -- NONE -- 450 LOCAL TELEPHONE MODEMS NONE 500 -- - TIP TERMINALS NONE 700 HOST MACHINE NONE 800 T LOCAL FACILITIES LOCBL IMP MM I NT 5 71 $ 4 0 -'MONTH 900 LOCAL MANAGEMENT - NONE Ct-'iFLEDftTft Hi -DC IiiTC- S - r- T r M F A EP c ' 6 I NC FhC-E fl-13 t 9 3 MM MI SlTE No 11 SITE INVENTORY EruiPiiEN STANFORD DEFTEMEER 26 1973 CLPSS i i 100 LEASED LIMES B CONNECT GDATE S DATE H SITE NO OADDED SPEMD ED 0 LINE LINE TD AMES LINE TD ISI COST TYPE 3-72 S $ 86 oMONTH 1 790 -MONTH 7-71 7-71 S S 425 -'MONTH 425 -MONTH ij 150 MODFrtS MODEM TO MODEM TO AMES ISI 180 MODEM INTERFACES MODEM INTERFACES $ 10 000 INVESTMENT 200 IMPS '-71 IMP $ 45 000 INVESTMENT t $ 125000 INVESTMENT 20 000 NON-PECUP oI 420 ''MONTH JfiO TIPS NONE 4 00 LOCAL TELEPHONE LINES NONE 450 LOCAL TELEPHONE MODEMS NONE 500 TIP TERMINALS NONE 600 HOST INTERFACE PDP-10 Hr-wp INTERFACE 7-71 PDP-10 SFTHR MOD 7-71 700 HOST MACHINE oNONE S00 LOCAL FACILITIES ' LOCAL IMP MAINT 7-71 900 LOCAL MANAGEMENT NONE FASE CABLE-DATA ASSOCIATES' IN SEPTEMBER 26 1973 ' - - im--m H-14 m SITE NO 12 SITE INVENTORY ILLINIOS SEPTEMIE 26 1973 EQUIPMENT CLASS i 100 LEASED LINES B CONNECT BDATE B DATE H 6 SITE NO BAI'SEI BREHOI E D B LINE COST TYPE 12-71 12-71 5 450 -'MONTH 6 597 'MONTH 150 MDDEMS MDLEM TO UTfiH 12-71 Mar-EM TO MIT 12-71 425 '-'MONTH 425 -'MONTH LINE TD MIT LINE TD UTAH 6 4 ISO MODEM INTERFACES 2 MODEM INTERFACES 12-71 10 000 INVESTMENT 00 - IMPS --IMP 12-71 45 000 INVESTMENT 300 TIPS NONE 400 LOCAL TELEPHONE LINES NDNE 450 LDCflL TELEPHONE MODEMS NONE 500 TIP TERMINALS NONE 600 HOST INTERFACE PDP-11 Hr-NP INTERFACE 12-71 PDP-11 SFTNR Mar 12-71 lStOOO INVESTMENT 20 000 NON-RECUP 700 HOST MACHINE NDNE 300 LOCAL FACILITIES - LDCAL IMP M INT 12-71 420 oMONTH 90 0 LDCRL MRNAbEMENT NONE CAILEUATA ASSOCIATES INC SEPTEMBER 26 1 73 ' o FftfE H-15 13 SITE NO SITE INVENTORY CASE 19 6 SEPTEMJ-EP EPUIPMENT CLASS 100 LEASED LINES B B LINE CONNECT SITE B BD TE DATE NO ERDDED EPEMOI ED COST B H TYPE 11-71 10-71 4 7 oI f 42 -MONTH 003 -'MONTH 5 450 MONTH 150 MODEMS ' MODEM TO 6WCT 10-71 MODEM TD RADT 10-71 MODEM TO CMU 10-71 o1 S $ 4S5 -MONTH 425 -MONTH 180 MODEM INTERFACES 3 MODEM INTEPPACES 10-71 $ 14 15 4 LINE TD CMU LINE TD PRUT LINE TO GWCT 15-000 INVESTMENT IMPS S00 IMP 45 000 INVESTMENT 10-71 NONE 400 LOCAL TELEPHONE LINES NONE -- 450 LOCAL TELEPHONE MODEMS NONE 500 TIP TERMINALS NONE 600 HOST INTERFACE PDP-10 HDMP INT 10-71 PDP-10 SFTWK MOD 10-71 S I lc'jOOfi IN'-'ESTMFNT NON-PECUP E'OuOOO 700 HDST MACHINE NONE S00 LOCAL FACILITIES LncfiL IMP MfiiNT 10-71 4 0 -MONTH 900 LOCAL MANAGEMENT NONE Cft LEi T SEPTEMBER - 1 M ASSOCIATES 6 IN 1973 FA E H-16 SITE Ma 14 SITE IN' -ENTDPY CAPNESIE MELLON UNIV 6 SEPTEMBER 1973 EQUIPMENT CLASS 100 LEfiSED LIMES B CONNECT HDATE 3 DATE B SITE NO EHI'DEI BPEMOt'E B LINE COST TYPE 1 1 434 'MONTH 643 -MONTH 150 MODEMS Mar-En TO CASE 11-71 1 MODEM TO PEL ' 11-71 $ $ 4 5 'MONTH 4 5 -MONTH ISO MODEM IMTEPFflCES MODEM INTERFACES 11-71 $ 10 000 INVESTMENT 11-71 oI 45'0 00 INVESTMENT 600 HOST INTERFACE 3 HDNR INTERFACES 11-71 3 SFTMP MDI 11-71 $ $ 36 0 00 INVESTMENT 6 0 0 0 0 NDN-PECUP 7 13 LIME TD EELV LIME TD CfiSE 6-72 11-71 T IMPS 00 IMP NONE 400 LDCfiL TELEPHONE LINES NONE 450 LDCfiL TELEPHONE MODEMS NONE 500 Tip TERMINALS - NONE 700 - HOST MACHINE NONE 800 LDCfiL FfiCILITIES LOCAL 900 420 ooMONTH 11-71 IMP HA INT LDCfiL MANAGEMENT NONE C A 7 L E r A T f - f o A TLMIITP P V C5 Z I A T E S ' C J I fi 3 97 5 EASE H-17 15 SITE NO JITE INVENTORY PADLIJPENN SEPTEMBER 26 1973 EpiJIFMENT CLASS I i 100 LEASED LINES B H LINE CONNECT EDATE 0 DATE SITE NO 3ADDED PPEMOI-ED LINE TQ STAN 11 3-72 B 0 COST T- CC $ 86 -'MOM TU E 850 -MOr-n-J 425 Mnr T-j 4 5 -MGN-J 425 oMDrrr -- 15 0 MODEMS 2 MODEMS TD ftHVHHERE 2-72 MODEM TO AMES TIP MODEM TO HAWTIP MODEM TO STANFORD 8-72 8-72 8-72 8-72 $ l X 1 3 0 MDBEM INTERFACES 2 MODEM INTERFACES 2-72 3 MODEM INTERFACES 8-72 8-72 $ $ 10 000 INV--7 Fr-T 15 000 INVF T NT 00 IMPS IMP 2-72 IMP 8-72 8-72 $ i 45-000 INVECr-- NT 45 J 000 INVESTMENT 8-72 S-72 X X X X 12 000 HIVE-T NT 20 000 NON-cc' - 4 000 IMVF f NT 40 000 N0N-- -' ' -v o 3-72 X X 420 -MONTH 420 MONTH NONE 40 0 LOCAL TELEPHONE LINES NONE 450 LDCAL TELEPHONE MODEMS NONE 500 -- TIP TERMINALS -- NONE --r' v v ' - 600 HOST INTERFACE 1 HDNP INTEPFACE 2-72 1 SFTWR MDD 2-72 HDWR INTERFACES 8-72 2 SFTWR MDD 3-72 700 HDST MACHINE NONE SOO LGCAL FACILITIES LOCAL IMP MAINT 2-72 LOCAL IMP MAINT 3-72 CAELEDATA ASSOCIATES INC SSPTEMJ EP 26 1373 fc L4 Uk 3ft PA- E H-18 900 LOCAL MfiMftGEHEHT NOUS CrttLEPftTft ftrSOCIftTESs EF TEMI'fiP INC 4 l'- ' ' JLMM M MMM I P 5E tf-19 lt SITE MO STTE INVENTORY- AMES TIP 6 SEFTEMIEP 1973 EPUIFMENT CLMIS I 100 LEASED LIMES LlME 0 CONNECT Jl'fiTE G G DftTE SITE ND MADDED LPEMOCED Cc T G TYPE G 8-72 oI 86 -MONTH 150 MODEMS MODEM TD AMES 8-72 MODEM TC SRI 8-72 f I 425 -MONTH 425 -MONTH 180 rr MODEM INTERFACES MODEM INTEPFPCES 8-72 S 10 000 INVESTMENT 8-72 I 92 000 INVESTMENT 600 HOST INTERFACE 1 HDWP INTERFACE 8-72 1 SFTWf MoD 3t0''67 8-72 l t 12 f 000 IN'-'ESTMFNT 20'OTiO NOH-RECUR LIME TG SRI 2 200 IMPS NONE TIP 400 LOCAL TELEPHONE LIMES NONE 450 LOCAL TELEPHDME MODEMS NONE 500 TIP TEPMIMALS NONE 700 HOST MACHINE NONE 30 0 LOCAL FACILITIES o Locf ii_ TIP MfliNT 8-72 535 --MONTH 900 LOCAL MANAGEMENT NONE Ci iiLEDi iTri ASSOCIATES SEPTEMBER - n A_ 26 INC eft-SE 197 m H-20 17 SITE NO SITE INVENTORY MITPE TIP 26 SEPTEMBER 1973 EQUIPMENT CLRSS 100 LEASED LINES I CONNECT ED TE LINE B SITE H DRTE ND BRBDED BREMDVED B COST T VPE B 6-72 5-72 $ 86 MONTH 172 'MONTH 150 r MODEMS MODEM TO flPpfi T 4-71 MODEM TO SDBT 4-71 oI t 425 -'MONTH 425 'MONTH ISO MODEM INTERFACES 2 MODEM INTERFACES 4-71 $ 10 000 INVESTMENT 3 92 000 INVESTMENT '3 26 LINE TD RPPT LINE TD SDAT 00 t r IMPS NONE 300 -- TIPS o TIP 4-71 4 00 LOCRL TELEPHONE LINES NONE 450 LOCAL TELEPHONE MDDEMS NONE 500 TIP TERMINALS NONE 600 HOST INTERFACE NONE 700 HOST MACHINE NONE 800 LOCAL FACILITIES LOCAL TIP MMIMT 4-71 oMONTH 900 LOCRL MANAGEMENT NONE C i i i Et t M T A A - s VJ C I P T E s SEPTEMBER 26 I N o 1973 PfiC'E H-21 IS SITE NO SITE INVENTORY PADC SEPTEMBER c't- ' 1373 EQUIPMENT CLASS i i 100 LEASED LINES B 0 LINE CONNECT EDATE R DATE SITE NO El AIDED REMOVED 3 Q CDST TYPE 10-71 t ' 003 -MONTH 150 MODEMS MODEM TO CASE 10-71 MODEM TO LL 10-71 $ $ 435 -'MONTH 435 -MONTH ISO MODEM INTERFACES 2 MODEM INTERFACES 10-71 % 10-000 INVESTMENT X 2 n00 INVESTMENT LINE TO CASE 13 00 IMPS - o NONE 300 TIPS TIP 10-71 4 00 - LDCfiL TELEPHONE LINES NONE 450 LOCAL TELEPHONE MODEMS NONE 500 TIP TERMINALS NONE 600 HOST INTERFACE NONE 700 HOST MACHINE NONE Poo LOCAL FACILITIES LOCAL TIP MAINT MONTH 10-71 900 LOCAL MANAGEMENT NONE CAI-LEDATO ASSOCIATES INC SEPTEMBER 26 o 1973 -1 PA-SE H-22 19 SITE NO SITE INVENTORY UBS SEPTEMEEP 26 o 1973 ECI'JIFMENT CLASS I 100 LEASED LIMES B G LINE CONNECT HDMTE SITE NQ BHr-rEr B DATE EREHOVEX COST B S Tvp- 7-72 t 87 oMONTH 150 nnDEtis - 11-71 natEURM TO AEP D 11-71 MDEEM TQ ETFlT I 425 'MONTH 425 oMONTH LINE TO flDPD 9 I ISO MODEM INTERFACES 2 MOI'EM IMTERPflCES 11 71 $ 105000 INVESTMENT 11-71 t 92 000 INVESTMENT 11-71 S $ 12 000 INVESTMENT 0 000 NOH-RECUP NONE UP 400 LOCAL TELEPHONE LINES NDNC 450 LOCfiL TELEPHONE MODEMS NONE 500 TIP TERMINALS NONE 600 HGST INTERFACE 1 HriWR IHTERFflCE 1 SFTNR MOD PDP-11 11-71 700 HOST MACHINE NONE gOO LOCAL FACILITIES LOCRL TIP UM I NT 585 MONTH 11 71 900 LOCRL MANAGEMENT NONE i tl'MTfi A -C1 TES' i I NC S EFTC M SI H 2 t 1973 CM L ' JL PI IGE H 23 20 SITE HO SITE INVENTORY ETAT 6 SEPTEMBER 1973 EQUIPMENT CLASS 100 LEASED LIMES B CONNECT ODATE 0 SITE ND-EADSED LINE LINE TO PMLT LINE TD APPT LIME TO RRPT I DATE 8 EREMDVED H COST TYPE o-o o Cr _o % 3 8 3-72 6-72 I 4 539 o'MONTH 143 MONTH 36 o'MONTH 150 MODEMS MODEM TO MBS 3-7 MODEM TO RMLT 3-7 I 425 MONTH 425 o'MONTH ISO MODEM INTERFACES Mot'EM INTERFACES 10 000 INVESTMENT 3-72 NONE TIP 3-72 S 92 000 INVESTMENT 400 LOCAL TELEPHONE LINES NONE 450 LOCAL TELEPHONE MODEMS ND J oo TIP TERMINALS NDN 600 HOST INTERFACE NONE 700 HOST MACHINE NONE S00 LOCAL FACILITIES LOCAL TIP MAINT 900 LOCAL MANAGEMENT NONE CftBLCfflTR SEPTEMEEP M A ASSOCIATES 26 INC 1973 PA3E H 24 21 SITE ND SITE INI ENTOPY TJMKtK 6 SEPTEMSEP 1973 EQUIPMENT Cl_i iSS I 100 LEASED LIHE H CONK-CT HD TE P DF TE u 0 SITE r'n BfiDt'EC EREMD 'zU 2 LINE COST TYPE NONE 150 -- MDDEMS --o MODEM TD 2-72 425 -MONTH 425 vMONTH 2-73 $ c-tc c-3 i- 2 72 -73 $ 10 000 INVESTMENT 00 IMPS IMP 2-72 -73 S 45 000 INVESTMENT 2-73 2-73 $ $ 12 000 INVESTMENT 20 000 NON-PECUR 2-73 'I 420 -'MONTH MODEM TD i' o 1Q0 MODEM INTERFACES 2 Mct'EM INTERFACES 300 TIPS - NOME 4 00 - LOCAL TELEPHONE LINES - NONE 450 LDCflL TELEPHONE MODEMS NONE 500 TIP TERMINALS HONE 600 HOST INTERFACE - 1 Ht'lJP INTERFACE 2-72 1 SFTNR MOD 2-72 700 rr KOST MACHINE NONE 800 LOCAL FACILITIES LDCBL IMP MA INT 2-72 900 --_- LOCAL MANAGEMENT NONE C' J-LCI'MTR ASSOCIATES SEF'TEMI EP 26 INC 1973 PFtSE H-25 i SITE SITE Na 22 II-H ENTHRY McCLELLRN SEPTEMBER 1973 26 EraUIF-MENT CLASS 100 LERSED LIMES H LINE CONNECT BBATE B LINE TD STAN LINE TD RAND LIME TO RMLT 11 7 37 oI 8-73 2-73 TTFE H $ $ 3-72 3-78 -73 150 - MODEMS Mai'EM TD ' 3-78 MODEM TO 3-72 MODEM TD PML 2-73 MODEM TD RAND 2-73 MODEM TD STAN 2-73 COST DATE 3 SITE ND BRDDED FJpEMQt'ED l $ oi S 1 790 oMONTH 229 MONTH 13 424 MONTH 425 425 425 425 425 -MONTH MONTH -oMONTH MONTH -MONTH 130 MODEM INTERFACES - 3-72 2-73 00 IMPS IMP 3-72 IMP 2-73 -73 SMODEM INTERFACES -- NONE 10 000 INVESTMENT S 45 000 INVESTMENT 45 000 INVESTMENT -- 400 LOCAL TELEPHONE LINES NONE 450 LOCAL TELEPHONE MODEMS NONE 500 TIP TERMINALS HONE 600 HOST INTERFACE 1 HDNR INTERFACE 3-72 1 SFTNR MOD 3-72 PDP-10 HDUR INT 2-73 PUP-10 SFTNR MDD 2-73 2-73 2-73 $ t 12 000 20 000 12 000 20 000 700 HOST MACHINE NONE CABLEDATA ASSOCIATES INC SEPTEMBER 86 5 973 t L jJh i FASE H-26 INVESTMENT NDN-PECl'P INVESTMENT NON-RECUP SCO sssssass LUV HL mULIll Locfti- IMP MftlNT LDCfiC IMP MfilNT 3-7 2-73 -73 ---- $ $ 420 oMONTH 4 0 'MONTH 900 LDCflL MANAGEMENT NONE CftVLEVftTft Rs DCI lTES INC PftGE H-27 A I C I IU o fc_ -_- SITE IHI ENTOP EQUIPMENT 26 SEFTEMBEP 1973 CLASS 100 LEASED LINES LINE 8 CONNECT EDATE B DATE fi JITE BREMDVED NO BRDDKD 2-70 4-72 LIME TO HURT LIME TO SBC 150 e c COST i MODEMS 4-72 4-72 MODEM TU SDC MODEM TO NDHT 180 MODEM INTERFACES 2 MODEM INTEF- _ CES 4-7 t TTC-E 4 376 -oMONTH ooMDriTH -o C o- 425 o MONTH 425 -oMONTH oI 10 000 INVESTMENT S 92 00f INVESTMENT $ $ IS P00 INVESTMENT 20 000 NGN-RECUR NONE TIPS too TIF 4-7i 400 LOCAL TELEPHONE LINES NONE 450 LOCAL TELEPHONE MDDEMS NONE SfjO TIP TERMINALS NONE 600 HOST INTERFACE 1 HDNP INTERFACE 4-72 1 SFTWR MOD 360 44 4-72 7 00 HOST MACHINE NONE S00 LOCAL FACILITIES LOCAL TIP MAINT 4-72 900 LOCAL MANAGEMENT -- NONE -- CAXLLDATA SEPTEMBER - A Ar CCTFtTES 26 ItlC 1973 - PflC-E H-28 SITE INVENTORY 26 SEPTEMBER 1973 EQUIPMENT C'LfiSS 100 LEASED LINES B CONNECT BDATE LIME B DATE R COST TYPE E SITE NO EAl DEI BREMOI EE B LINE TD CASE LIHE TO NDAT 13 25 4-72 5-72 150 MODEMS MODEM TD CASE 4-72 MODEM TD NDAT 4-72 l $ F $ 5 450 oMONTH 1 721 'MONTH 425 -o'MONTH 425 -o'MONTH 190 MDDEM INTERFACES 2 MODEM INTERFACES 10 000 INVESTMENT 4 72 IMPS 00 NOME 00 TIPS TIP 4-72 S 92 000 INVESTMENT 40 0 LDCflL TELEPHONE LINES NOME 450 LOCAL TELEPHONE MODEMS 500 TIP TERMINALS NONE 600 HOST INTERFACE NONE 700 HOST MACHINE NONE 800 LOCAL FACILITIES A -- o LnCHL TIP MAI NT -MONTH 900 - LOCAL MANAGEMENT - NONE CAJ LET'ATA SEPTEMBER Ate J J M AssrciATi- s INC c'b 1973 mmmm FA E H-29 5ITE INI ENTDPY 6 SEFTEMEEP J 1973 EDUIFMENT CLASS 100 LEASED LINES B CONNECT BI TE E PP-TE 3 0 SITE Ma fcAime BPEMDVED B LINE LINE TO 6UCT LINE TO USCT 4 3 5-7 -70 150 MODEMS Mar-EM TD 6MCT 5-7 MODEM m USCT 5-7 COST TYFE J 1 731 'MONTH 4 876 'MONTH I 4 5 -'MONTH 4 5 'MONTH oI ISO MODEM INTEPFACES MODEM INTERFACES 5-72 10 000 INVESTMENT NONE 300 TIPS TIP 5-72 $ 92 000 INVESTMENT 400 LOCAL TELEPHONE LINES NONE 450 LDCFIL TELEPHONE MODEMS NONE 500 TIP TERMINALS NDNE 600 HOST INTERFACE NONE 700 HOST MACHINE NDNE Son LOCAL LOCKL FACILITIES T IP MftINT 5-72 -'MONTH 3 900 LOCAL MANAGEMENT NONE CoruEPriTfl SEPTEMBER i J mlh ASSOCIATES 6 J INC 1973 PI IGE H--30 c't- SITE NO SITE INVENTORY sum 26 JEPTEMIEP J 1973 EQUIPMENT CLASS 100 LEASED LIMES B B LINE CONNECT B BII TE NO Efln EP SITE DFITE BREMOI'ED CDST B B Tree 6-73 5-72 % % 237 -'MONTH 172 MONTH 150 MODEMS MOKEM TO BELV 5-72 MODEM TO NSAT 5-72 MODEM TD MTRT 5-72 % % % 425 VMftNTH 425 MONTH 425 MONTH 7 17 LINE TD EELV LINE TG MTRT 180 MODEM INTERFACES 3 MODEM INTERFACES 5-72 15J000 INVESTMENT IMPS 200 NONE TIP 92 000 INVESTMENT j c 400 LOCfiL TELEPHONE LINES NONE 450 LOCAL TELEPHONE MODEMS NDNE 500 TIP TERMINALS NONE g OO HOST INTERFACE 360-'44 HKNR INT 360''44 SFTMP MDI 5-72 t 5 12 000 IN'-'ESTIFNT 20 5 0 00 HON-PECUP 700 HOST MACHINE NONE 800 LOCfiL FACILITIES LDCRL TIP M I NT 900 LOCfiL MANAGEMENT NONE CfiFLEr'MTM H DCIi lTcS SEPTEMBER o i 26 ItK 1973 FfiffE H-31 -l ilTE no SITE INVENTORY 1'L UVU J SEFTEMEER 26 137 EPUIFMENT CLMSS I 100 LEASED LIMES CONNECT EII TE LINE 8 SITE 8 DFITE NO EADDED BSEMO ED B COST TYPE E 6-73 6-72 I oI l 430 -MONTH 287 'MONTH 1 434 'MONTH 150 -- MODEMS -- MODEM TO HERD 6-72 MODEM TO CMU 6-72 MODEM TO SDRT 6-72 $ oI 3 425 -MONTH 425 -'MONTH 425 'MONTH 29 26 14 LINE TO AERD LIME TO SDRT LIME TO CMU ISO MODEM INTERFACES 6 72 oI 15 000 INVESTMENT t - f d oI 45 J 0 00 INVESTMENT 600 HOST INTERFRCE CDC 6600 Hi'UR INT 6-72 CDC 6600 SFTHR MOD 6-72 I $ 12J000 INVESTMENT 20J000 NDN-RECUP 3 MODEM INTERFACES IMPS 200 IMP 300 TIPS - NONE 4 00 LDCRL TELEPHONE LIMES NONE 450 LDCRL TELEPHONE MODEMS NOME 500 TIP TERMINALS NONE 700 HDST MACHINE NDNE 300 LOCRL FACILITIES P LOCAL IM MftiNT 6-72 420 -MONTH 900 LOCAL MANAGEMENT NDNE C rLED Tft RiSaciftTES' IN SEFTEMFER 26 J 1973 -A- F SE H-3 28 SITE HO ITE INVENTORY HKKH I II- SEPTEMBER 26 1973 EQUIPMENT CLASS 100 LEASED LINES B CONNECT BDfiTE B DftTE I SITE NO EADDED SPEMOfED E LINE COST TYPE 3-72 6-72 6-72 t $ $ 14 3 MONTH 86 -'MONTH 86 'MONTH 150 MODEMS MODEM TO MTRT 6-72 MODEM TD ETfiT 6-72 $ $ 425 -MONTH 425 MONTH dO 0 17 LIME TD ETfiT LIME TD ETfiT LIHE TO MTRT 130 MODEM INTERFACES 2 MODEM INTERFACES 6-72 -- NONE 10 000 INVESTMENT -- 30 0 TIPS TIP 6-72 $ 92 000 INVESTMENT $ S ' 000 INVESTMENT 20 0 00 NON- ECUP 400 LDCflL TELEPHONE LINES NONE 450 LDCflL TELEPHONE MODEMS NONE TIP TERMINALS J NONE 600 HOST INTERFACE PEP-15 HDWP INT 6-72 PIiP-15 SFTNR MOD 6-72 70 0 HGST MACHINE NONE 800 LOCAL FACILITIES LOCRL TIP MftINT 585 MONTH 6-72 900 LOCAL MANAGEMENT NONE C Ft r L ' T M T M A s a c i B T E I N SEf'TnNi FR 26 J 1973 Jh FMCJE H-33 Na SITE 2 SlTE IHI-EWTDFY EF'JIFMENT ABERDEEN 26 SEFTEMFEP ' 1 973 C'UftSS i i 100 LEHNET LIMES LINE ' H CONNECT EDRTE E DOTE SITE NO Bfir-r-Ei- GPEMO' EE Co T i S 19 LINE TO NBST LINE TD HAPV LINE TD EELV 8 fi I T FE 87 -'MONTH 2 151 'MONTH 430 -MOMTH 150 finpEMr NONE lyij MDDEM INTERFACES 3 MODEM INTEPFfiCES t 15 000 INVESTMENT S 45 000 INVESTMENT S X 12 000 INVESTMENT 20 000 NON-PECUP IMPS c U LI IMP TIPS un NONE 4 00 LOCAL TELEPHONE LINES -- NONE 450 LOCAL TELEPHONE MODEMS HONE 500 TIP TERMINALS NONE 600 HOST INTERFACE PDP-11 HEMP INT PDP-11 SFTNR NDL 700 HOST MACHINE NONE 00 LOCAL FACILITIES LDCML 420 o'MONTH IMP MftlNT 900 LOCAL MANAGEMENT -- NONE CftTLXt oT SEPTEMBER - ASSOCIATES 26 INC 1973 ii PrtGE H-34 30 SITE- MD SITE INVENTORY EQUIPMENT BEN TIF SEPTEMBER 6 1973 CLRSS 100 LEASED LINES 8 CONNECT ED TE H DOTE H E SITE No-BRr-fEP BREMOVED B LINE COST TYPE NONE 150 rinnEMS NONE ISO MODEM INTERFACES NONE NONE 300 TIPS TIP 7-73 T 92j000 INVESTMENT 400 LOCAL TELEPHONE LINES NONE 450 LDCRL TELEPHONE MODEMS NONE 500 TIP TERMINALS NONE 600 - HOST INTERFACE NONE 700 -- HOST MACHINE NONE So LOCAL FACILITIES LacftL TIP MRINT 7-7 B BSsas 900 - LOCAL MANAGEMENT -- None C ft ELF T ' ft T ft A S OCIflTEft I N ' EFTEMtER '6 IS'73 - - - -1 Fft e H-35 SITE Mo 31 SITE INVENTOR EPUIPM NT U MI SEPTEMIER 26 1973 CLASS I i 100 LEASED LINES B CONNECT RDATE H SITE NO BADDED LINE 5 6 LIME TD BI N LINE TQ MIT B DATE EPEMOI-ED H COST oMONTH 57 ooMONTH - 1 8-72 150 MODEMS Mnr-EM TO MIT 6 8-72 MODEM TO EBH 8-72 I 180 MODEM INTERFACES 2 MODEM INTERFACES 2 MODEM INTERFACES 3 $ 00 - IMPS T 'PE 425 oMONTH 425 M0 1 H 10 000 INVESTMENT 10 000 INVESTMENT NONE 92 000 INVESTMENT TIP 400 LOCAL TELEPHONE LIMES NONE 45 0 LOCAL TELEPHONE MODEMS NONE 500 TIP TERMINALS NONE 600 - HOST INTERFACE PDP-10 HDWR INT PDP-10 SFTHR MOD % oI 12 000 INVESTMENT 20 000 MON-PECUR 700 HOST MACHINE NONE 800 LOCAL FACILITIES LOCAL TIP MA INT $ 585 -'MONTH 900 LOCAL MANAGEMENT NONE CAPLEDATA ASSOCIATES INC SEPTEMBER 26 1973 - o - A PA' E H-36 32 ITE NO SlTE Iut'ENTOPY XEROX SEPTEfSEP 26 1973 EQUIPMENT CLASS 100 -- LEASED LIMES S B LINE CONNECT BDATE i BATE SITE NO EADDED oREMQVED LINE TD SRI LI HE TD FNI IT 2 33 10-72 11-72 150 MODEMS MODEM TD SRI 10-72 MODEM TD FNMT 10-72 1 B CD-T TYPE I 86 -'MONTH 344 -'MONTH t t 425 -'MONTH 425 -'MONTH $ 180 MODEM INTERFACES NONE IMP 10-72 $ 45- 000 INVESTMENT 300 - - TIPS NONE 400 LOCAL TELEPHONE LINES NONE 450 - LDCAL TELEPHONE MODEMS NONE 500 TIP TERMINALS NONE 600 HOST INTERFACE 3 HDWP INTERFACES 10-72 3 SFTHP MOD 10-72 36 000 INVESTMENT 60 ooo NDN-PECI JP 700 HOST MACHINE NONE 800 LDCAL FACILITIES LDCKL IMP MPIHT 10-72 $ 42 - -'MONTH 900 LDCAL MANAGEMENT NONE CriVLEDATKi A SOC lfiTES If SCPTEMIEM e V j 1973 o -1 1 PM E H-37 33 SITE MD SITE INVENTORY FNMC TIPSEPTEMEEP 26 1973 EPUIPKENT CLASS i i 100 LEASED LINES B CONNECT BBATE V DOTE B 8 SITE Na-BHi't'Er- EFEMOVED Q LINE COST TYPE 11-72 11-72 I I 1 147 oMONTH 344 oMDNTH 150 MODEMS Mnr EM TO XROX 11-72 MDI EM TO UCSB 11-72 oI oI 425 MONTH 425 MONTH ISO MODEM INTERFACES 2 MODEM INTERFACES 11-72 f 3 32 LIME TD UCSB LIME TD XROX 10 i-O00 INVESTMENT NONE 11-72 92-0 00 INVESTMENT 600 HOST INTERFACE CDC 6500 HDWR INT 11-72 CDC 6500 SFTWR MOD 11-72 12K000 INVESTMENT 20 000 NON-P'ECUP TIP 400 LOCAL TELEPHONE LIMES NONE 450 LOCAL TELEPHONE MODEMS NONE 500 TIP TERMINALS NONE - o $ 700 HDST MACHINE NONE C no LOCAL FACILITIES LOCAL TiP MAINT 11-72 900 LOCAL MANAGEMENT NONE CABLEBATA ASSOCIATES INC SEPTEMBER 26 J 1973 - o -iW PASE H-38 ilTE I1U -T SITE INI'ENTOPY SEPTEMBER 26 o 1973 EQUIPMENT CLASS 100 LEASED LINES - LINE B CONNECT BDATE H DATE B B SITE ND Sfll'I'ED EPEMOI'EI- B 12-72 12-72 LIME ID SRI LINE TD UTAH 150 MODEMS Mat-EM TO UTAH 12-72 MOJCEM TO SRI 12-72 180 MODEM INTERFACES CDST TYPE otoo I MONTH 3 442 oMONTH $ % 425 V'MDNTH 423 'MONTH 1 12 72 % 10 000 INVESTMENT 00 -- - IMPS IMP 12-72 oI 45 000 INVESTMENT I oI 12 000 INVESTMENT 20 000 NDN-RECUP 2 MOt'EM INTERFACES 00 TIPS NONE 400 LOCAL TELEPHONE LINES NOME 450 LOCAL TELEPHONE MODEMS HONE 00 TIP TERMINALS NONE 600 - HOST INTERFACE CDC 760 0 Hi'MR INT 12-72 CDC 7600 SFTWP Mac 12-72 70i i - HOST MACHINE NONE SOG LOCAL FACILITIES LOCAL IMP MAINT 12-72 420 -'MDNTH or 00 -- LOCAL MflWBGEMENT NONE C'Al Lt'l'ATA ASSOCIATES I M St PTErir-ER 26 1973 A PAAE H-39 SITE INVENTORY 6 SEPTEMBER 19 EQUIPMENT CLASS 100 LEASED LINES R CONNECT BDATE L I HE n DOTE O COST TVFE B SITE NO HADDED BPEHOt'EU B 1 643 MONTH 360 MONTH 150 MODEMS MODEM TO UCLfl MODEM TO RflMD $ $ 425 -MONTH 425 MONTH ISO MODEM INTERFACES MODEM INTERFACES -73 S I 1 LIHE TO UCLfl LINE TG RAND o 00 10 000 INVESTMENT IMPS 45 000 INVESTMENT IMP NDMF 4 00 - LOCAL TELEPHONE LINES NONE 45 0 LOCAL TELEPHONE MODEMS NONE 00 TIP TERMINALS NONE -- 600 HOST INTERFACE E6700 HDMP INT -73 E6700 SFTWR MOD -73 $ $ 12 000 IN'-'FSTMFNT 20 000 NON-PECUR 700 HOST MACHINE NONE Soo LOCAL FACILITIES LOCAL IMP MMINT 420 MONTH 90 0 LOCAL MANAGEMENT NONE CAJELEDATA ASSOCIATES SCFTEMSEP p 1 J -W 26 IN 1973 FA5E H-40 SITE SITE Ha 36 IMI'ENTDPY H w ii SEPTEMIEP I IK 6 1973 EQUIPMENT CLPISS 100 LEASED LINES i CONNECT SDRTE S B TE B i SITE ND EflSDEC EPEMDV-'EI' 8 LIME CDST TVF F NONE 150 - MODEMS Mnr-EM TO AMES 12-7E 425 MONTH 130 MODEM INTERFACES lc - C S 5 000 INVESTMENT 300 - - TIPS TIP 1S-7I S 9ESOO0 INVESTMENT 1 MDt EM INTERFACE NONE 400 LOCAL TELEPHONE LINES NONE 450 LOCAL TELEPHONE MODEMS NONE 500 --- - TIP TERMINALS NONE 600 a- --- - ---- HDST INTEPFACE NONE 700 HOST MACHINE NONE o300 LOCAL FACILITIES LOCAL - llll 1 c'-- f c' TIP MPIINT i 535 o-MONTH LOCAL MANAGEMENT NDNE CiiJ Lf 'j'fiTt i A fjc IMT f - INC Sf-f-T Mt-Ei' Et i 1973 P ' E H-41 SITE INVENTORY EPTEHJFP 6 o 1973 EQUIPMENT CLASS IO1 1 LEASED LINES COST H CONNECT M'ATE H DOTE 0 H SITE NP-uJADIJED REMOVED U LINE TYPE -73 -73 13v4 4 --MONTH 4 589 -MONTH 150 - - MODEMS MODEM TO ISI -73 MODEM TD ETAT -73 425 -MONTH 4 5 -MONTH LI HE TG IS I LINE TD ETAT 0 ISO MODEM INTERFACES' MODEM INTERFACES -73 10 000 INVESTMENT S00 IMPS NONE 300 TIPS TIP -73 $ 9 i n00 INVESTMENT 400 LOCAL TELEPHONE LINES NONE 450 LOCAL TELEPHONE MODEMS NONE 500 TIP TERMINALS NONE 600 HOST INTERFACE NONE 700 HOST MACHINE NONE S00 LOCAL FACILITIES - LOCRL TIP MM INT 5S5 -MONTH -73 900 LOCAL MANAGEMENT NONE CMELEDMIM ASSOCIATES I r4 -- SEFTEMVEF 6 ' 1973 FI I- E H-42 OENEPftL ANI I'El-ELOFMEHT FACILITIES E BATE B BATE B Hflr-r-ED SPEMDI EI B FMCILITV 000 DEVELDFMENT FACILITIES - NETWORK ANALYSIS CnRF NfiC Eot_T EEPANEK t NEWMAN EEN 1-70 12-75 1-70 1-70 050 GENERAL FACILITIES BEN NETWCRK CONTROL CENTER 1-70 SRI NETWORK INFORMATION CENTER 1-70 MITPE 1-72 PAUSE MEASUREMENTS LAB CRNL 1-72 RRPfl UASHINSTON 1 70 UCLH NETWORK CAELEUATA RCfl FLORIDA MEASUREMENT 12-75 12-75 12-75 12-75 12-75 CENTER 1-7012- 4 1-73 1- 4 12-69 12- COST I 105000 -'MONTH $ 1 0 0 0 0 0 0 0 TOTAL I l % S 1 T $ i 50 000 oMONTH 43 000 oo'MONTH 10'0 00 MONTH 5 0 0 0 oMONTH 3 0 no F c-e 'MONTH 2 0 0 0 0 oMONTH 1 0 0 0 0 -MONTH 0 -MDNTH ' ' 'Ati tiifiTfi Mi nzT hTC5 o IN 197 SET-IENIf'R C TYPE H-43 Appendix I FINANCIAL ANALYSIS OF THE ARPANET PROGRAM 5 RONZ by RONALD C CRANE PREFACE These programs form the heart of the analysis framework used by Cabledata Associates to estimate the value of the ARPANET at any point in history under various costing and depreciation schedules and with any combination of elements added and or removed They use the same data base files used in Appendix H ARPANET INVENTORY by Ronald Crane The reader should review Appendix H to better understand what is and is not included in thj definition of the resources at any site For example at the start of this Cabledata Associates project for ARPA A'VA chose to have the study limited to the network per se lines IMPS TIPS modems etc and would not at this time include host machines and or other ARPA-owned facilities These programs are written in BASIC for the Hewlett-Packard 2000F timesharing system because access to this fast response system is only $3 00 hour including cpu during the hours when this work was done However it appears that these programs have a longer continuing applicability to run on an ARPANET computer I-i They could be altered easily PROGRAMS RONl RON2 FUNCTION These programs will calculate the book value of the network monthly cost in the requested month as well as per site values of the previous quantities for the ARPANET- Parts cf the network may be deleted and the cost of the network with and without these parts and the difference of these costs may be obtained Dif- ferent costing bases and depreciation rates and periods may be used HOW TO U Log on the computer system type GET-RON1 type RUN The program will take you from there It is helpful if the user has at hand the ''ARPANET Inventory Listing Program RONA CAWP# 107 Appendix J and a logical map of the network SPECIAL NOTES The depreciation rate applies to both the investments and the one-time charges such as lease set-up charges The scrap value is applied only to investments Non- recurring costs lease set-up charges all go to zero since they have no scrap value Month M is not charged for equipment bought and sold prior to month M It is assumed that the equipment was sold for its book value at the time the equipment was removed I--1 Up to 97 sites can be examined provided they are present in the data base file FDATA and the name file FILE4 Programs RONl and R0N2 work together to provide the financial analysis It is separated into two program blocks because the system storage capacity will not allow it to fit under one program name Any changes in either of the programs involving renumbering must include a change of the other program so that it chains to the proper place in the program being changed 1-2 VARIABLES USED A Date equipment added A l Marginal Cost option flag input by user of program A 2 Value of $ per month per site in first run A 3 Value of $ per month total in first run A 4 Value of Total investment to date in first run A 5 Value of Total depreciation to date in first run A 6 Value of Book value of the network in first run A 7 Value of Total investment in month M in first run A 8 Value of Average investment per site in month M in first run A 9 Value of New rent added in month M B Date equipment removed E l -B 98 Flags denoting that all the leased lines at a ncde are to be deleted B 99 The number of nodes with all lines deleted C Cost type l monthly 2 investment 3 non-recurring D Cost D l Day number D 2 Year number D 3 Leap year option E Equipment number F$ Equipment name G Site number H$ Site name I Connected Site number used for leased lines otherwise 0 K Type of cost l short term marginal 2 long term marginal 3 fully allocated L Margin constant horizontal positinn of output on page L l Depreciation rate specifier l straight line 2 sum of the digits 1 2 Depreciation period months L 3 Scrap value fraction of purchase price L 4 Dummy used to make non-recurring costs have 0 scrap value 1-3 L 5 Number of applicable depreciation months L$ 20 Dummy character variable used for entering answers to inj it routine M Month number of evaluation M l Internal dummy for month of evaluation calculation M 2 Year of month M M 3 Month # of month M M 4 Beginning year 1970 M 5 Beginning month # in beginning year 0 M$ 15 Character variable for month names N Cumulative investment through month M N l -N 97 Flags denoting that a node is to be deleted if the flag 1 N 98 The number of nodes to be deleted N 99 Node number corresponding to general development facilities 0 Cumulative rent in month M 01 Output variable for number subroutine P Cumulative depreciation of purchase Q Investment in month M Ql Instruction request flag Yes l No 0 R l -R 99 Flags indicating that the site was used in the calculations Rl Flag variable - if equal to 1 it denotes a change in the leased line rates R2 R4 Rate r Rate Rate R5 Rate for 1001 - 1500 miles R6 Rate for greater than 1500 miles R7 Dummy for leased line calculation section R8 Dummy for leased line calculation section S Number of sites used in the calculations R3 for 0 - 250 miles - r r i for 251 - 500 miles _ __ - for 501 - 1000 miles Leased Line rate changes otherwise assumed to he _ _ _ a flat rate ofc $5 83 per mile per month S l -5 99 Flags indicating that the site was used in the ca i -ulations T Total of one-time ncn-recurring costs lease set-up costs U New non-recurring or lease set-up LSU costt in month M 1-4 V New rent in month M W Amortization of investments and non-recurring costs applicable to month M Y Dummy variable used for DO-LOOPS Z Dummy variable used for input routines FILES USED IN PONl File #1 FDATA Contains the data base It is created by another program FILMAK frcm the primary data base programs DATA-1 and DATA-2 File #2 FILE4 Contains a list of the names of the nodes in order It is created by the program FILER with user interaction 1-5 o jni 4 H o t ' o o f L U ' ' E ' L -o' o j - h' ' I - I r L L Lei o L L jJ ' i - 'l o f'i j o ' f f o j D ' - i r ' ' o o ' 1 ' ' i - i o _ 1 R i i ' 4 '' ' r '_ L '1' s - o - n REM 1- Th' Y ri-Z iTi-- T r--PNT r UM5f r I M l z '0 r i- U' -f-- ri I 7O 0 M f 4 L 15 FTL S FDi T 'FILE-i Ic END 1 THEt-i 13 10 FC V l TU 9 - - - T Y u-'l o EMD LIF I'HTM 411 i r 1 - 0 K MT Da ocj MMfJT iMtTPucnari ' V M ' o I r - U T LS I - L'I lJ v IHEH 3 03 IF Lltn Y' THEM 00 IF LLL1J- N THEM 33 0 IF L'$Cn H THEM 30 FT-'Iff INPUT MUST P 5IN UTH Y CP N FLEH E FYTC 130 J M IJ 01 1 c o 0 3 0 VJTO 30 n i j - I I THZr 1 'o'-F r0 '9 0 f I PE TF- o y-JU 3 0 '' - 0 xECTTCii v - ' - o oo' - - IN THE rERI' 1J v THIS Fi'1 J TH FOIi-T UHtPF TH C-'V- CL'L HT I ONS ThfJT PaQ' fiH 1 IF T - H ES'SINNIN'3 DF THi2 Cfl'_CULflTIOf4 LOOP IF TYP- l 3 THEM 1 I Q ' - ri D 1 h D C o D J E F i h H i 1 F HtlJ l THEN 410 I- M 031 1 THEN 1190 I'- t 150 nf'-I1 E 100 f'HEM 360 'V7TG 410 ' 5 U V-0 4U0 410 44 0 430 44 0 450 46 0 470 4 SO IF BCS1 1 THEM 1190 if BCI 1 THEM 1190 IF R1-1 THEM 4 00 bJTD 410 FGYUE 433 0 IF E 100 THEM 4 3 0 i QID 44 0 IF K 3 THEM 1190 PEM IS THE ITEM IM THE FUTURE WITH RESPECT TD MONTH M IF H -M THEM 1190 IF C l THEM 520 Ic OS THEM 520 IF C 3 THEM 520 1-6 49 f 500 51 0 5S 0 530 -0 Jt I ' e rri o 0 0 E1 0 6 c '0 to i 0 FPIN T iHEPE IS Sfl P E IN THE IMFUT I'riTP C l OR 3 -P I NT U11NS 51C AoB C B'E Ft '3 HI I IM EE St lz o oo i o EI J o o i ' K I' o 1ER v rrr-' - o w f4T CALCULATIONS PEM U L- Pfc iT V NEW PENT IM MQNTH M E EQTH DP ABOVE AP CUMULATI'-'E G'-EF ALL ELEMENTS SEAFCHED IF o THEN 57 0 EdTG 65 0 IF M THEM USO IF --' H'EN 610 ECTO 116 0 V V ti 6 60 o-' o 1 ri 'E-iTM I T CALCULATIONS c En h RLL INVESTMENT Q NcM INVESTMENT IN MONTH N FE 1 BOTH OF ABC'-E R-E CUMULATIVE GVEP ALL ELEMENTS SEARCHED IF C-S THEN 67 0 SETS 7S0 6 0 N if- 8I IF - 'i THEN 0 0 64 0 SS 0 r i p 69 0 700 FEM RN 4 IJ QNE TIME NGN PECGVEFRILE COSTS o LEA it SE T UM COSTS CLSUJ oSI 7i T HLL LS'J U MEL' LSU IN f'iOfiTH M BOTH OF THE REEVE RFE CUMULATIVE OVER HLL ELEMENTS SEPFChSD 75 0 17 ' rn o 77 o 730 79 0 ESTE EE0 T TH IF R M THEN GOTO L' EO X I 0 IJ I THEN 770 POO c 'i ' o MONTHLY RNOPT IZRTION -o o o cU FEN r EN M CUMULATIVE CHftPGES RPPLICRILE TO MONTH M _ ' ' 1 S4 0 S50 DeFK CIr--iTTOrj 1-STPF M-HT E-SUM CF THH r E 0 L ' E ' DEFP CIHTION ' o r 0 L' 'j i-CHnr o' H i U L'4 - DUMMT Ui-KC 91 0 9 ' 0 9 'o 0 94 0 - i i -j - i j - 0 1 000 5FECIPIEP Die-I i r PERIOD TIME ' F 7 T 1 CC T' 1 t H - l OF TO SCP'HF 1 'ftLUE FIJFCHASE FPICE HOM-PECuF I No COST HAVE 6 - I ' 0 0 Kft'S LI NE HEM REM L 5' IF - THcN IF E M THL- N LC53 M ft Ier LI5 i_ t 3 IF r-1 THEN S C - THEN L ' 1 - 0 E QIC ill 0 94 0 10E0 P ITEr n CPfi v'HL_L' NUMEEP O H -FL_I CHILE IEFFFCIHTIDN HHNTHF TOTfiL r-EF-P CI Tl3N OF HLL NVE TMt NTi TD MONTH M THEN 107 0 107 0 1 '00 i_ Lt J 1 1-7 1010 I i J IF LCI 5 THEM 104 0 i-t - - 1-L 4 -LC63 I 0 J TG 1 050 n - l-L C4J L L6 J -L TT i 66 rG i 07 0 o L Lc'J 1 ' --L 63 t i i 1 V'- o 107 0 10 0 1 0 30 1100 1110 1160 1130 1 140 1150 1160 1170 1160 11 ij 1600 1610 r C 3 THEN 1160 IF LC53 L 61 rnEH 114 0 IP LL13 c THEN 1 1 30 P P B LC5J ' 1-L C 31 o L 23 GOTO 1160 -PH 'l-L C33 ' oo1 'L 23 L 3 1 ' o I 61-L I5J ' '-L 1 L C63 J G6TC 1160 o-c 1_L J r-EPi o o o o o 1 I V EI USEP Ir4 THE ChLCULPTlGNS - - C-5 1 FEM SITES LXMl'lHED o 3 63 1 GOTO 69 0 EN r-r Tf-i CHLCULBTIOf-i o -EM ti 4 6GI-r-1 5 7E6P T LC 5 F'-H 1660 PEM 13 30 164u 125 0 1660 127 0 1260 126 0 1300 1310 136 0 13 30 13 0 F 1143 1970 M 53 U M T3 M MC53 SEM MG6'' 16 THE YEHF OF MONTH M M 63 lriT'M 13 o16' Mf43 2N M'6 ' I THE MONTH IN YEHP M Z OF MONTH M MT 3 f'Un-12 ' Ilfi -ML13 -16 ' If MC33--0 THEN 166 0 60T0 134 0 M 33 16 H 63 63-1 PEM hjME P cr sues CFILCULRTICM M' 5 MONTH 0 D NE MONTH FEFOFE EEGfNNlMb 135 0 136 0 13T0 133 0 1390 1400 1410 143 0 1430 144 0 3 0 FOP Y l TO 93 3 3 Y3 3 NEXT Y PEM END OF CfiLCULPTICN SECTION PEM OUTPUT SECTION # PPINT LIM 3 ' IF AC 13 2 THEN 1630 PPINT TRP'L U PPINT USING 145 0 1460 147 0 146 0 1465 143U PPINT TRIVD PPINT ARPANET COST ANALYSIS MODEL PRINT LIM a' 5 PEM CHAIN STRTEMENT CHAIN PQN6 o460 1500 1510 156 0 PPINT TPti-L' SPh' 60 ' M153PH'1 ' ii 13 PPINT LINa RERIi 1 1 1-8 J D 63 15 30 If 4 0 1 - SO 15 ij If 7 0 If - ij 15 90 it 0 0 It 10 It EU it 50 it 4 0 it 5 0 it t 0 1- 7 0 it 8 0 11 9 0 17 0 0 10 17 c'O - - J i 17 4 0 17 5 0 1 so 1 7 0 1 7 r ' 17 3 0 7 t 90 - - i - - o- IS 1 0 It 0 If 3 0 1950 199 0 1 - 7 U 1 3 '3 0 1 39 0 1 300 1 31 0 i ' Ij 1 -1 -1 ' 1 i -1 M 1 3 j 'J 1 3 r l i 1 3 P IJ 33f 3 J IJ 0 'j 0 01 0 1 i 1 - i c' I ' 3040 PFAr i H t c ro E F-J '5 H i PPInT TfiB L i PA-cO FT PRINT LIN' 1' i UHMfiPv DF INP UT S ' PPIrJT THt'L-' ppi'iT LINCI o M PRIMT TAS'L' Tr TH CF ELPLUMTJCN CM M DP M 33 - 5M FPINT TH o -L o C-DSTING EHSIS o 5 PRINT J IF K l THEH It50 IF r d THEH lb7 0 PRINT p -- FiJLLr At_LaCPTEI '-G'Z It 30 PR'I NT 3 -- Inn T TEF M MAP'5INI II_ GuTD 1630 FT I iT L -- Lor- VEPM i'V P'3iNf L PPJNT T '-'D 'I'SESEcisTinH PP-TE IF LC1 1 THEN 17 30 PRINT USING 1710 IMA9E YEP 3' 'M DF THE E'l'SITr 5 L IE oo'IE' 3 TO 175 0 PPINT USING 174 0 I't -rE vE 3 iTBAiaHT LINE PRINT TR3 f'L 5 IM' TM NT ICFPP VPUJE 5LC31 1 Rl 1 THEH 17 30 GOTO 1790 - P I -o o o CHAIN STATEMENT L'HA 3 H PEuL ' ' 34 0 JF A 11 1 THEN 1-10 GOTO 13E0 PRINT TRB'L 5 MARGINAL CO T CPTIDH INCLUDED FPI NT IN- 1' PC'INT IRB'L ' PEIHI LINING 3EO IF hCU 3 THEH 1970 GTS'I'D 1'3 0 PRIH7 LIM'E '5 P T MT TP ' L ' i F -1rJ j II3IN3 1310 PPI ir LUNl 5 IH'- GE o # i -p-It - '_ Cn T DPTICN - Car riPIrnN OF TWO FTC' 'I OUT I - A IJ - 1 THEM J 39 0 Ji 1 ' fHEH E 000 Gum EG50 P I HT TAB L o PPINT Ui-INi'- E0 30 PR ini i ih'i ' GEPO E050 PPIHT rHB' L - HP IHT U ' 1HG E040 hPJHT LIH' I o Iti-H'v- o _i_ ITEMS HPE INCLUK ' N THL- CUTFUT -o ' IMHAfc i-w i_E'Tti ITEMS R E MCIT TMCUUI-EI' IN TH I nuTfuT 1-9 D5T-r 050 060 070 030 090 100 110 l 0 E1 0 14' 150 16 0 17 0 l 0 190 00 10 0 0 4ij 5' - -i 70 IF 0 THEh 090 PPIUT THt L'i 0 Hat'ES NEPE USEt' IN THE CFlLCULftTICNS CHECK PRINT THE ' L ' I'MT E C 5' EVSLUiiTION nNJv-OP IEUETIGN OPTIONS ' 5DT2 5150 PF1MT THB'L 'I FEP MONTH PEP SITE 5 IF ML11 3 THEN 130 01 'o' oo 3 9CTC 140 D1 H - C -M' ' bG 'JP -05 0 PP'INT THI _''' ' i ER MONTH TOTPIL 5 IF 9 El 3 THEN 190 C1 0 W 9GTC 0 0 0 Ir 9 1 THEr1 10 00 10 0 l' 4 0 5 0 360 7 0 3 0 4 0 0 3GTG 3330 01 F 53-f EuI-uB 050 FK'INf T9L 'L' Lac ' I -PUJE CF TKF NETNOPK IF Htl THEH 70 Dl h-P 6CTG 33 0 Dl -9C6 - 'l -P'- GGGl'B 3 05 0 PFIMT TF1 ' ' TCTI IL INVESTMENT IN MI NTH Ier H 11 'HEN 43 0 90 D1 H 3J-Q-M GGI' fc 05 0 FF ITJT 19 L ' 5 TCTPL IF 9C1I-3 r--M 50 U1-1-GGTG 6 0 G - C4 -r GG L'E 050 PPTf'-T THB ' Lv 5 TCT-'_ INVESTMENT TD UFITE fcPfeci TiDM TC DATE 5 4 M 5 4 1 0 01 0' 4 0 4 0 9 ' 'T G 4 4 0 G 1 4 171-0' 44 0 450 C G 050 PFIHT T9F' L 1 9V'EPI E INVESTMENT FEP SITE IN MONTH M 46 0 47 0 43 0 IF RC13 3 DK' -i 9GTG 50 0 4-1 'j1 9 - oo i 5 j 0 9 u I J E 3 0 5 0 510 5 0 PFIHT T9B o _ ' MEN FENT RDTEB IN MONTH M IP nil 3 THEN 55 0 ' -TG - '000 57 0 THEN 430 - o 'o' 56 0 o Li 0 0' 0 IF ft 13 1 THEM 590 I-IO 5 5QTC c'i G 0 0 RC3J D 't i C -' A V - 0 10 0 'o34 0 650 6 60 67 0 E6S0 69 ' ' 70 0 7 10 7 30 3 74 0 750 77 0 7 30 790 0 0 310 -J 0 c 0 35 0 33 0 3 70 93 0 090 -00 31 0 9 0 I' 0 MO o- ''70 01 0 0 3 ij 0 3 'j 0 4 '05 0 T'r- 0 11 110 S 51 P ftt M-P P 71 0 R 6 0-i hC9 V I- til 1 1 THEM 700 IF HC1J THEN 7 0 I 27-J 0 HL11 _' J T u ci 0 HlD JTQ l 5u FEF P t1 PEM p-PM FEM SiTtS ENTMINET OUTPUT SECTION L - MfiPSir-J CONSTANT - D-UET II LINES DSL ETEI' Motes - E H-UNEJ VJtES - Pares user IN THE CALCULATIONS E -- n o-- F -- ' o -- -1 r T Lin' o PUTT T9 'L b SITE STATUS IN CALCULAT TONS ' PI 'FT LIM i l -i o- '- T T Mo UsEt DELETIONS 15s 3 i ra NANS -i -i T - i l TC 6 0 IF E'- J i-ci FPEM 3 0 0 PEHli L I- L'tCi'33 ENt THEM 3020 IF R CV i -- -1 THEM 910 97 TE 30 0 ' 5 THfc o -35 ' Pf I t-'T IF P ' L 1 o ' LJ5 T i 'Li-30 IF Yj ' FH M 9 0 OLF'L 9' i f R 11FT TFL rL 35'' 5 ir M Y 1 THl'H E TVj PP MT iPt 'L 4F H IF BCV3- 1 Th-i 3000 1 77 7 l 'i F'' I FT Thl ' L 'V3 ' 'L ' IMT o ri - T v SITES 9717 r-'EM DUTFUT hurn-EP o- t L'L'-TJt'l XI'T'JJ PP I MT i F- I I'F 3OF05 01 IF 71-11' h THEM 316 0 i-- Q1- l L uF PHEH 3140 R'FLIE PPIMT EXAMINE 313 0 ij IfiM 31 oO 1--11 i-ui ROUTINE OUTPUT SECTION TO 3170 'o -YE i- r i 71r -5 51TU 'u 1 1 7 0 f T HjPN -j cr r DF rjijrp-i 'T fvJ'Mrlf ' u PGUT T N r 7U 4-57 0 - 01 0 THEN 3490 o-'INT o # CF- F LEI f H Hi 7GC IFiTE-S o 10 oZ iO f--TNT c' i 0 F-1NT LIN 8 - INT THIS PFGG 'P i I-ILL CFLCULPT' - - lj _Z0 60 -1' 0 1 C I ' l ' 300 o l o 330 3 ' 0 34 0 3J0 I 3- 0 0 30 390 400 41 0 4c 0 i M t-t U 450 46 0 ' -t 480 4 0 500 51 0 5c 0 33 0 34 0 53 0 560 57 0 53 0 59 0 600 6i o 6 20 630 NT hPFpf-ET CGiT HNPLYSIS MCLEL MPr 7H F ITE FD- Tu3 r-L-r-lLc' D ITE E ' INEI- THE H ii'KCLE f CUTP'jri PF5 F IEB CK TRE LICHEE INFCFr-7 IG- cD'jr T iE NETVGF t o'PMi H SELECTED I'EPRECIHT ICTi RPTE HNfj TG71 IH'5 E'-i 13 H NODE MM-I EE DELETED r r flt-iswepmc TE QN TH- IELETICIN PECUEET FPCr- TH FPO' PPM LEM fD LINE COSTS Cr ' F _ 0 I E tELETED FOP THAT NODS IF TH LINES l C'JLIi NET F S THPCUE-i ' THS r QIE FDP M DIRECT CONNECTION UTH PMIIE MODEMS FCP THPiT N3IE WILL iE 'IELETEI- 1 r ifHEP DELETION DFTION Tl lP-2INi iL COET IHrCP ST ICN CONCERNING THE P DD I T I CNPL COST PEP NODE PND TO THE NET PS P WWQLE CPN T- E OI'TfllNEE BY 3H'IN5 fiN ftPFlPM TII E hNTWtP TO THE It-KC-IN 1 CC -T DPTI3N IN Cftl CUL T I M 5 THE MftPiriNfit COST DP P NODE DP NODES -nE r oac-P-vi FIB T DETERMINES U-MT THE NETUCAK CDE-TS P--S WITH THE rCttr I NCLL'D I FPINT TH PE5 'LTr ICZ TM CDFTJ NI TH THE NOTES LEFT C' 'T PPINTJ TH RESULTS rMD THEN 5t-J t Th-E t' I Fc PEtJC BETWEEN THE TNQ Ci LCL LPTIGr PS THE MfiPCINnL COST Cr THE NOIEs SELECTEC IN THE DELETION OPTIONS - INT -- INT -INT '-INT r INT F F I NT FFINT --INT - -INT ' - INT --'INT --oINT -- INT F- INT -PINT -PINT -- INT LIN 3' c-piNf PEPRECIftTIGN PHTE CG3T1N5 1-HiIS ND NGnTH G EVFLUHTICN r F'INT pPE STPNT'PRD WITH HLL DF THE OPTIONS HEt MHK UP THE t-PIHT FOPMEWOPK IN WHICH THE l HRIOL'S OPTIONS PF E CP LCULP TED THEPE FIRE TWD T' F-E OF r-EPFECIHTICN PfnTE fiVfl I LPiELE FFINT I - I TPI 'I'5HT LINE PRINT D - 3UM DF THE I'l ITE F INT PRINT DEFPECIFITIDM RATE o i DP D o INPUT Li IF Li CU t THEN 357 0 IF Li Cn D THEN 357 0 IF Li Cl - s THEN 359 0 IF Li Ci S THEt 339 0 PRINT INPUT MUST BEGIN WITH L OP SJ RETYPE GQTG 35 0 0 LC1 c GGTG 3600 L L13 1 IF 01 0 THEN 364 0 THE PEPRECIRTIDN FEWIOB IS THE TINE RFTEP WHICH THE PRINT PRINT INVESTMENT HriS DEPPECI fiTEE TO ITS SCRftP 'o'flLL'E SFECIFV THE r'EFRECI iTIC3N PERIOD IN TEFIRS PRINT 1-12 4 U PR 1 fir II iT LC E0 '- 3 7 0 '83 0 890 r ij o LL J 1 LC TF 01 0 ThEri i t 0 7 2 0 $74 0 ' t -' U 780 7 r o T 8 ij 79 0 ' 30 0 l o 3 3' 0 ' 30 ' 4 0 o50 30 0 - fi c P ' I r- T F'T'Ii-T PIN T Pc'IrlT PEP IGE YEARS 5 37g0 THE SCPPF i-'HLUE OF P J rWESTMENT IS 1 'ftLUE va WHICH IT DEF RE IMTES TC EY THE END DP THE I'EPPECIftTIDM PERIOD IT RETAINS THI l-'PLC'E FOR E ER AFTER THE SCPPF VALUE IJ FRACTITN CF THE OF ISINPL PURCHASE FRICE icppp VALUE ' DECIMAL FPRCTION ' 5 PRINT IM-UT LC3 IF Ol 0 THEM 5840 p ' o i_ IrJ v i F'R INT THERE PRE THREE PRINT FPIHT PR I NT PRINT FPI T '3 PWILMELE TYPES OP COSTING SP-IS iHORT TEFM MARC INfiL COST INCLUDES ONLY MDMTHL'i jHTflL COSTS DEPRECIATION AND DEVELOPMENT COSTS APE NOT INCLUDED - LDN TE M MA SINAL COST ' ONLY DEVELOPMENT COSTS MRE N i LiCTEIi - FULLY ALLOCATED COSTS ALL COSTS A E INCLUDED -' 7 ' L P''Ir-T F PPINT LIN' 1 PRINT CasriN I'M sis 'o 3 o L DP F IrFuT Li IF Li L1 ' S THEN 394 0 IP Li J3 oo THEN 3 34 0 IF L' Cn L- THEN 393 0 IF i_ l L THEM 3980 IF L'i L1J F Th- F4 393U I- L 13 F THEM 3330 Fpi''7 INPUT MUST BEGIN HITH 3 L DR F PLEASE RETYPE TL MO 5 0 ' 30 ' DEFPECX TICN 385 0 k' l 9DTD 389 0 K -' GDTD 399 0 fi IF 01 0 THEN 4 0 30 FPir LI' 1 J cc 'IrO - 8 T FPI i r i ii ' i ''i - L EI- 71 Tt'FE I'l THE NUr'FER OF THE MONTH IN WHICH THE NETWOFY n ES o'ILUFITEI' ON THE FCI L HINS LIME 'MONTH 1 1-1970 o MONTH IF 01-i Tr'7 f PRiNT LIN1 39 pr s f - ' H 4 -90 4 I ij I'I 4 1 0 -11 8 0 4 1 30 4 1 4 ij NUMSEP 1 DUE tELETICN OPTION p Ir- T LIN' 1 F'FlNT IF DU U iNT TC H i 'E M NODE OP NODES DELETED' ENTER THE J-HIMT N'jMf fi Dc NODuS TO DELETED FIRST ' 0 IF NONE rFt TO I E F'PINT DELETE' I ' AND THEN hCH EJF THE NODE NUMBERS PS PROMPTED 1 r F ' INT ' i DUP TEFMIMHL PFINT Hcu MANY MODES H F ro v DELETED IN UI HL931 i r- r- 98 3 -i T t l 4 3 7 0 4180 1-13 ' r 1 r - 1 i 0 iU 4 - Mj 4 4 3u - 4U 4 1 -t ' 00 10 -u o1 ' 0 4 340 4 u 4 r-lj ' I 4 E'O 4 -1' 44 ' 0 'o4 i0 4 0 44 1 1 44 4 0 o44 5 0 Jt -X C 0 44 ro 44 8 0 44 9 0 o- 0o 4 5 10 45 ZO 30 45 4 0 4 TO 45 60 4 i 7 0 45 -J 0 45 O l 4c 0 0 4- l n 4c 0 46 3 0 4- 4 0 46 5 0 4t 6 0 4c 7 0 4c S 0 4fc 90 47 0 0 ' THIS LGEF FPHrF- T -rn R PL 5 tuT'-T-Ihr USED IN n1-' - THE- o INT THE 3 'oo MT u IN 9 4i io v 'o -9E o DD ' ' i' l DP Y l LIP Y 3 1 THEN SD IP Y' Z O1 ' Y-Zci CP Y c THEN 430 J IF Y 3 UP Y Z3 DP ' - 5 GF Y 4- THEN FPINT TH 5 i r ra 4 3 30 fPlMT ST 9JTC 4330 FP'IMT tic 5 SD TO 4 330 rPIfST f D ' FPINT Noi'E TD IE DELETED IS CALCULATION l F n MOULT te FECTION i r t OF THF fHP 4 3E0 I -DUT Z N ZJ 1 EXT Y IF 01 0 T tU 1440 PPINT LIN' D PPINT LEASED PPINT THIS OPTION PF'INT FP'IMT PRINT PPINT LINE DELETION OPTION ALLOWS VOU TO PEMCUE I I_L_ THE LEM EI' LINES THAT APE CONNECTED TO 1 1 NOTE THIS IS USEFUL FOP DELETED NCISS NHUEE LINES LINES WOULD NOT It CONNECT ST I'SHT THPOU 3H IF THE NODE UZ P - NOT TMEPE HOW MANV NODES f PE THEPE WITH PLL THE LINES DELETED INPUT E 99 I E 99 0 THEN 46S0 PPM C 'EH g EM THIS LOOP SETS A FLfis user IN THE CALCULATION SECTION OP FPD'jPftM TO DETERMINE WHETHEP AN ITEM SHOULD BE USED FOP THE CALCULATIONS pgp Y l TG E 99 PPINT THE 5 PPINT USIN'3 45 J Y If'tHf-E DD IF Y l DP Y l DP Y 31 DP Y 4 DP Y 51 THEH 459 0 IP Y Z DP Y DP Y 32 DP Y-42 DP Y 52 THEN 4610 IF Y 3 OF ' E3 DP Y 3 3 DP Y 4 3 DP 'r 5 3 THEN 463 0 PPINT TH ' 9DTO 4640 PPINT ST 9DTD 464 0 PPINT MD GOTO 464 0 PPINT PE j PRINT NODE TO HAI-'E ALL LINES DELETED IS INPUT Z 23 1 NEXT Y IF Nt' - S 0 THEN 47 0 IF 99 0 THEN 4720 fi 13 0 1-14 THE 47 10 47 0 47 3 0 47 40 47 50 47 it 0 47 70 4 i 80 90 43 0 0 z 10 43 9 0 4t SO t 1 4 40 4 c 50 4360 4' 70 4830 4 8 90 49 0 0 -191 0 49c 0 4 33 0 GDTD 4S50 IF 01-0 THEN 474 0 F'h'lMT I'D YD WANT THE Mftfi IMflL COST CATION AS rEICPIFEI' PtDv'E Ch-IN PONE 50 bCT'J 940 95 TG 5150 r ti'' - LEASEE LINE CALCULATION SECT IDU # FE'- rE E0f'I' e TMH NiJMI EP Dr MILES IN THE LINK IY CI l 'I I' IM5 THE COST F H tv Tr-E FLiiT S TH GF 'I '03 PEP MILE PEP MCNTH r -r r 5 95 IF r 7- 950 0 THEri 4950 o 494 0 495 0 4960 497 0 499 0 4990 5000 01 0 5 09 0 5 030 5 04 0 5 05 0 5090 s u ii Ffi'I T ' MPP SINAL COST OPTION ' Y fl o 5 IM uT LI IF Li l v THEN 4 4 0 IF LSC1 Y ThEfl 4 40 IF Lim h THEN 4 0 IF L'I 1 THEM 4890 F'FINT INPUI M'-ST E'5IN HITH ft Y DR ft FLESSE F'ETYFE i -ZTG -1750 H L1 0 95 TG 4' 5 0 HtlJ l FEi o CHftlH STATEMENT ' l o o C- 7- F 9 95 '9 IF P7-950--0 ThEH 5000 F'8 c 9 F7 F'3 63ID 5130 P7' r-950 F'8 98 850 F 3 IF C--5 0 0 0 THEH 5 05 0 F'9 F ' F 7 F4 5 TH o 3 0 P7 F'7-500 F 3 F 3 C 0 0 c 4 IF F7-500 0 THEN 5100 f - o -c 51 U 0 51 1 Li 51 7 0 51 3 0 514 0 515 0 F 9 - ' 3 5 0 0 IT 5'' ' P 7 F ' fET'JC N FFi' ' - Eiu ' ac LEA JET' 1-15 LINE CALCULATION SECTION -'J ' ' ' M L J fc t l ' ' ' ' Co' oMC53 f f ' o r -r 1 F z ' o f f' _ LJ- L' I J o c' f i Ii ' ' ' i - - ' i ' ' 4M COM M1E15J JA T r-3'1 THEPE APE 3 CHAIN ETI-VEMSNTE T I HI FPJ -F PEN HtY PE JTATtHEMT hi 'MtESi 3 3 0 5 450 o ANI' lOlO PE ' LEF'- c r LINE HOC'IFIOTTIDN INPUT PDUTIUS o IF 01 0 rHEh 40 ij -1 11 0 180 1 -10 150 1 0 170 I 180 190 cuu 810 88 'j 830 40 u' -_50 ST 0 70 2 SO 89 0 300 31 0 380 340 350 330 37 0 330 3 3 0 4 00 41 0 480 430 44 0 445 450 46 0 470 430 PRINT TH'E LErtfEC LINES USSt FC THE NETHCfH 1 'P P'T IN C ICE PRINT WITH TIME F-Nt' THE CLWMTITv LEVEED LINE OHAF'SES APE JAEEl PRINT DN FUIMT TO FCINT AIPLINE hILtWE Pf ICING IE GN A Ei_ir TN- PRINT SCftLE EUCH TH r THE PEP MILE PER MONTH CHfiP'jE ' CES IC- 'i M fc INT THE PRINT rPlMT PRINT SEOUEiTEI I T FLAT ATE Er 35 3 3 rEP MILE PEP MON TH iNIEFENI EfiT OF DISTANCE MILL IE LEEI BG TC'J l NT T'J SPECIFY NEW PATES 5 nsTSNCE IN--UT Lf IP LI ID 'oo IF LI 1 V F 1 0 GOTO 33 0 81 1 p P INT PRINT INPUT PRINT INPUT PRINT Iff-'UT PRINT ir- UT C-CEE U FOP A SU'EN UNLESS OTHERWISE THEN 800 THEN 800 ' SPECIFY THE LEAEEB LINE COSTS 0 - 85 0 MILES 'J-MILE-MCMTH R8 ' oo ''MILE--MONTH 351 - 500 MILES R3 I'MILE-MONTH 501 - 1000 MILES R4 'I' --MILE-MONTH 1001 - 1500 MILES P5 PRINT CIRCUIT Dt E 1500 MI LEE IN X PEP MILE PEP MONTH - 'I-MILE-MONTH INPUT R6 FRINT LIM 1 PEN CHAIN STATEMENT v CHAIN PCNI 4830 fi M LEAEET LINE DUTFUT SUIPOUTTNE o PRINT TAB''L 5 THE COST OF LEAEEI' LINES IN i PEP MONTH PEP MILE PRINT TAE'L 5 5 is AT FOLLOWS PRINT LIN a - PRINT THB'- L ' 5 PRINT USING 400 IMAGE MiLES 0-85 0 ci51 -50 0 501-1 000 1 0 1-15 00 1500 PRINT THp'L lO ' PRINT USING 430 PR TAB'-L 80 P3 TAE 'L 3U-1 P4 TPB L 40 P5 TAE ' L 50 PG IMAGE ' o ei' ci'xxxx PRINT LIN a ' 5 PEN CHAIN STATEMENT CHAIN RONI 1790 PEN -t- TOuA -i o' DATE IF TIM 3 o'o 4 1 MT TIM N3 4 THEN 49 0 GOTO 510 I- 16 4'- IJ 50U 510 5 r 330 540 55 0 560 57 0 58 0 53 600 61 0 2 0 630 r-4 0 65 0 6 U DC 1 6DT0 58 0 D 3 0 D J TIM' 3 lS00 I Cn TIM' IP TCI 31 THEN 69 0 IF r en IF Ml IP DEH IF LCI IF 0 1 IF Dm IF Dm IF Dm IF Dm IF DEI s Dm-DCiJ- 6ro 68 0 59 TUEH Cm TrtEN 130 151 181 818 10 40 770 800 83 0 66 0 89 U 9 0 95 0 TMEM YH3ri THEN TH8r THEri THEN 3 04 THEM 334 THEN o3 fi -4 F 3DTC 1000 M - jHhiL'HP'l r J o 710 78 0 77 0 7 0 75 0 76 0 77 0 7 SO 790 300 61 0 380 5' 3 0 34 0 35 0 36 0 37 0 83 0 39 0 900 91 0 98 0 9 0 94 0 95 0 i 0 v r v 3CTD 1000 Dcn p i -ji Yt FEE PURR o 5CT0 1000 rm D U-59 L- OTu 10 00 B i DU -9 0 Mi- flFPIL' 9GTC 10 00 DC11 D 11-59 M i '1 Y 8DTD 10 00 L'CIJ-D 11-151 MI JUNE GOTO 1000 Dm DtlJ-lSl Mi j'JLr 6DFQ 10 00 1 L 13 1' C13-81 8 M' T- fl 3uST 9DTD 30 00 Dm BCi3 -a4i ' SEPTEMIER 6C-TQ 1000 D 13 D 13-873 '13 ' QcTOPER SQTQ 1000 Dcn Dm-304 93 0 990 M3 Nat 'EM EP 1 0 0 0 END OF Toiler's IIMTE SECTION PE11 CHHlN 8TRTEMEHT 10 05 1 01 0 CHHITI RDNI 1500 108 0 END 1-17 _Appendix USING THE PROGRAM RON I FOR 3031 AND VALUE DETERMINATION RONALD C CRANE PREFACE This appendix shows some preliminary examples of the use of the R0N1 and RON2 computer programs described in Appendix I FINANCIAL ANALYSIS OF THE ARPANET PROGRAM RONl RON2 by Ronald C Crane This program written in BASIC for the HP200F is selfdocumenting The computer output itself is used in this Appendix J paper This program run together with some simple outputs forms a sufficient description of its general use The program has been designed so that different parameters such as different communications lines tariffs may i e used and the differential results compared The output is arranged in narrov column form to permit many different runs to be made and compared to OK another by simply using scissors and glue to produce an effective infinite width printout The detailed lifting of facilities considered in each run is described in Appendix H ARPANET INVENTORY This program has been made operative only recently so it is entirely conveivable that there yet may be undetected bugs Thus this program is still subject to expected reviews and changes J-i The examples tested here are for 1 10 73 Fully allocated 5-year straight line depreciation scrap value 10% With one node deleted for test 2 6 71 Long term marginal 5-year straight line depreciation scrap value 10% 3 6 71 Fully allocated 5-year straight line depreciation scrap value 10% 4 10 73 Fully allocated 5-year straight line depreciation scrap value 10% 5 10 73 Long term marginal 5-year straight line depreciation scrap value 10% 6 10 73 Fully allocated 5-year sum of the digits scrap value 0% As we nave made no additions to our present data base of the ARPANET for the month of October 1973 this month will show a zero for nev additions J-ii Peru -IQ v'C ' ' I' 3TP' ' TICMS I IMT T' ooo'' H ' o v CRFLELhTH rf' SOC IhTEL o o pp - jrT cQzi flNPLvsii r-toriEL THIS FCGFPM '-ILL CFILCULUTE THE T PEP MONTH CP THE HETI I JW hi H MHCLE PriD PEP IITE FDP THE NUHTrEP C SITES FXfiMINEF THE C'JTPUTS P E I'A'Er I -' 'ME ITEPED ir'PGPfifTTDH Hl'DL' T THE NETUISPS' HHT H SELECTED I'EPPEC IP TIDH PflTE RND 'SOJTIriG PRSIS H NOTE MH- IF IELETEI EV M I PP - EV DN THE tet E T I DM PECI'JEET crDp H FP '-l T HT THE NEI E I'lPECT IF IN MDI'E hM rr Ti CDET Nt- T Fr-i 'T-- TV- a F-PTNT- THr r - ' ofii'XTj' P F P P NDIE F'OEE PE CHN SL H THROUGH c L CQ'iCEPMM-J NHOLE Cf -' TH-'-T GP H3D i n CDETE T --K EE 'ic r tt c-iMri THE COSTS FHF IE THE liDI'E DL riCTE ETET' -CP FE UILL rTIDN- NETNOPt- E ULTi ' CALCULATIONS IN THE NET LINE HOT MOBEME INF PH TTON ETF ' N' ' TitID G7 f Pi- I -E TH M 'I - - R HPPEINPL LINES IELETIDN CCET r HFP R'iHTiv r THE hDULt' EITHER f'1pP TNH L E LEIHET TM C 'JtiCCTIDC T 'ELETEII r FPD'-PftM JFTJDM iiTh P CITJETE JTH PNI TS N FHON T S THE MflfSIMiL CDET IKE IN THE FI RF r D rT j EETMEF - r- TjrSS J INLLUISI LEFT T I FEFENCE CEFF 51 o ' I o P CMLCILHTTN NCI'F NETIFE THE Lr ' F D PAN tr-E L THE fiir-ITIC'ifl C3 TfllNET THE 5ELECTEP DFUETICN OFT IONS DEPPEClHTIUN F-'HTE- CnSTING BHSIS WMII MCHTM IF EVPLUHTIC'M E TMI-U - P-f'SWCriK IN HITH flO JHE E P L_L lihir H OF THE TNN THE DETIONS i APID' i Cr IVF'ES THE' QPTIENE - -FE tEPPEC fiT J ON MH f JF Thf CAL CUufiTFI' RfiTE M ' o I LAS'l c S - IT PIEHT LINE r Ti EC I o - I i TJ Jt-1 GF THE ' ' GP i i PEPrCI IF TH- -' p T E F N - - FF C I o ' DIGIT o' '' S Tr-'E FTFP IN E ' 0 i T H H 'o'c P E C I A T E P T Hi F P A F THE lEFFFCISTIfJ'l - FPIOI IN l'EftP Tr- -FEINT INN -S PIOV ''' lEAFF 5 ccp-- 1 -rtuE DF AN INNESTMENI THE t'EFPFClr TE PCTF'INF IE H TO T-'7F FRACTION SCPAF- VML'E I 1 I'Mi FJP THE I 'E ENC FDP THE CF El-EP TMt J-l T E PURCHASE FRACTION ' I PLL'E ' 1 Tt-if rEC I f TO NHICH LFi-PEC IhT I DM MFTFP ORIGINAL I'ECIMflL IF UH I C-H I o' A L U SCPAF PICE PEP1DI5 I 'PLUE IT IT PPE THREE o' 3 ' pl PI_PILLE T FEE O COSTING P r'I - kiCFlT TE M uMB- I t ri _ COST X NO ' o j C'-L MRfiTHL'i PENTPL COST I'E F C IPT I CM PNL JEifLCf fTI'T COST PPF MJT INCLUDED '-ON TEFM MPPSINPL COS ' Ol JL DE' ELOF f Er IT CDrT PFE NEGLECTED J HFPF - L F FULL i ALLDCPTED corrj ALL COETS MPF M rj t iti iNCiurei' 1 COSTING EPSIS I-LI-MJE T 3 EE T'-iFE IN T HEL El'PLUPTED CN M'J' EP TM OF THE PCLLOWUiJ -j- n ' I'i iHT i f' ONT 1 TM t-T L l-l'a7' ' 24 MONTH NU EEP NODE DELFT I DM OPTION Jp VDO ilRtiT TO l-Mt-'E P NODE OP NODES IFL ETEI' ' E TEP THf NUMEEP DP NODES TD E tELETEI' FIPST l i NONE PPE TO EE lELETEtt PND THEN EPCH OF THE NODE Ml'MJEP- ME FKOrfTeP' ET D '-' TEPMIN L HOW MPN-- r ai'EE MPE TO EE IELETEI' 7' 0 LEPEEI' LINE DELETION LVTION THIS OPTION PLLOK I DU TO PSJCI'E LL THE LEPSED LINES WE CONNECTED T - P NODE THIS IE USEFUL FOR DELETED' NODE'NHOSE LINES LINES HDULD NOT EE CONNECTED STFPI'SHT THPDO' H THE NODE NTPE NOT THEPE Men NPNT NODES P E THEPE MTTH PLL THE LINES DELETE 7'0 Example 1 PUN mm Do v'OU UP NT INSTRUCTIONS BEFPECIPTION F'PTE BEFPECIPTION FEPIOD ' S V oo N OP ' 'N D 'o 's VEPPS 'o' T 5 SCPPF- V LUE DECIMPL FPPCTION ' 1 C COSTINE I'PSIS S L CP ' ' ' F MDNTH NUMEEP '46 HOM MRNY NODES PRE TO EE DELETED 71 THE 1ST HODE TD IE DELETED IS 5 HON MPNY NODES MPE THERE KITH _L THE LINE DELETE 0 MPPSINPL COST OPTION V -H 'v J-2 -- CfiPUer'ftTia HJ -DCIRTES - o flPPwNET H'iiiLi n - r-lnrei_ CEST JCTSEEP I ' 197 DATA BfiiE ACCURATE RS OF 6-73 iurvinP t 1n-jTH E E' ijjftTia i 3JTir ' HhFI CEe sLCI iTICH INVESTMENT D r M 4 r PhTF i'cPne MftC'biriHi crrr INPUTS DP FULL 1 0 i i EP T' - 1 7 HI_LOCHTET TOMI MT LINE 1 'MLUE DPI ION INTLUTEP n -r r DUTFUTS o o Fii_L ITEMS E INCLUDE ' IN THI - CUTF-UT o cEp HDMTH Ftp ' ire I pep f' nT4 TDTML iDTh' INVESTMENT ra Di-'-Tg- -- --- I -- i IcFPFCIf'iTICN TD DPTE Pn f o _o- C 7HS NETHC0-' T TDT L INVESVMEN IN finr TH i'1 TET'- fl'-'EPP E DELETEri PEP I PEP Tat L TnTflt f'DNTH IN I ' EMS FSP f n PFS NOT INCL' T't'D IN THIS IlTE INVESTMENT IT prt ri i il-IrJHL MfjNTH FEP ir-i F SF IN MC1 -JTH P DCTl T- - Can - ptP I SEN I I TE PpTS PEPPEC JHT ran in DPTE Faa -- ''PL'JL' EF THE NETNDP-' TOTM' IN' EST'-r T Ttl NENT M i-'t'HT IN' - EST'lENT flM SIi If i 4ft1 I 4 f I - FEP MCMTH SITE f HI'I j O 3r c o fi I 0 'I TDTML H'-SWHGE E T PI O TH CCST -- K SITE 1 '- i % i MCNTH IN' ESfflE'fT r-r i r OUTPUT DpTE IN ' Fs MENTM TDTPC TOTP'_ IN' r-TMENT Til HEN 5 ' M MENTH nO-JTH TUTML lili'tiTr Ef r TC Ri'EP SE j 2 41 0 TaTMi- IJEF rcrftiiSN rr I HTCBorv- i_' OP THE KETWCV MEN i 'j _- - 4 7 0 I r r T rs s s OUTPUT - - -- i i INVESTMENT FEP SITE IN -io-r f'i'rVN IT Pirrer K 1 - J-3 IN MSNIH M CF TNC FPE'-IDHS t 11 4 I T 4 c'05 10 j00r' I 'I 4A '5ri r f NT C ' ' ' ' ' 1 0 o LITE _Tr T _t riT rc i Trj t T T nn-r U LP PI U B 7 UTRH PErJ MIT FKND SDC HRPV LL ThN I-L CR E CMU fi-CS RMES MITP PFit'C NBS ETRT TIN ' McCL U C SIC rOPT 4 cr TIHT EELV appR RE 'D EBM CCfiT XPG - FNWC LBL -- - 3 10 11 1 4 - 14 13 16 4 -S A 1 i 20 21 iZ-C l_ _o 54 T' Cw o c' -' ' 0 31 o c T ' ' 4 ucsn PMLT o 3b - i J-4 Example 2 FON 1 DD ' au --M -JT i j T 'jrTiunr ' Y N I'EPPECI T aw F' TE 3 DR I 7 N T DiEP-crciiSiTinri PF r5ir r- ' ' ve s 'o '5 1 -W F VALUE EECIMAL PP CTIDN' ' 1 COST if f isis 3 L OP F ' Ti_ f awTH NUMSEP ' 13 How MAN Nnr'E i -e TD E I-ELETEI 0 HOW IAN V NDIEI MPE THtPE WITH hl L THE LINE - RF'P'Hr-ET I'E- LETEI' 7 0 o' Cz-r ftunuijri Mnr-rL CT3fEP 11 197 DftTfl EfiSE RCCURRT3 R3 OF 6-7 3 SuMri- -V f'lCMTH CF E' U-'r-TTOH - '' fT' IS D 6 '''T o o T J I T-v-fj-i FEp FE r- -P TDTPU ITF-MECIRT DN te H Etn rn IV-ITE U ''EET'-iENT Firt'ED 3 - 1 71 TSHI HT f I I pOD - I'I- -JJ CF TMH r- THD- '-' TCT L In' r JThFNT I r J MCNTH c E PR 5 TDTR'_ r DP LIN 'MLUE ''i-fJTH F ' I Tc f MTH TDT L Ifli-'E TMENT TO flPTE Rl'ERni - Itir 'i T5 IN i T r FEP 5 I TE llnNTM IN rlD'-ITH M f'1 3 TRT u r 3 ITC M 3 J J -f 3 -S't l 343 fll f I 47 640 I 3 S' 'J J 19 J l i 0 fl ' I iH Nril E C L C U I T i n r J ffj U rEri I' L'l C T i r U JB IJTHH 3 4 F3H M IT F'Hfip 3 DC 3 3 7 3 10 17 HHR' ' LL HTTP J-5 DFL ET I Dr-J 1 o ' fl o J 3 033 Example 3 RUN PON I Dn - ou U NT IKSTPOCTIBNJ V V M ' 'N BEFPECIftTtCN PPiTE '' 1 OP I' ' ' z DSFFECIftTION FEPTOt' ' vEfPI ' 'o '5 iCHftS V L'-'E ' DECIMHL FBMCTIDH'1 7' 1 COSTIN'3 EPSIE ' S ' L OR F F MawTH N'JMEEP 18 M Q -J M N I rlOFE fiBE TC IE IELCTEI'O HOW l NV NGtES PiPE THEPE MITH t i_I T E LINES I ELETEL U -- - LMELEI' T MyiCCIfiTE RC'PRNET CC -T H'iHL 5 DITIEP KSSSSSISSSSSI f CT'Eu 1 'a73 11 DiVTfl EASE RCCUPRTE RS DF 6-73 i'JMMnP'i 'o'n iT-i GF El '1 LiJ'iT 13M ' Y ' COSTING EPS I DceF ClMTlSfJ 'P TE INVESTMENT 'C HF 'V'RUJE OP Ir FUT 19 F j 1 3 6 - t l ' I_L ' Rt LDCflTEt IEHP ' T Pr '3 - T L I r-J r - - 1 -t OUTPUTS s s s tra 5 I - 3 I' E f f'ln- rw PF SITE $ CEP MDNTH ToThL TDT L INVESTMENT TO DPTE TOT- U LEFPECIMTISN TO I'MTE Eao- IuL'JE c' THE I NETNO K INVESTMENT IN M JNTM M R'-'EP -SE Irli EJTME ir FJF 5ITc TOTH MEN C 'ENT PUI-EU IN 7ITE MONTH M LTHTU SITE 'JC'LR RI U33E UTRH EEM MIT PPiMD - T - IN MONTH M HRME 39 o ofi 1 9 0 0 0 1 8 ft'5 1 f IN CSLCULFTIONJ Ma 1 c i - SEI l 4 cr j 7 HHPV fj LL 10 17 ll i 1 p o % 30 d34 0 7 3 4 r rio 4 40 J-6 DELETION 3 ri33 L EL EI MT 1 1 hr 0 iftTEJ RFPRNET r MtihLvrii Hol EL COST Example 4 De rarEf 11 197 7 l f m ERSE RCGURRTE ftS QF 6-73 S iJ M M H B v MONTW DF M ' ' EI'MLUHTICN F 5 ' 'HL'J'f IT3FT I NPUT 5 4i CDJTIU6 rn ir DEPRECIATION PftTE INI-'SSTNEMT OP OP -- 10 197 FuLLT PLLOC TEX' vEhO- STPMINHT i INF 1 CUTFUT 'J PEP MONTH FEP SITE i 1 3 -5 3 l FEP MONTH 1 5 ''c' c'57 1 3 45' '' nn I t 'I 1 033 470 2 419 j53 ri 0 TOTMI TOTHL INVESTMENT TD TDTSL DEPRECIATION DATE TD BATE Bno I-HLUE CF TH METHDP TrjTFiu INVESTMENT It-j MONTH Hv AGE HEM P'EtJT INVESTMENT FitT'EI - IN SITE FEF o' SITE TPTUS ITE IN f' h CALCULAT IONS ND T f PM MIT RflND D'DC HRFV l L 3TRH ILL CRSE UMU R'IES RMES MI TR PR DC il _l 7 o 1 0 l r nrur 4 11 12 13 14 15 ETRT 4 r4E - o 'JrFI 1 U HH i M 'l0NTH f'1 UCL H SPI UC F IN MONTH f 1 18 19 - ' o' 21 flcCL U 'C cc 'j 1 u' 24 fiGHT SHRT BELV R'-PR RF PTi EE'D CCRT o1- ' j ' ' Z- Z ' C l' u c c '9 ' ' 31 PG FM' M j LEI 34 U D - er '- - Mf C ' 1 T J-7 1- 0 l- 0 I'ETLF - - -- ----- -71- t rlF Ll C'r ' RrPRNcT Example 5 Co T HF CO TINi3 H -e Ei'HLUiiT '' CF ' f'V' INVESTMENT PflTE SZP-F V LUE QF 6-73 Iflri 'T 46 EriFIS jEFPECInTIO't -- i Tc_'_ f'tlBLi-t Dfljfl EftSE HCCURhTE Ft 'Jrl'inP i MONTH i JC 1 rj i_ LOt -1 E I 1 - l 4 Te TM I' a-il'iHl ' T-nlGnT i_ IC' 1 OUTPUTS - - - _ _ - I FEP MONTH FEP ITE I ' ' T S PEP MONTH TDTML TQT L E 3DST TfiL CTAL INVESTMENT TI B FPECII T CN I 'MLUS CF THE INVESTMENT Hi P '5E TO D TE NETNOP IN rN'-'EsTMtNT NEW PENT EXED P TE MCTW FEP SITE IN MONTH IlTE M IN tf MOUTH M STfiTuS tr ITS NAME Sft-CUl-FiTIONS N J UCLfi 1 US HE i UC E 3 UTRH ZBt M IT RNP 4 5 6 7 src g HRPV i-L 3TRN ILL 9 10 11 1 CR E I CfVJ HUES' RME MITP 9RDC MBS ETRT TINK McCL USC 31 i C HURT SIiflT EELV RRPR flBPD BBM CCRT XPDX FNI 'C LEL UCSD HRI H PMLT 14 15 16 17 IS I'5 '0 21 IB 23 24 25 26 27 23 29 30 31 32 33 24 35 36 37 J 8 1 f I I fu - 'J 5J unn 1 S I o So '5 2' 1 6 1 1 6c' I f 0 I '' DE ETIC i Example 6 PUN Farn Da T uv - r-tiT iri -TBi iC'iCN - V M PEFRPCIATID J PATE o S DP I' IlEFPEClHTIUf FEPICf t-lDW MANY HDW M N NOTES Mai'F - fiPt HFE Tn E THERE HPF'RMET ' ' 'H I1 'EWS ' '5 CELETEt 0 H1TH l i_ THE COST OcTDEEP RUHLV5I5 1 LINE IEl_E TEI' U MDI-EL 19 rftTFi BASE fi'XURRTE 1 1 DF 6-73 EVriLUflTia 'M tf- i Drrcsr ci T i -jto r' rT I N oo E r T r t T Z CPP P ' o - _i i MC'ITH a- 46 F ---13 r ' CD -TIT i E T F-EP h 'n rtTM tlDUTM FB ' iTc TZJTPL ToTfii IH' 'E-THETJT TDTML PE FECIPTIDN Enni TOTAL r-'' 10 - 1 7 ' ULLi HL SEATET t W' SU T PP L- r TG IlPTt TO DpTE ' 'HLt'E CF TM riETwapiIm-E VTnErjT IN MDMTH o ft -'EPP'Z-fc IH 'E Tr ENT FEP EITE D NEU vn PITET ' IM KDHTM M J-9 X r 4 MtHiiTH '1 ri r-iTC o 11 u 1 44c ' 14 T ' ' 45 l l I ' M % l 3 4jf-54 1 l 1 6 1 444 0 1 ' T 0 SITE -STATUS 'ITS UCLR - r-i _ 1 U IE UTRH r kMe IN CSLCUL TICW No Ur 1 i tl 4 EBN er _i MIT c riD t' r 3 '- 10 11 WHPV ITRM ILL CRvE CMi i RM S RVE MIT PR DC r F ETRT TINK M-CL UIC 6 MC 14 15 1R 1 13 1 30 1 4 NDHT IDRT PCI ' o RBPD BBN ' CRT 'PCX FMMC LBL UCTD MRU I PULT 30 31 34 J-10                     National Security Archive    Suite 701  Gelman Library  The George Washington University    2130 H Street  NW  Washington  D C  20037    Phone  202 994‐7000  Fax  202 994‐7005  nsarchiv@gwu edu