aJUVl J aJl1 1 Bmwvu aJ I WI lrUlWV lD13UlWlD13 lD61 W13a1 13 f WalW I1 ilIIi lBrn Urnl1rn J BU JI I Q 1J0B1 2 P L '1 LEADERSHIP A PERSONAL PHILOSOPHY U ooooooooo WHAT'S THE GOOD PASS WORD U o o o o o o o o HUMAN FACTORS TEXT EDITORS U THE REALITY OF COMMUNICATIONS CHANGES U 86-36 57 i i 1 6 o oo oooooo 9 oooooooo ooo ooooo 12 PUZZLE U ooooooooooooooooooooooooooooooooooooooooooooooooooooooo oo l il o o o o o o 14 SIGINT 1990 Part Two U ooooooooooooooooooooo 1 o ooo 16 ANSWER AN OLD PROBLEM U oooooooooooooooooooooo o oooooooooooooooooooooooo 29 NOT SECRET ANYMORE ooooooooooooooooooooooooooooo 1 29 r 'fIllS BOCUMBN'f CON'fl' INS COBHWORB Ml' 'fHRUb eLASSIPIEB BY lisA essn 123 2 R o d Agefte a BetermlnatlOiiequlZEUR NOT RELEASABLE 'FO CONTRACTORS Declassified and Approved for Release by NSA on 'I 0-'1 - O'I pursuant to E O 135 6 rvlDR Case # 54778 DOCIO 1 Published by PI Technique and Standard VOL IX No 10 PUBLISHER OCTOBER 1982 ss BOARD OF EDITORS Editor o o 1 8322 7119 Aut Editor o H' r--------IOl03 Produc t ion 3369 Collection o1 8555 Cryptanalysis o o 1 5311s Cryptolin gui tiCo 1 1l03s Informat ionSgience 1 5711s Language oo o i 8l6ls MachineSupport J 468ls Mathemati98 oo 1 1 8518 Puzzles oo o ooooo David H Williams 1103 Speda Re e rch o Vera R Filby 7119s TJ aUic AnJlylti8 o oo o Don Taurone 35738 d ooo For ub criptions send name and organization to CRYPTOLOG PI orcllU 13369s P L Can two walk together unless they are agreed And can they agree unless they communicate Many of the tasks we find ourselves confronted with these days require a multiskilled approach From the manager's point of view it would be nice to have a lot of multi-skilled people around so that when a problem came up one could decide just what mix of skills were needed and then order the proper multi-skilled person s to go in and solve the problem We might even set up a data base and query the computer o Although managers continue to ask for multi-skilled people to do these tasks it seems clear that most people learn only one kill at a time Therefore mo t of these tasks requiring a multi-skilled approach must be attacked by using two or more people each with different skills Supervising a group of people with similar skills and perspective is hard enough but trying to manage a group of people with various skills and viewpoints is often much harder especially if the supervisor doesn't have all of those skills That is probably why managers keep asking for multi-skilled people 86-36 To submit article or letters via PLATFORM mail send to cryptolg at barlc 5 bar-one-c-zero-five note no '0' in 'log' Part of the problem lies in communications When we talk with people in other skill fields we may think we use the same language Mostly we use the same words--with different meanings Each kill area develop it own working dialect borrowing words and changing their meaning to fit the skill area The answer probably lies in listening to one another NOTE The front cover of the September should carry the caveat Editoria' 1982 issue Not Relea able To Contractors peR 0'1 1P iW hf dht- epPle HS 8MhJPY DOCID peR epPf9f h YS eHbY LEADERSHIP A Personal Philosophy u P L by 1 II _ e ers ip has received much at en- tlon In r cent years from varlOUS writers aDd r rch groups The question of how we get people to cooperate on a ca on task for the purpose of achieving a hared loal has always been a major problhl for _nkiDd Concern with this question has mounted as the complexity of tasks in our world has increased to the point where relatively few jobs can be accomplished by an individual working alone Leadership has been defined in several ways by tho e who have worked in this area Leadership is the exercise of authority and the _king of decisions Dubin 1951 Each member of the Class of 1982 Army War College was required to prepare a paper setting forth a personal philosophy of leadership which best met the challenge of a senior leadership position in his organization or agency This paper was selected by the faculty for publication and is reprinted here by permission of the author can be no leaders in isolation If you want to know whether you are a leader see if there is someone following you The difference that makes the difference between successful and unsuccessful leaders is of great importance to me and I'm sure to the other students here at the War College as we move towards higherlevel civilian and military positions The leader is the person who creates the most effective change in group performance Cattell 1951 THEORETICAL CONSIDERATIONS OF LEADERSHIP The leader is one who succeeds in getting others to follow him Cowley 1928 ''Leadership is the procesa of group activities toward goal etting and goal achiev_nt Stogdill 1948 infl encing Tvo important threads run through all of the above definition The first is that leadership is a relationship between people in which influence and power are unevenly distributed The second observation is that there Oct 82 peR Leadership is always a relative process To be effective and to communicate as intended a leader sust always adapt his behavior to take into account the expectations values and interpersonal skills of those with whom he is interacting There can be no specific rules of leadership which will work well in all situations Broad principles can be applied in the process of leadership and furnish valuable guides to behavior These principles however must be applied always in a manner that takes fully into' account the characteristics of the specific situation and of the people involved CRYPTOLOG 9PPfe Page 1 Be BHbY 86-36 ---rtO-CID 4009853 FQR QFFIQI YSK QN Y o high-initiating-structure low-consideration o high-initiating-structure high-consideration low-initiating-structure high-consideration low-initiating-structure low-consideration productivity and employee satisfaction The findings were neither consistent nor definitive The major conclusion that can be drawn from these studies is that as previously stated there is no one best style that is appropriate for all situations HOW WE CAN GET THERE FROM HERE OR HOW I WOULD RUN MANAGE MY AGENCY It is possible to state principles of leadership in such a way that they are incapable of empirical refutation and appear equally consistent with quite different forms of leader behavior To say that a leader should manage in such a way that personnel at all levels feel real responsibility for the attainment of the organization's goals Likert 1967 or alternatively that he should exhibit concern for both production and people Blake and Mouton 1964 is not saying a great deal about what he should do in the concrete situations that he faces daily The increasing tempo of our social evolution requires improved human relations techniques and increased knowledge of the social sciences in contrast to the greater emphasis on mechanical techniques and the natural sciences in the past To cope effectively with these shifting demands the successful leader must have a keen sense of timing and adaptation He must learn to roll with the punch and yet maintain a gyroscopic guidance for the organization keeping it within the confines of fundamental principles and headed toward a sound objective My initial thoughts on how I would run such a massive intelligeQce organization as the National Security Agency coincided to o large degree with those functions of an executive as delineated in numerous writings on administration These leadership duties essentially are as follows 1 Planning determining what should be done--including clarification of objectives establishment of policies mapping of programs and determining methods and procedures 2 Organizing grouping the activities necessary to carry out the plans into administrative units and defining the relationships among the executives and workers in such units 3 Assembling resources obtaining the personnel capital facilities and other things needed to execute the plans 4 Directing issuing instructions Indicating plans to those who are responsible for carrying them out and stressing personal relationShips between the boss and his subordinates 5 Controlling seeing that operating results conform as nearly as possible to the plans--the establishment of standards motivation of people to achieve these standards comparison of actual results against the standard and necessary corrective action where performance deviates from the plan Over the past three decades research has been conducted into the relationships between the four styles of leadership Oct 82 CRYPTOLOG FQR 9FFIQlhb YS8 Page 2 QH JPY DOCIO 4009853 FeR eFFIe lAb HBB 8lfU However after some intensive reflection on this subject other salient aspects began to emerge which embrace if you will my personal philosophy of leadership Firstly I believe that leadership is the process of influencing a group or getting others to accomplish things A leader doesn't necessarily have to solve problems rather his job is to get problems solved TO HAKE POLICY IS EASY TO HAKE IT HAPPEN IS DIFFICULT During my twenty years as an analyst at NSA I have been exposed to a variety of managers some good and some not so good The manager I have the greatest respect for is an individual who a delegated responsibility hesitation to me without and recom- would cite some inspirational remarks made by the following outstanding leaders who were truly men of vision Ask not what your country can do for you but what you can do for your country John F Kennedy I have a dre8lll Martin Luther King Some men see things as they are and say why I dream things that never were and say why not Robert F Kennedy b trusted me fully c listened intently to problems I would then ask my executives to provide a frank assessment of their individual leadership skills style and effectiveness with particular emphasis on the following general areas mended solutions and d supported me completely If I occupied the top position at NSA I would make every effort to emulate my favorite manager's leadership style to the maximum extent possible One of the most important things that I learned from hi was that the higher your position is the more no less dependent you become on others a Working with people b Personality and character traits c Making decisions d Developing and selling ideas e Delegating responsibility Secondly I believe that a leader must be courageous and possess the ability to make rapid but sound decisions As you are aware intelligence information come from a variety of sources and is often fragmented and incomplete Even in the absence of complete and vital data I would not hesitate in making the critical decision required Remember the moment of absolute certainty never arrives One can never be 100 percent sure of anything The next step would be to compare their evaluations with the following criteria that I have personally established for these general areas Thirdly I can not overemphasize my feeling that leadership style must rise above day-today activities I strongly believe that a leader must have the mental agility and the creative vision to conceptualize how things come together As the Director NSA I would make a concerted effort to instill in my subordinates that in order to become effective leaders they must possess the courage to dream the ability to organize and the strength to execute My first step would be to host an informal gathering of my top-level executives Office Chiefs for Planning Operations Budgeting Research and Development COMSEC etc at some secluded location that was conducive to relaxation and reflective thinking To set the tone for this session I Oct 82 CRYPTOLOG Page 3 DOCID 4009853 FaR aFFiSiAh HSS QNhJPY outline their duties and responsibilities preferably in wr1t1ng in terms which the subordinates can understand Planning the work and having a program are great assets in effective leadership A leader should learn to inspire confidence and develop enthusiasm in subordinates He should be a good listener show a sympathetic understanding and try to see the other fellow's viewpoint He should keep the group looking ahead with a well worked out program He should take an optimistic view whenever it is feasible and make subordinates comfortable when the going is tough He should show confidence by expecting much and letting subordinates know that he does expect much He should compliment people when it can be done appropriately and sincerely and he should in fact look for opportunities to do so Working with people Personality and Character Traits Working with your superior Dependability and follow-through constitute two of the strongest measures of effectiveness in working with your superior Initiative and doing more than expected make the best sales approach for a subordinate in building himself up in the eyes of his superior It is always constructive to help the superior by supplementing his weaknesses particularly when this can be handled with a reasonable degree of finesse and understanding I consider purposeful energy progressiveness intensity and health as being particularly necessary to the successful leader Purposeful energy is commonly reflected enthusiasm initiative and optimism in Progressiveness is often reflected in the ease of generating and accepting new ideas Working with your associates Make yourself easy to get acquainted with and to work with-loosen up--be adaptable Be alert to your associates' attitudes--develop sensitivity Learn how they react and what motivates them-develop objectivity Show interest in your associates seek advice from them and go out of your way to help them Working with your subordinates In working with subordinates the necessary respect prestige and confidence to warrant acceptance as a lead r must be earned Such acceptance results from a consistent application of high personal standards and sound principles and from a good quality of decisions and ideas A leader should seek the advice of his subordinates on problems and keep them posted on new policies and programs He should help them to start new projects and carefully evaluate results of completed work He should also Oct 82 FQR Intensity and drive are characteristics of the outstanding executive which make him and his job almost one A real leader reflects his job and the job reflects his personality Health is a factor of growing importance The requirements of leadership are becoming more exacting and complex and the tempo continues to rise The increasing frequency of hypertension heart ailments and stomach ulcers among executives particularly those of middle age should be viewed with alarm More real relaxation better planned vacations and greater hobby interests are certainly desirable CRYPTOLOG QFFISI1 Page 4 YSS QNhY DOCID 009B53 POll OPPieIlIrt ijSt i etltfl Making Decisions Sound decisions are generally based on a combination of background knowledge available facts and exercise of judgment One should feel and express conviction n selling ideas The old saying '''be sure you're right--then go ahead certainly applies At the same time to avoid creating negative resistance the degree of aggressiveness must be moderated by the proper amount of tact finesse and understanding Delegating Responsibility The knowledge element in decision-making is primarily derived from theories principles and practical experience Good executives will coordinate group and use other people's knowledge for their mutual benefit The facts element is highly important in arriving at correct decisions The ability to recognize facts to know 'what facts are available in connection with each particular problem to be able to assemble them and then to analyze and interpret them is fundamental to good leadersltlp The judgment element includes drawing sound conclusions based on knowledge and facts It involves thinking effectively--sifting the important elements of the problem from the superficial and exercising perspective Developing and Selling Ideas A leader must be willing to delegate responsibility and to decentralize authority You have to trust your people to do the job required In working with subordinates a leader should remember that one of the greatest human desires is to create or produce something and to feel essential Furthermore as I previously stated the higher your position is the more dependent you become on others I truly believe that from the discussions generated in the exercise described above leadership strengths and weaknesses can be identified and corrective actions can be taken to facilitate promote effective leadership at my agency If a leader develops his executives he will in turn be even stronger You can handle people more successfully by enlisting their feelings than by convincing their reason A leader must get ideas e sily Hence he should try to develop his u agination Ideas must first be obtained thentllsted and timed for presentation and finally sold to others for application Ideas usually originate from knowledge or research 1 believe that every important office in my agency should continually apportion some of its time to research for new solutions and new ideas A leader must learn to sell ideas--to get his thinking across to others One approach is to learn to write well o ganized clear convincing and hort morand s Another is to learn to conduct himself well in a conference to go to a conference with a plan of attack with suggestiQa and not just to sit in And still anothe is to learn to speak to small and large grouP_ this not only builds personal presti UiIt h irtually a must for present and fut l aderoo Oct 82 CRYPTOLOG Qi 9FFHHMJ Page 5 sse etlta DOCID 4009853 WHAT'S THE GOOD PASS WORD U byl 1'4 P L DDT Senior Computer Security Coordinator computer security compromise was recently discovered here at NSA by an evaluation team from Cl the Computer Security Evaluation Center The team was evaluating the computer at the request of DDT II 7' the following is the recovered by the team LIST OF RECOVERED PASSWORDS abyss aesir again apple april backward bandit barba baseball beaver beetle bible bingo bluetop brogue bushed candy canoe chance check cjunk confused converse copper crazy design digital dogwood donna drifter dumbo eagle elephant empire The compromise was the result of an administrative error The system administrator had made a copy of the password file which included the identifier and the encrypted password of all users of the system However the protection bits for the file were improperly set and the file could therefore be read by anyone on the system The evaluation team attempted to see how many passwords could be discovered from this file They tried guessing passwords sending them through the password encryption process to see if they matched they recovered some passwords The team then sent a dictionary of common words through the encryption process and compared the results still more passwords recovered Then the team tried all five letter values from aaaaa to zzzzz in three days they recovered a total of 107 of the 255 passwbrds in the file All users of the system have been directed to change their passwords and the system administrator will be more careful about protecting the file in the future In order to give the general user some idea of the kind of password that is easily guessed or discovered Oct 82 CRYPTOLOG list Page 6 peR epPlSUb fSE QtlbJPY escort eyeball fairway fishs flower freedom funny geisha genie golden green happy hawthorn holiday horse intel ironside joewood kelly kingfish kitten lambda landing lemon locked logic login lumpy major michael mikey mouth muffin murphy passwords 86-36 DOCID 4009853 fQIl QfFHnAl5 ISS BlihY ---------flower nroha omega panel panelist perch permit pmrmd poets popcorn puissant rabbit radec redbud romance sail ing scooter scrabble security shari silver kitterr canoe snowball softball ssret stage striper summer sunshine susan tacos testing tracing update vision vkjrd westward wicks wildlife window wizard redbud Passwords are encrypted to prevent casual recovery by someone else but you must cooperate to make this effective You must choose a password that will not fallout through a simple analysis effort The simplest factor you control is the length of the password The longer the password the better If the password vkjrd in the list above had been one character longer it would have taken about 26 x 2 1 2 days or 65 days to recover it The table tells you that if you use only a single lower case letter as a password it can be recovered in at most 26 tries If you use eight characters of lower case upper case numerics and punctuation then it could require 1 000 000 000 000 000 tries to recover your password A second factor that you control is the alphabet size If you mix in both upper case letters and numbers you increase the number of tries needed by the exhaustive method i e trying every possibility Note that there is not a single upper case character in the above list of recovered passwords A good password is eight or more characters in length At least six different characters should be included You should select uncommon words Mixing in some upper case characters along with numbers and punctuation characters helps defeat discovery through exhaustive search EXHAUSTIVE TRIES NEEDED Password Length Alphabet Size 2 6 4 Lower Case Letters 26 26 Upper and Lower Case 52 52 2704 7 3x10 UC LC and Numbers 62 62 3844 1 4x10 UC LC NR Punctuation 75 75 5625 3 lxl0 7 456 966 676 Oct 82 6 7 CRYPTOLOG 8 3 08x10 8 1 9xlO 10 5 6xlO 10 1 7xlO ll Page 7 FeR eFFIE IAl5 ISS erll5Y 2 08xIO ll 5 3xlO 13 2 lxI0 14 I OxI0 15 _DOCID 4009853 FOR OFFU I JIai ON Y __ t _ __ I _ _ _ t _ _-- -_ _ liOlo _'_ --' __ _ - 1 0011' I Has someone else already finished the puzzle in this issue Are all the letters filled in In ink Have you ever wished you could hold a particular copy of CRYPTOLOG for a while so that you could get a chance to read one of those longer articles Have you ever thought about getting a personal subscription to CRYPTOLOG It's really a rather complicated process but if challenges appeal to you then hereJs how you can do it l find a gray phone 2 dial 3369 3 ask forl _ 4 give him your name and org P L 86-36 OR 5 send your name and org to _ _ _ _ _----Jlp14 I - Oct 82 -- CRYPTOLOG Page 8 FOR OFFIOlkb l JSE OnbY ' DOCItL 4009853_' _ _- L- 1 P6R 6PPfe SH 6NtJPY HUMAN FACTORS Text Editors u by 1 --- P L lp13 REVIEW HOW DO PEOPLE REALLY USE TEXT EDITORS J Whiteside et aI Proceedings of SIGDA Conference Philadelphia 21-23 June 1982 This study carried out by five researchers at DEC Digital Equipment Gorp compared knowledge workers and secretaries performing their normal work with two different text editors Keystroke statistics were gathered during normal usage of the editors Both of the editors were screen editors One called EPT Editor Prototyping Tool was an experimental editor which contained a built-in keystroke logging facility It permitted a simple set of editing functions including cursor movement anywhere on the screen search text relocation and text deletion Text could be inserted or overprinted in modes activated by a function key It did not offer the formatting functions included in most word processors but it did provide automatic word wrapping The other editor WPS was DEC's 200Series commercial word processor It was a character-oriented screen editor in which editing was done on the bottom line of the screen The cursor could be moved anywhere within the boundaries of existing text It had logical equivalents for all EPT functions but also offered many more functions including sophisticated text-formatting Oct 82 Six knowledge workers were studied as they used EPT to create and revise their own reports and memos they had at least three weeks experience with EPT before the test began Eight secretaries provided the sample of WPS usage as they carried out their regular secretarial and administrative duties in a DEC R D department Their experience with WPS varied from one month self-taught to two years and three training courses EPT keystrokes were logged by a built-in facility while WPS keystrokes were captured by a PDP 11 34 computer For both editors the logging of keystrokes was accomplished in such a way that editor performance was not affected Each keystroke was logged along with the time of occurrence The results of the study are highly interesting and I believe they are directly relevant to our work situations Neither EPT nor WPS is probably directly comparable to the text editors we use EPT may be somewhat similar to the UNIX Rand Editor Still it seems probable that the essential editing functions printing characters cursor movement are comparable across different editors The methodology of the study has even more to offer than the results in my opinion I would very much like to see analyses of this kind carried out on many of our interactive systems text editors and otherwise CRYPTOLOG Page 9 peR epp reMh 8'9B elir 86-36 - nOCID 4009853 Inactive time was logged whenever no keystrokes occurred for a period of more than 150 seconds The secretaries usually turned on their machines as soon as they came in in the morning and left them turned on all day while the knowledge workers were using EPT on a time-sharing system and were less likely to leave the editor running when they weren't actually using it The actual work hours and total keystrokes are quite comparable in the two samples in fact the six EPT users generated a few more keystrokes within slightly fewer work hours than the 8 secretaries This seems highly interesting in itself I would' like to know more about this unexpected showing in which six people who were not presumably trained typists but professionals doing their own documentation turned out more keystrokes in a shorter time than eight people whose primary task was skilled typingl Unfortunately the paper does not address this point so we are left in the dark about it It may have been related to the text editors in that WPS may have provided more power with special functions so that the secretaries could accomplish more with fewer keystrokes It raises a question about assessments of productivity based on keystrokes per person hour in any case The data was first studied at the level of individual keys What proportion of total keystrokes for all users was taken up by pressing each key What proportion of the total time was used pressing each key What was the average time per keystroke Transition probabilities were computed for pairs of keys How likely is it that a user will press key B after pressing key A The degree of randomness or organization in the pair transitions and the degree of predictability in longer strings of keystrokes were studied in various ways Finally keystrokes were grouped into larger classes to provide a more global analysis in terms of states making it easier to compare results with different editors in terms of basic functionality for the user Here are some highlights of adapted and simplified from tables the the results author's CHARACTERISTICS OF USERS AND WORK SAMPLES EPT Type of Editor WPS Simple Screen Editor Sophisticated Commercial Word Processor 6 Knowledge Workers 8 Secretaries Total Person-Hours 212 482 Inactive Hours 127 390 86 92 510 513 406 102 Users Active Work Hours Total Keystrokes Oct 82 CRYPTOLOG Page 10 iOR OnICU I USE Ow y DOCID pO OPPIehH gSS 9N1 JPY INDIVIDUAL KEYSTROKES IN EPT AND WPS 1 1 1 EPI I % OF KEY STROKES I I I P-r i-n-t-C ' h-a-r- 1Left-Arrow 1Right-Ar row IDelete IDown-Arrow 1Up-Arrow IReturn 1Tab IWord IAll Other 1 WPS 1 I % OF TIME 53 58 12 90 12 52 7 35 5 67 3 15 1 38 1 01 34 2 10 KEY I 1 % OF 'STROKES 1 I I Pr-i -n-t Ch -a-r- -1-- -54 - - 8 1- 46 84 5 08 5 49 7 38 9 89 5 89 4 38 1 66 39 13 00 _ IAdvance ILine I Backup 15 48 10 30 '6 59 IRub Char '3 33 IReturn '2 83 1Tab I 1 52 IWord 1 24 IDelete Char I 1 11 'All Other 1 2 79 The grouping of keystrokes into states produces results that modify these proportions somewhat as the table above shows Typing uses up a bit more than half and cursor movement more than a third of the keystrokes Erasing accounts four four times as many keystrokes as the all other category in both editors It is interesting to note that the knowledge workers gave twice as much weight to erasing than the secretaries did No explanation is offered for this difference in the paper Perhaps the secretaries made fewer typing errors on the other hand the difference may be related to creation of documents by the knowledge workers Composing an original document might be expected to call for more erasures than transcribing or updating a document originated by someone else I recommend this paper to all readers interested in human factors in our Agency In my opinion it represents the kind of welldesigned practical study we should be doing to obtain more hard data about user system interactions in our offices STATES IN EPI AND WPS IPERCENT OF KEYSTROKESI EPI WPS I I I 1 I I Cursor I Ty-p-e--1 I 55 97 I 34 39 1 IErase 8 27 I I I I IAll otherl I 47 48 11 91 6 27 7 41 3 35 8 08 1 79 65 90 12 16 ----------- Typing in characters accounted for a little more than half of all keystrokes and nearly half of all time in both editors Arrow cursor movements in EPI accounted for about a third of keystrokes and a quarter of user time while Advance Line and Backup cursor-movement keys in WPS accounted for very similar portions of keystrokes and time Nine key types out of a total of 29 in EPI accounted for 98% of keystrokes and 87% of time while nine key types out of a total of 64 in WPS accounted for 97% of keystrokes and 88% of time Thus even at the level of individual keystrokes these two apparently quite different editors were used in a very similar way by two highly different sets of users The authors of this paper summarize these results as follows A rough rule of thumb that emerges from our data is that free usage text editing consists of about 1 2 typing about 1 4 cursor movement about 1 8 deletion and 1 8 all other functions put together This rule has been shown to apply in two different situations knowledge workers creating documents with an experimental text editor and secretaries transcribing and updating documents on a commercial word processing system p 37 I StATE % OF TIME 59 16 I 1 37 I I I 34 45 I I I I I 4 44 I 1 95 I I Oct 82 CRYPTOLOG Page 11 PeR 8PPfetAb 5SI l Olft Y DOCID 4009853 SE8RE'f 3' QO DER OL G 4L Reprint from DRAGON SEEDS June 1972 - Oct 82 EO 1 4 c 6 CRYPTOLa p ' I a 6 ItMHlbE VIA 88t1IUT 8IWRlSbS 9NbJPY p L 86- 3 6 EO 1 4 c - DOCID 1 4 a 09 8-5-3-- -------- J 'J ' -Il 3o - 888M' ' iI I Oct 82 CRYPTOLOG Iii SRB' ' Page 13 IIMfB bI n S9tHN' ' 8HA1iJlilU _ 4 0093 5 3 86 36 fyM g d k M43 Oct 82 Page 14 DOCID 4009853 Oct 82 Page 15 L 86-36 DOCID 4009853 i'8P 9B8Rf li' EO 1 4 c P L 86-36 SIGINT 1990 Part 2 u E L 86-36 OPTICAL FIBER What new problems will SIGINT have to face by 1990 What do the new trends in technology tell us about the not-so-distant future The author has adapted this article presented here in the second of several monthly installments from his presentation at a January 1982 session of CA-305 OPTICAL FIBER COMMUNICATIONS NEC U Initial applications of optical fiber were for specialized applications such as the communications circuits see figure 21 that were installed with buried power cables The fibers were unaffected by the powerful electrical fields and gave reliable wideband circuits Other specialized applications used the resistance to electrical noise and interference for shipboard and military applications despite the fact that neither the glass nor the electro-optical components were very good U Rapid advances in glass technology and in laser and LED light-emitting diode and detector technology have occurred during the past ten years These improvements have made short-distance circuits e g within Oct 82 computer nets economically feaSible and the telecommunication operators are beginning to install short-distance and medium-distance up to a few kilometers circuits in parts of the existing networks U By 1990 better glass and more reliable electro-optical components will make long distance trunks economical and by the mid 1990's optical fiber loops will be introduced from the local switches to subscribers' premises By 1982 the various Bell System companies had already installed over 25 000 miles of optical fiber trunk A typical application will be the Northeast Corridor a 400-mile trunk from to Boston using 3C digital Washington transmission at 90 Hbps fiber Some fiber systems will use three separate light wavelengths at once to give 270 Hbps fiber Typically an installed fiber system will have 12 fibers with repeater spacing at 35 kilometers Since coaxial cable repeater spacing is CRYPTOLOG 'P8P SBBRBi' Page 16 IhI NB B VIA B9flHf'P 8IlMfNB S 8N' DOCID 4009853 c I 'Pep SBaRB'f 10 10 aj bO l I QJ QJ 10 8 QJ 'IS c - 10 6 e' II J V o et 10 4 10 2 Special applications Power suppliers plants etc '--- -1- --1 'tn -1- 1- _ OPTICAL FIBER COMMUNICATIONS Low Density polyethylene copper-Clad Steel Wire CORE DE'II-IL CABLE UNDERSEA FIBERGUIDE CABLE Oct 82 CRYPTOLOG Page 17 1l1n'lbU nA eem li' ellfdflfl tJs 9NbY DOCID 4009853 'fep SHeRE'f about 4 kID improvement s and cost the optical fiber offers big in system design maintenance U NEC predicts a million kilometers of installed fiber by 1985 and 100 million kilometers by 1990 Since the Japanese are among the world leaders in both the technology and the marketing they are in a position to make this prophecy materialize UNDERSEA OPTICAL FIBER CABLE U One of the most dramatic illustrations of the advances in optical fiber is the development of a transatlantic optical fiber cable fig 22 to be operational by 1988 which will carry 72 000 two way voice channels between the U S and three foreign countries over a 6500-km undersea cable Transmission bit rate will be 375 Mbps per fiber and a TASI Time-Available Speech Interpolation will be used to triple the traffic capacity U Power will be supplied over a copperclad central steel cable and a dozen singlemode fibers will surround the core SG cable repeaters will be used with a regenerator for each fiber spaced at 35 kID The system is expected to have a 25-year life The first cable will probably be the start of a revolution in transoceanic communications just as the first transoceanic coaxial cable began the explosion of transatlantic telephony COST COMPARISON FOR UNDERSEA SYSTEMS U Analysis by BTL has shown that compared to the older 30KHz coaxial cable systems the 274 Mbps optical fiber cable will give a 5-to-l improvement in circuit costs The very high bit rate will make special services such as video video conferences and data services feasible U While satellite technology will improve over the same time period satellites are a stable technology while optical fiber is still going through explosive growth in performance The critical choke point in satellite communications is at the midocean points where demand for spectrum and services is growing while orbital and spectrum resources are fixed The development of intersatellite links will relieve this problem to some extent but in the long term the transoceanic capacity of optical fiber submarine cable--and its resistance to antisatellite weapons and electronic warfare--will make fiber the primary transoceanic medium he impact on SIGINT of the development of transoceanic and overland optical fiber trunks is that traffic which now must go primarily by satellite will disappear onto fiber The extreme publicity given to SIGINT over the past eight years by the revelations of World War II COMINT and by the excoriations of U S and British SIGINT agencies by governmental bodies by journalists by former employees and by COMSEC and cOimnunication protection advocates has made the telecommunication authorities highly conscious of satellite interception and microwave interception Optical fiber will be a preferred transmission medium because it is so difficult to intercept U While the AT T undersea fiber trunk will use digital regenerators the Japanese are now proposing a 5000-km undersea optical fiber cable with no regenerators which will use 1inear optical amp 1ifiers along the fiber This will allow changes to the digital technology at the cable ends as technology improves Experiments have already shown that power can be sent along light fibers to operate remote devices so a completely optical system with no electrical power or copper wire should be feasible Oct 82 Ne'f aBhShShBbH 'fe eeN'f S'feRS CRYPTOLOG ap SfleMl'f Page 18 Ihltlfet fA eellflf'f en lflffltS aNt t P L 86-36 EO 1 4 Ie DooID --'- ------ ---- - --- 'Pep BB6RB'P P L 86-36 EO 1 4 c U One oltha'main attractions of optieal fiber Jocal nets from the point of view of the PTT's is that over the air broadcasting could be almost eliminated and this would put all information flow completely under the control of the PTT's BIGFON OPTICAL FIBER LOCAL LOOP GLASS U A number of European countries are now experimenting with applications of optical fiber to the local network The BIGFON network see figure 23 will be tested in several German towns over the next few years to determine how various subscriber services are used An optical fiber pair with a capacity of hundreds of millions of bits will run from the local switch to the subscriber premises and will carry various two-way services including telephony television facsimile data communication telex teletex and stereo reception At the local switch selected channels of video stereo etc will be connected to the individual loop Current estimates are that a fiber pair can carry three TV signals and a variety of other services U Glass qual ity and cost are important considerations since the total amount of glass fiber will be large replacing the copper wire loops of current local nets Optical fiber glass is made primarily by two processes viz the crucible method which gives a cheap and stable method of producing low-quality glass 10 dB attenuation km and the MCVD multiple chemical vapor deposit method which gives a very flexible technique for producing high quality but expenaive glass 0 2 dB km loss Expert opinion Midwinter expects industrial techniques to produce a stable low-cost glass at about 1 dB km which will allow fairly long local loops The better glass would also allow economical higher bit rates j The effect of low-cost optical fiber local networks would be twofold viz a much wider range of services including videophone and conference nets with high grade voice security at 64 Kbps would be available in the local nets and the plant would be cheaper and more reliable than the conventional copper wire networks leading to expansion of local urban telecommunications services in less affluent countries Oct 82 PURITY U The driving factor in the development of optical fiber communications has been the improvements in glass purity In 1966 the first article proposing monomode waveguide operation of glass fiber was published but the glasses available at tha t time made the proposal appear absurd because the attenuation was too high for any useful system However blocks of pure silica were found commercially available with losses as low as 10 dB km and this spurred enormous progress in glass chemistry U The crude glasses of Egyptian times 3000 B C had attenuation losses of 10 million dB km and over more than four millenia glass purity was gradually improved to the quality of Venetian glass of 1500 A D with losses of 10 000 dB This repreaented a significant decline Over the next 450 years through the development of modern chemistry and industrial quality control a further improvement was made to give the optical top quality glasses of 1970 with losses of 1000 dB per km Then suddenly in a decade of explosive development completely new methods of purifying glass and drawing it into fibers were invented with the result that a 10 000dB improvement in glass attenuation was made in only ten years U The ratio of 10 000 dB corresponds to the number 10 raised to the power 1000 or a 1 followe by 1000 zeros a truly phenomenal improve Dent U To visualize the effect of this improvement on optical communications if all the power generating stations in the world converted all their power into light with perfect efficiency and these many gigawatts of light were transmitted into a glass fiber with 1000 dB km attenuation then an observer one kilometer away would have to wait over a trillion years for the first photon to emerge The glass fiber of course would be vaporized in an instant by the light absorp ion By contrast modern glass at 0 2 dB km loss can transmit a few milliwatts of light 100 km without a repeater at high bit rates over 100 Mbps CRYPTOLOG 'Pep BBe Page 19 IWfBI JB 'II' 8em ' 8lhtdftiBb9 lib'l ' OOClO 4009853 'PElP SHeRS'P BIGFON broadband integrated optical fibre local telecommunications network Telephony r--'_ _'_ --' - - o - Subscriber line 1 tiber subscriber up to 5 km length star network I x ' W ' -- 'I CONTROL UN IT i nttwork i - --' I I I I o 1-- - --- trunk ISDN I ATV L-- SWITCH i - Telex Teletex Stereo reception S H o - I VIDEOPHONE --------_ -t Data communication II 0 I I i Picture teleph ony Television Facsimile transmission - e FM REO WI SWITCH L _ _ _ _ TV FM- STEREO - progroalllls various subscriber 11ne transmission-methods - optical vavelength-diviaion multiplex time-division -ultiplex frequency-diviaion multiplex apace-division multiplex jig ISS' 81 SDM ZJ ere MonJrlta 21-25 sept Oct 82 WOK TDM FIlM CRYPTOLOG 'PElP 888MT 1981 Page 20 HMIB 8 VI' ssunl' Slh'dfNB 8 Stl JPY DOCID 4009853 ep U At the same time that the glass has been improved the light transmitters and detectors have also been improved especially in matters of lifetime and reliability A few years ago the lasers and light emitting diodes had lifetimes measured in a few hours but this has been improved to 100 000 hours with further improvements to more than a million hours expected Efficiency has also been improved so that very low power consumption will keep a long series of repeaters operating This is important to submarine cables where the repeaters are inaccessible U A comment about fiber manufacturing is in order because it implies certain limitations on SIGINT operations against different fibers The manufacture of high-quality optical fiber is done primarily by two processes viz Modified Chemical Vapor Deposition MCVD and Vapor-phase Axial Deposition VAD Both MCVD and VAD produce a large glass rod that is subsequently drawn into very long fibers in a high-temperature furnace A subtle chemical process known as thermopheresis is used to allow various chemical gases to penetrate into the hot glass while it is a tubular form turning on the lathe and this process captures unwanted molecules and also deposits desired chemicals in a systematic way The glass tube is then collapsed into a rod and drawn into a fiber The lower-quality optical fiber is produced by a double crucible process in which a mixture of chemicals is put into two crucibles and melted giving two kinds of glass of uniform composition but not as pure as HCVD yields Glass from one crucible is allowed to gravity-feed into a thin fiber and this is drawn through a gravity-formed tube of an outer glass still in molt n condition to form coaxial fiber The light travels in the inner fiber and is reflected by the outer cladding E8RS'f best glasses the losses of light energy in propagation have been greatly reduced but some losses still occur from a phenomenon known as microbending in which the light rays hit the interface between the inner and outer glass at a high enough angle so that some energy is transmitted into the outer cladding and escapes The fibers are currently stored inside small-diameter plastic tubes so that they are not bent too sharply as the cables are laid around curves The critical angle seems to be about six degrees At higher angles of incidence microbending losses occur In the design of the transmission systems an allowance for such microbending losses is provided The geometry of light propagation down a coaxial fiber is very complicated because of coupling between different propagation modes so that the exact light path is unknown In the graded-index fibers produced by the HCVD process an axial focusing and defocusing of light takes place but even that propagation is very difficult to describe The result is that in general light energy is launched into one end of a lightguide and some of it emerges at the other with random losses due to absorption escape etc 24 HICROBENDING U Because of the very high purity of Oct 82 Me' R8b ABJ iS 'F9 89NH 'S'f9R8 the CRYPTOLOG ap iGU r Page 21 H llmlsK VIl SaHliA' SIwmBb5 eNtN P L 86-36 EO 1 4 c DOCID 4009853 'PElP sllsae'P --ts't The central aim of optical-fiber SIGINT would be the concept of proven reserves derived from the petroleum and m1n1ng industries rather than the current journalist ic concept of hand to mouth ilDD1ediate exploitation of whatever is easiest to get or fits current consumer requirements P L 86-36 1 4 c lEO 25 MOLE Oct 82 CRYPTOLOG YO SI 8Rfi'f Page 22 I i NBhB VIA eEltIlN'f 81Wftfl bS 6fH iJPY DOCID 4009853 _ _ _ _ _ _ _ _ _ _---1 -- I I 'fOP SE6RE'F systems this after the fact access is not a workable scheme If it took several years to get access to several fiber cables in an urban area a crisis could come and go before any traffic could be collected 5 The proven reserve concept does not only apply to interception but to analytic and exploitation capabilities as well so that the resources to attack and exploit systems should be developed and proved before there is a desperate need Arguments that this is unaffordable should be evaluated by looking at the enormous success and wealth of the oil and mining companies who do find this system affordable affects the entire SIGINT activity rather than just hardening one problem at a time -- 5 QCQLThere is a historical precedent for this in the World War II context where the German traffic security which at first thwarted Allied efforts at efficient collection was solved for most of the war and theop L 86-36 at the end of 1944 became so secure that Eo 1 4 c high-level cryptanalysis on the ENIGMA problem almost came to a halt because the traffic networks could no longer be identified or analyzed EVOLUTION OF EUROPEAN SWITCHING U In the keynote papers at the Internat ional switching Sympos ium in Montreal in 1981 ISS 81 the rapid shift of the telecommunications plants of the Western nations from hard wired electromechanical switches to digital electronic switches was described Although the old fashioned switc hes had been designed for a 30- to 50-year amortization there is a growing trend to earlier replacement because the new switches are cheaper to operate and make more efficient use of the network As a result some 35 percent of subscriber loops will be connected to digital electronic switches by 1990 and 65 percent wi11 be tied into the digital electronic switches by 2000 As the new switches replace the old the impact on SIGINT will be that the networks will become more flexible and efficient and less rigid in the way traffic flows and in the services they provide Some of the digital services including those that use end-to-end encryption will also have an impact on SIGINT BEL L SYSTEMS STORED PROGRAM SYSTEMS IN SERVICE I laM SWITCHING _ 26 BELL STORED PROGRAM CONTROL SPC SYSTEMS IN SERVICE Oct 82 NO'f REfJEA5ABfJE 'FO 60tl'fR A6'fORS CRYPTOLOG 'FOP SE6RE'F Page 23 IJ I NBbE VIA 6OtfHl'f 61fANNEbS ONbY DOCID 4009853 'lap SlleRIl' ' U In the U S the Bell System has also been very active in building and installing computer-controlled switches Currently 47 percent of the local subscriber lines are covered by SPC switches and the coverage in the Bell System by the early 1990's is expected to be 100 percent There are now 2800 SPC systems in service U The SPC switches allow the network to provide new services and also allow more efficient flow of traffic between switches 27 SWITCHED DIGITAL CAPABILITY able The CCIS information may flow over coaxial cable where the traffic itself flows over analog radio digital radio or satellite circuits -rst-rhe new Switched Digital Capability is an example of the new kinds of services which the digital electronic SPC switches will provide In addition to the encrypted voice encrypted high-speed facsimile and electronic mail and encrypted data services will flourish Because the CCIS or other Common Channel Signalling will separate traffic addresses and routing from the traffic itself the wide band transmission systems can be efficiently filled to capacity with a continuous stream of bits or in the case of analog radio links successive talkspurts without any indication of who the sender or recipient are The toll switches are also capable of instantaneous automatic rerouting of traffic without any break in service so that a given message or session may flow over various different channels The effect is that the switches tend to act as transposition scramblers on relatively featureless and unidentifiable analog and digital traffic The dedicated or leased channel will be a bookkeeping notion rather than a oredictable physical circuit I One of the significant changes in local services which the new switches will provide is high bit rate traffic over the subscriber loops These high bit rates are made possible by the improvements in channel equalization and by sophisticated modems that can be implemented cheaply with the high density microelectronics Some of the European nets are experimenting with 80 Kbps and 140 Kbps links and the Bell System will offer 56 000 and 64 000 bps between the subscriber station and the local switch Encrypted voice as well as facsimile and other data services will be available throughout the network where lA ESS or No 5 ESS switches are used The Switched Digital Capability SOC will be introduced in 1984 For some subscribers optical fiber will be used in the local loop The 56 000 bps service will be end-to-end and will obviously provide high quality encrypted voice without the distortions that speech coders generate This should lead to a rapid growth in encrypted voice for business traffic with DES and Public Key interconnection between subscribers U Between the local switch and the toll switch wireline or optical fibers will be used under control of Common Channel Interoffice Signalling CCIS Coaxial cable analog radio digital radio and satellite links will be used at the toll switches No 4 ESS and CCIS signaling will be used where it is availOct 82 Digital Switching No 4 ESS INTRODUCTION 1 18 78 067 offlcoo now In erYlce e Characteristic o Time dly' 'on wltchlng network o Powerful central proce sor o New technolog o Full duplication o Disciplined oftwa e Iilethodolog eFeature o Large capaclt o Wide range of SPC network fe tures o ExtenslYe O A M feature 28 No 4 ESS TECHNOLOGY U At ISS 72 the No 4 ESS created a sensation as a novel and ambitious functional and technological development It was introduced into service in 1976 and 67 offices are now operating However the hardware technology of the switch has been completely replaced because of the technical superiority of newer microcircuits No 4 ESS uses time-division CRYPTOLOG 'l 9P 8ReM' ' Page 24 IIMfBl5 VIA S8UIlIt 91Is VRR3hS 9UhY P L 86-36 EO 1 4 c DOCID 4009853 'fe seene switching in its internal logic to give it very high resistance to blocking i e refusing a connection where a path actually exists in the switch U After the first shocking experience with switch software in the No 1 ESS the Bell System turned to disciplined software methodology and now they can produce software which does not cause frequent switch outages In spite of this a new switch software product must run in for several hundred switch years before all the bugs are removed plan has been set to provide a size range from very small to large offices of at least 100 000 lines with a large complement of customer network and administrative and maintenance features U The flexibility of the No 5 ESS and its production in various size ranges is typical of the new developments in switching The main producers of electronic switches are the French Japanese and L M Ericsson The Swedish company produced a modulaiizedAXE switch in the early 1970's wtlich in its original form routed signa1siri analog fon but was converted block by block to all digital operation The first AXE switch was installed in Sweden i01918 but has sold well abroad Outside America there are only about 100 PTT customers for switches The attraction of the AXE was that an analog network could be gradually converted to digital and this technical lead enabled L M Ericsson to win the biggest telecom contract ever awarded for a $5billion Saudi telephone system The conversion of existing analog networks over to digital operation which involves the integrating of new switches into mixed analog and digital networks is one of the most demanding problems in switch and network design ADVANCES IN TECHNOLOGY FOR ESS o o o o U The No 5 ESS is a new local switch designed to handle up to 100 000 lines and designed to be used in rural areas This switch applies time division switching to local nets In contrast to the No 4 ESS which is bssed on a powerful central processor the No 5 ESS architecture is based on distributed control Powerful microprocessors are used in all of the peripheral modules while the central processor performs more globsl control functions as well as overall administration and maintenance of the system U The switch designers have found it necessary to move into the forefront in development manufacture and application of the mo t modern hardware design and software technology The network functions call for very complex logic represented in VLSI hardware and software Now the switch designers are converting software functions into firmware where the critical algorithmS are designed in silicon from the start U No 5 uses modern software concents and all elements common to 512 lines are duplicated for reliability An evolutionary Oct 82 VLSIMicroprocessors Memory Custom logic Other Hardware l Optical fibers High-voltage semiconductors Display technology Software Architecture l Operating systems l High-level language Development methodology CAD Program development support l Automated testing CRYPTOLOG tSP SH8Rflt U Because of the high speed of modern Page 25 WWgU HI z SQHII Qllt HNHhS Slfh'l P L 86-36 EO 1 4 c DOCIO TOP 8l iSRl iT computers various real-time telecommunications functions can be codified in higher level languages Software has become a major part of switch and telecom plant production At Bell Telephone Laboratories in 1950 no software was produced Now half the staff works at software production The advanced programming techniques enable software production at 100 times the rate of 15 years ago The testing and debugging and particularly the reliability of modern telecom software is a major technical factor The other side of this coin is that once the software is designed and tested it is very hard to change because hundreds or even thousands of switches and transmission systems are made interdependent and interconnected by the software Hence even if the hardware is replaced and parts of the software are codified into firmware the basic software system will probably have a long operational life The software programs are often very large running into hundreds of thousands of lines of code and the developers are frequently protective of their source level program scripts COMMON CHANN L SIGNALLING U In Common Channel Signalling the addressing information used to specify the call and the routing is not sent as a pr fix on an idle channel Instead a separate dedicated channel is used to c rry all the addressing and routing and signaling data The resul t is that thetransmiss ion fac il ities are used more efficiently because idle channels do not have to be held open f r some minutes while successive channels are found and then until a called party finally picks up the telephone On long distance calls signaling and call setup used tp ake about half the channel time The a lication of common channel signaling allows network efficiency to be practically doubled without building new plant ' U Three major systems are No 6 CCITT No 7 CCITT and the Bell System CCIS Common Channel Interoffice Signalling Oct 82 NQT RSr S ' S ' BMi '8 8QU'l'lb 8'f8RS CRYPTOLOG TQP 81SMT Page 26 lIlJiIBl5E ' ll Set SIIIo NlIEI aS eUr iY 'P L 86-36 EO 1 4 c DOCID 4009853 ' gp SI iQIHl'F an SPC computer into existing crossbar switches so that the crossbars can be controlled by common channel signaling This will enable the entire French switched network to be operated by common channel signaling The PUCE retrofit will allow the existing switches and transmission facilities to be operated more efficiently without the expense of replacing the existing large investment in electromechanical equipment This will also prolong the life of the cr ssbarplllrtf A plan publishedby CNET shows the effect of PUCE in extending the life of crossbar switches P L 86-36 EO 1 4 c CCIS NETWORK PACKETS PER BUSY HOUR x 10'J JOO ----------------- 31 30 eCIS NETWORK PACKET TRAFFIC FORECAST rUHCH SWITCHING EQUIPMENT ASSETS U The jor Irovth will be in timedivi ion electronic witche temporel with rotary and manual pha eout by 1983 and ordinary crooobar dwindlina after 1984 The crooobar with the PUC retrofit will remain at a fairly steady level up to 1990 presumably as a result of converting the ordinary crossbars U In the Bell System Common Channel Signalling is already in use between some switches and a packet network is employed to transmit the switching data Currently some 20 million packets are sent during a peak hour By 1990 the packet traffic rate for CeIS will rise to 700 million packets per hour The dedicated CCIS packet network operates to set up the circuits switched network that will actually carry the traffic U While the new digital switches can be designed to implement common channel si nal- in the older electromechanical switches e crossbar are desi ned around a control svstem that uses in-channel orefix si2nalin2 The French have developed an innovation for this problem called PUCE which will retrofit Oct 82 U Summin up the switchinR trends timedivision has the 2rP-RteRt 2rowth ootential and the wide use of time-divi ion di2ital switches will reduce switchin2 from 50 percent to 20 percent of total plant value but many CRYPTOLOG Page 27 IIldfB'b1 ' liz 8eUIN'F ElI NKl S gm JPY j i DOCID 4009853 'Pep SHSRe'P years will pass before most countries can afford this On the other hand the low cost and high reliability of microprocessors and computers will make it increasingly attractive for PTT's to seek cheap retrofits and hybrid systems which will give more efficient operation and more or newer services without the expense of replacing still serviceable equipment U In most poor countries the problems of maintaining outside plant will be as significant a factor as shortage of money in retarding the successful introduction of digital electronic switches defects and noise in the outside plant subscriber loops and trunk transmission cause errors that can disable the more critical digital technology Another key factor in the introduction of digital switches in any country is the distribution of timing standards because close synchronization through the networks is needed This is already a problem in the U S where different suppliers provide equipment to hundreds of independent operating companies Once a small country starts to introduce digital switches the difficulties of time distribution and digital interface will cause special problems for any rival manufacturer who wants to sell in that market Oct 82 ue RElbHi SliBI JH 'P8 S8tRlli'tS'P8RS CRYPTOLOG TOP SKeUT I I Page 28 L 86-36 1 4 c DOCID 4009853 SBSRB'f SPSU I I I I I I I Not Secret Anymore P L 86-36 May I add a few words to I I reminiscences in the August 1982 CrypI tolog especially to his comments about the I initials NSA standing for Not Secret I Anymore Like Brotherl I recall those days of M M as we lovingly called Martin Mitchell and I also remember figuring out that expansion for myself but 1 doubt if I was the only employee to come up with the clever meaning for NSA II ll I ll I remember including a cartoon in my book NonseNSA showing President Eisenhower denying everything that M M had said ending up with and there is no such agency all NSA at Fort Meade Maryland In fact there is no such place as Fort Meade Maryland and scribbling on the bottom of the page NSA now stands for 'Not Secret An Qrel' n fortunately I loaned NonseNSA toL J who used 2 or 3 cartoons from it in early issues of Cryptologic Spectrum and then retired without returning it to me so I don't know if the book is still in existence ANSWER An Old Problem U CRYPTOLOG August 1982 ll But at least I'm glad to see that somebody else used Not Secret Anymore since my other great discovery about what an agency's initials meant seems to have been restricted to me alone When President Reagan appointed William J Casey to be the head of CIA and then got our Admiral Inman to be his deputy I remarked on several occasions that the CIA was the Casey-Inman Agency which I thought was pretty clever--in fact I was tempted to send it in to The Ear at The Washington Star--but to date I've never heard or seen it used by anyone but little 01' me ooo and now that Admiral Inman has left I doubt if I ever will o _ _ _ _ _ _ _ JI P16 When you finish send up a flare P L 86-36 EO 1 4 c P L 86-36 Oct 82 o CRYPTOLOG o Page 29 SBeRe spon PI-Oct 82-53-21537 I DQi ooo o o _pog dfiE'PB53 - - v -----r r ---- TIllS B9EUR1 f 1BllIT T 69NTzYNB 699BW9RD AU T This document is from the holdings of The 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