GJUlUV0GJUll1 5lBVJWVU Ul 5GJlB lrW U 15W 15 W15 fil15 f W l1 Wfil I Brn urnl1rn 3 JU 300 D 3J iJ0Bf2 P L I ----Ih o 1 '---- V NORMANDY 1944 U oooooooooooooooooooooooooooooo TSS REVOLUTION U ooooooooooooooooooooooooooooo SIGINT 1990 Part One U oooooooooooooooooooooo1 86-36 8 1 13 LETTERS U oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo 29 MORE FREE GOODIES U oooooooooooo ooooooooooooo ooooooooooooo oooooooooooo 29 TillS BOeUMHN'F eON'FhINS C098WOK9 l ll T8KI U Classified hj NSA CSSH 123 2 Qslas81fy ea Q 181aat1ag A@easy's eteFm1a8tlea Re lEee Declassified and Approved for Release by NSA on '10-'1 2- 20'1 2 pursuant to E O '135 26 vl DR Case # 54778 DOCID 4011953 Published by PI Techniques and Standards VOL IX No 9 SEPTEMBER 1982 Editorial PUBLISHER BOARD OF EDITORS C Edit -in-Chief Asst Editor oooo Production ooo 1 ' 8321 7119S _ _ 1103s 3369s Collection oooooooooooooooo 1 Cryptanalysis oooooooooooo 1 Cryptolinguistics ooooo 1 Information Science 1 Language Machine Support'1l Mathematics Puzzles oooooooooooooo David H Willia lls Special Research ooooooooo Vera R Filby Traffic Analysis ooooooooooo Don Taurone 8555s 5311s 1103s 57lls 8l6ls 468ls 85l8s K1103s 7ll9 35738 For subscriptions send name and organization to CRYPTOLOG PI or calli 369s The purpose of CRYPTOLOG is to help you to do your tasks Our emphasis is primarily technical but we don't always limit ourselves to the purely technical Helping you do you job might be achieved by showing you how to do something or by letting you know that someone else has worked on something similar It may also be done by making you more aware of what is going on around you both in space at the other end of the building and in time last year on a related problem because many of the best technical people are basically curious people For some of you CRYPTOLOG offers an opportunity to stretch yourselves by writing you might be surprised at how few readable writers there are in certain fields whose output is intelligible to readers outside their own territory Developing a reputation as a writer could give your career a boost If you are thinking about writing something and are waiting to be asked call me and I'll ask you To submit articles or letters via PLATFORM mail send to At the end of the firs t year as ed i tor I ought to have some profound thoughts but I don 't The magazine is coming out more or less regularly each month and that was our first goal Some of the other goals have been met including the matching of the distribution against the locator file Some goals remain to be met I cryptolg at barlc05 bar-one-c-zero-five note no '0' in 'log' P L 86-36 In years past CRYPTOLOG often skipped a However as an experimonth in the summer ment we are offering instead an issue containing three somewhat longer articles Their subjects interestingly enough are the past the present and the future FeR eFFIe 11ft 'HIE eliLt DOCID al1953 TOP Sl eM T tftlllfblt NORMANDY 1944 U byl -' P UB REVIEW Six Armies in Normandy Keegan Viking 1982 NY by John U American English Scottish Polish French and German armies clashed in a decisive land battle in Normandy from June till September 1944 in one of the major campaigns of modern history Half a million German troops were lost most of them killed their 50 infantry divisions and 12 panzer divisons had been reduced to 24 infantry divisions at quarter strength and 11 panzer divisions had shrunk from 150 to 10 tanks each Even the destruction of the Army Group Center in June and July 1944 by 140 Soviet rifle and tank divisions which had cost 300 000 German soldiers was a lesser defeat The Western Allies had committed only 34 divisions to the battle against the 62 German divisions When the Germans finally lost in Normandy they lost all of France and fled in disorder to the Dutch frontier U John Keegan a lecturer at Sandhurst whose previous book The Face of Battle became a minor classic has written synoptic accounts of six different divisions in their most critical battles in Normandy to illustrate not only the key events in the battle seen from battalion down to squad level but also to illustrate something of the national character of the different armies Where did the troops and weapons come from Why did they fight How did they fight How did they react to the stress of 1944 warfare Sep 82 P L IPI3 iLl d _ The salient advantages the Western allies had in the land battle were seapower airpower plentiful supplies and COMINT As a result the Germans had to defend everywhere and this left them too weak at many points and unable to move forces to the battlefield When they did try to move the radio communications were deciphered in time to paralyze the movements and prepare defenses The cardinal contribution of COMINT was that it allowed the Allied armies to defeat the German Schwerpunkt tactics by giving prior knowledge of intentions 6 In general the Allied armies were no match tank for tank or battalion for battalion with the first-line German divisions General Alan Brooke British Chief of Staff considered the Germans the best soldiers in the world and tried to keep the British forces away from them by a peripheral strategy while the Americans sought a direct cross-Channel attack to destroy the German Armies in the field Normandy was the result of this prolonged dissent on Allied strategy U The British had sound reasons for regarding the German Army with caution In World War I the German Army was the only one in Europe that did not have mutinies It marched out of the Rhineland in 1919 in perfect order still willing to fight In World War II the German soldiers fought until the last few days of the war and Germany had to be invaded and crushed they would not surrender as long as they could still fight At Dieppe in 1942 an invading force of 5000 Canadian CRYPTOLOG Page 1 8F SeeRS l JIIBRA 86-36 DOCID 4011953 SP SSeRS tfllBR r V At the end of June the Panzer divisions were committed to the battle and the 9th and lath SS Panzers who later destroyed the British and Polish parachutists at Arnhem 3 were taken from the Russian front and moved to Caen where they smashed into the 2nd Argyll and Sutherland Highlanders in the Scottish Corridor west of Caen An ULTRA decrypt revealed the assembly area of the German attacking force and naval gunnery destroyed the German armor before it could mount its attack in force This COMINT assist was helpful for the Scots had no tanks and only sixpounder guns and PlAT antitank weapons for their inexperienced infantry in their first battle to fight SS Panzer soldiers who began their war in Poland and had spent several years in armored warfare in Russia soldiers attacked a coastal town defended by a garrison of 400 German soldiers and nearly all were killed in spite of 12-to-1 odds The German Soldiers in 1942 were in France training for or rehabilitating from the Russian campaign In 1944 many of the defending divisions at Normandy were bodenstaendige divisions equipped with captured foreign weapons or other inferior equipment without transport and the soldiers were old or unfit Thanks to the excellent Allied intelligence most of it from COMINT although Keegan does not mention this the initial landings were aimed at these relative weak spots V The German command system was able to react very quickly by converting spare personnel into infantry formations and putting weak units in front of SS regiments to stiffen their will to fight German officers did not wait for orders from above but put their forces into battle very skillfully in anticipation of orders from above Orders gave objectives rather than detailed directions so that German reactions were tailored to the battle situation--rather than premeditated The troops obeyed orders to the end because they had confidence in the Officers 13 The German Army ended the war with only half its billet ted officer strength because officer casualty rates were much higher than enlisted casualty rate and the Germans would not commission unsuitable people because it would have destroyed confidence in the Officer Corps as a whole 4 Sep 82 V Because the Germans had no navy left except V-Boats and no effective air force they were unable to drive off the supporting task force which fired on coastal targets for weeks with long-range indirect fire These Allied assets were not always well used however for the helpless city of Caen was reduced to rubble by naval gunfire and then further devastated by a massive strategic bombing operation because Montgomery did not want to expose his troops to heavy losses by the Germans The Germans were not in Caen so there was no military result The French trying to survive in cellars were buried under more rubble An insane asylum used as a refuge and hospital in the outskirts of Caen was devastated by Allied bombing The Germans however had to keep huge formations tied up defending coasts from Norway to Italy because they could not match Allied mobility and force at sea or in the air As soon as they tried to move forces COMINT read their orders and the transportation facilities were wrecked Hence Normandy was as much a war of interdiction as it was a war of attrition or maneuver V For a representative American division Keegan chose the 101st Airborne Division which landed by parachute and glider before the invasion to protect certain bridges so that the trQops at Vtah beach could get inland The aircraft navigation and the formations were thrown off by clouds so that the troops were scattered allover the Cotenin peninsula and some were dropped in the sea With typical staff planning the communication equipment--essential to coordinate and regroup the scattered airborne troops--was dropped separately from the troops and virtually all lost As a result almost all the operations were impromptu little clashes by tiny packets of troops Ad hoc formations and chains of command were set up as circumstances CRYPTOLOG Page 2 'PSF S BeMJ'P tfllBlbIt DOCID 4011953 dictated and sometimes several different elements attacked the same target at the same time without knowing who the others were Runners were used to find higher and lower units and deliver messages Small groups of paratroops were adrift for weeks out of touch with everyone Many units vanished presumably exterminated by the Germans Virtually no one spoke French and the local population was reluctant to get involved in the combat The resulting disorder apparently confused the German command who could not decide where the point of the attack was The results were bloody but successful with very few prisoners Keegan compared the American airborne troops to pioneers ignorant of the language landmarks or dangers around them who simply relied on their own weapons and skills The metaphor is apt for the American troops frequently outfought the German troops in the light infantry battles and never waited for orders or refined intelligence General Maxwell Taylor was cut off from his command and from the invading force until after the invasion forces had gotten inland but the lower level elements achieved their objectives anyway U It has since been said that Germans could never understand why they lost a battle to the Americans at any point in World War II because they tended to perceive American attacks as disorganized and confused construction which were wrecked in one night by a storm after which the Americans unloaded massive quantities directly onto the beaches and did far better than the British achieved with their rebuilt Mulberry harbor 18 very U German defense was frequently tenacious At Carpiquet on 4 July 1944 two Canadian regiments fought all day to take an airfield from 50 young soldiers of 12th SS Division In Operation Goodwood in mid-July a massed formation of British Guards armored divisions attempted to break out from Caen following a raid by 1000 bombers Battle Group von Luck of the 21st Panzer division threw the attack offstride by hitting its flank without any orders from above and then half a dozen Tiger tanks blocked the advance of a Guards Armored Division Soon the First SS Panzer Division Liebstandarte Adolf Hitler which had been in combat since Poland brought the British attack to a halt Montgomery was severely discredited by this defeat The British Guards divisions were actually Horse Guard regiments from the yeomanry and gentry of England who reluc tantly gave up their horses and switched to tanks as World War II become imminent British tanks were unreliable and no match for the 1930's-era German tanks but the British fitted a new gun and turret to the American Sherman tank and the result was a fairly good tank However it could not defeat the Tiger tank which the Germans used defensively in Normandy The Guards regiments had experience in desert fighting but in the terrain of Normandy the German tactics of using 88-mm antitank guns in conjunction with tanks that fired from cover proved superior At the same time the policy of fighting on a rigid line as ordered by Hitler and required by the' Allied interdiction of the German rear area left the Germans vulnerable to the breakthrough and encircling operation that occurred in late July and August American military operations are mentioned elsewhere in the book from the preinvasion buildup in England where Keegan as a schoolboy was impressed by the wonderful cornucopia of vehicles and equipment the Yanks brought with them to the breakout at Avranches and the sweep across France The Americans were much more willing to take casualties than the British because they had not felt the devastating effects of the trench German warfare of the First World War By 10 July 88 mm 1J70 the Americans had suffered 40 000 casualties Weight 4 tons still short of St Lo Then they lost 11 000 Muzzle velocity men in 5 days taking St Lo On the third 3 700 fps Length try carpet bombing was used to break the thin calibres 71 crust of German defenses infantry crashed Weight through the stunned defenders and armor ofshor 22 21bs poured south into Brittany American footbaIlers would call this a power play COM f INT revealed the exact details of the German t - -- j Co ' ' ' ' response and the American defenses pulverized I - 'y_ the counterattack Then masses of rocketC 7 i ' o -_o o _ firing P-51's kept Patton's flanks clear as _r -l i - 1 J 5 jl 7 t his armor swept eastward toward Metz where a J - __ '-' J tiny German defending force stopped him in hi S ' o J '- -'f' -' i' I tracks 16 18 The Americans also - 0 '- ----' j did well on logistics The British J ' ' - ------ __ - ' ' favored complicated gadgets such as ' J - t- the Mulberry harbors two years in T-nfT r r - -'5 1 1 ' -' ' - Sep 82 CRYPTOLOG Page 3 'SF sSeRS' lIBRA tH ' DOCID SCALE OF MILES o I O itS 'Fep SBSRB'F UlIBRf 4011953 100 U Keegan chose the landing of the North Shore regiment of the Canadian 3rd Division at Juno Beach on D-Day and contrasted it with the slaughter of the Canadian 2nd Division two years earlier at Dieppe The Canadians generally played an unsung part in the war Their battle doctrine was British but much of their equipment and logistics were American particularly the indispensable tanks and trucks Canadian regiments trained for four years before they fought in Normandy and had to compete to be selected for overseas service 12 Because of the long training and association in the same battalions and regiments the Canadian forces were able to withstand the savage battles with the SS Panzer Divisions from Caen to the Rhine By contrast the U S stripped 60 000 cadets out of flight school in Summer 1944 and 75 000 from advanced training programs and shipped them to France for immediate combat duty as infantrymen 15 Canadian losses were high a single 2nd Division Regiment the Algonquins suffered 1300 casualties about one third of its strength from August 1944 to May 1945 12 Similarly the U S 35th Infantry Division lost 3000 men in a few days in the St Lo breakout 17 U In an effort to trap the Germans in the Falaise pocket the First Polish Armored Division got onto a hill near Chambois and fought off repeated German attacks although it was itself cut off At this point the Polish Home Army in Warsaw was facing annihilation by the Germans while the Russian Army paused not far from the city The Polish parachute briSep 82 gade went on a hunger strike because the Western Allies would do nothing to help the Home Army in Warsaw The First Polish Armored Division had risen from the ashes of successive defeats After the invasion of Poland in 1939 100 000 refugees got to France and an army in exile was formed in France Only 17 000 got to England with no equipment after the fall of France in 1940 They recruited Poles from around the world and manipulated American Lend-Lease to equip and train an armored division An AustroHungarian officer Maczek was put in command This re-formed army kept the titles and colors of various Polish regiments long since destroyed It formed part of the Army of the Republic of Poland rather than being merely a division in the British Army The Poles were notoriously disdainful of danger and when given an order to close the Falaise gap set off without waiting for resupply of fuel and ammunition They took the wrong road and wound up in an excellent military position on high ground ready to block the German retreat but cut off Twenty German divisions were trying to break out past them On the same night that the Home Army retreated through the sewers of Warsaw the Germans began to fight for control of the hill at Chambois The Poles' resupply was airdropped five miles away They fought without supplies or support for three days against massed German infantry tanks and antitank guns They also had to guard 800 German prisoners Seventy percent of the survivors of the Westheer who had tried to escape at Falaise had been stopped by the Poles and taken out of the war In Warsaw the survivors of the Home Army had to surrender to the Germans After the war most of the surviving Poles from the First Armored Division stayed in England their battle to free their homeland had been in vain U The French Army was commanded by General LeClerc a nom de guerre for Vicomte Jacques-Phillipe de Hautecloque LeClerc was actually much higher up in French society and military life than DeGaulle his nominal superior His 2nd Armored Division which fought its way into Paris after the Normandy breakout actually began its existence in Southern Chad in 1940 The cadre consisted of black Senegalese soldiers and outcast French officers from Fort Lamy This swashbuckling crew of officers and bloodthirsty soldiers soon made a trek 1000 miles north across the Sahel to the Sahara where they fell upon the Italians in Southern Libya In 1942 they made further attacks in collaboration with the British Long-Range Desert Group and then left the base in Chad to move into Tunisia They joined with some Fbreign Legionnaires and Levantine soldiers serving under Koenig They CRYPTOLOG Page 4 'Fep S8SR8'F l I lBlb DOCID 4011953 'SF Sl3eM' UUBlblz filled out the cadres with Algerian and Moroccan mercenaries and it was these Berber mountain troops who broke the defenses at Monte Cassino in Italy The Moroccan troops by some accounts were paid three centimes for bringing back trophies cut from the enemy corpses For political reasons it was important to recruit native Frenchmen to fight in the reconquest of France itself The Americans promised equipment for a full armored division Former Vichy troops and officers in Africa were permitted to join on probation The Senegalese were not regarded as suitable liberators of the white French nation and were sent back to Chad over their protests Christian Arabs from North Africa replaced some of the Senegalese The officers were largely St Cyr graduates all known to each other LeClerc himself an aristocratic hereditary leader indifferent to danger was the central element in healing the divisions between the Vichy and anti-Vichy officers and men French military pride had een deeply wounded by the ease with which the German Panzers brushed their armored units aside in 1940 so they worked diligently to be up to battle standard by D-Day They went to France on 1 Aug 1944 under Patton's command man defense in Normandy as a model for how NATO might fight a conventional war where the rear area was disrupted no territory could be given up supplies were scarce and air superiority was held by the enemy He does not think much of the idea for the Allies were able to use seapower to attack a coastal point without warning while a Soviet advance would have to follow some obvious land route However there are several factors in the Normandy battle that are worth considering even if the particular tactics are no longer pertinent The Allies with only 34 divisions defeated some 62 German divisions destroying half of them By the book it should have taken 180 Allied divisions to push the Germans out of France Despite the inexperience of the Allied divisions and the great tactical ability and experience of the German divisons such as Panzerlehr etc the Allied casualties were comparatively low Airpower was critical especially in interdiction and in protecting the Allied logistic flow but one key factor was the Allied ability to avoid slaughters U Keegan mentions COMINT citing Bennett's book Ultra in the West but his treatment is fairly sketchy ULTRA COMINT was generally of only limited tactical use because it was either too late or the Germans did not send detailed tactical orders over ENIGMA There were some exceptions especially when Hitler sent tactical plans directly such as the counterattack at Mortain after the Avranches breakout Many of the read ENIGMA messages were unusable because the British General Staff had thrown away the only copy of the French military map that the German Army was using which gave the names of many little towns 51 Reconnaissance patrolling prisoners and captured documents gave most of the timely tactical intelligence 14 The French Poles and Canadians were not direct ULTRA recipients U The battle for Paris was fought for domestic political reasons rather than for military reasons Eisenhower wanted to bypass the city but the underground Communist and maquis was too strong politically so it had to be liberated by the Free French in order to establish DeGaulle as a force in the French government The German commander Choltitz had been ordered to destroy Paris and intended to give an episode of resistance in order to salvage German honor before surrendering the city The French ran armored columns through the streets of Paris in three columns to reach certain politically important objectives Their aim was to establish DeGaulle's appointees as the heads of key ministries by fait accompli They did not want to destroy Paris and were criticized by the Americans for advancing on a one-tank front Within a day they achieved their purpose The Parisians expecting to be liberated by the Americans were thrilled to see the insignia and pompoms of French regiments A huge crowd turned out to rush along with them just as they hit a German strong point Battles raged around the Arc de Triomphe Place de la Concorde Hotel Meurice and other parts of a battlefield deluxe After the honor of the officers was satisfied the troops were ordered to stop ting A European civil war has its roto- __ __ r r eF---C'I -- o- 1'1 _ 1 -c -__ o ' ' - U In his epilogue Keegan remarks military analysts are now looking at the Sep 82 TOP CRYPTOLOG o Page 5 ii RE YIHlIh' - - ' n _ l __ l '''' -- a _' _ DOCID 4011953 TOP Sl eRBlf ijflillU se ULTRA did not set the strategy That resulted from a long conflict between the British and Americans over the second front described by Keegan in a chapter The plan to fight the German Army after a cross-Channel invasion set the conditions for COMINT which then had to work on the Army keys and other ciphers pertinent to the German defenses 10 ULTRA from many sources gave an incomparable picture of the German disposition fortifications and preparations for the invasion It also showed in 1943 that the Foreign Armies West Intelligence was channel blind and thus vulnerable to the strategic deception that tied major forces in the Calais area 8 COMINT on German SIGINT and Intelligence and Agent systems showed that strategic deception was working and spared the Allies from a devastating surprise in the landing 9 After that COMINT continued for several months to give a lot of operational intelligence about major ground and air movements and high-level reports so that the Normandy breakout and the sweep across France could be pursued without fear of strategic surprise Although strategic deception worked well COMINT showed that tactical deception schemes were virtually always penetrated by the Germans 5 Someone always gave the game away Knowing that deception was NOT working was probably very valuable in holding down the butcher's bill in Allied attacks Sep 82 Some of the in points of the COMINT effort are interesting No German Army Enigma keys were read from aid-1940 until the spring of 1944 The solution of the cipher teleprinter TUNNY in Kay 1944 was the biggest success of the year in giving German strength dispositions plans and appreciations in the West Before D-Day and for three months afterwards COMINT was the primary or only source of a vast amount of information about German dispositions preparations etc Almost all of this COMINT was obtained from German military traffic COHINT gave routine intelligence about the German forces and actions up to the American breakout at St Lo after which it became spectacular In the counterattack at Mortain messages from the XLVII Panzer Corps and from C-in-C West were read within hours giving intentions which had been decided by Hitler The messages were out of date before they could be acted on because of Allied actions German intelligence on Allied intentions at Falaise was known from decrypts German fears of encirclement and Hitler's refusal to permit a withdrawal were known from COMINT German confusion was reflected in the decrypts which were often translated and sent to the users within a few hours When von Kluge vanished on the battlefield at Falaise Hitler personally ordered a counterattack which was successful but the message was not read for ten days The order to withdraw behind the Orne River given on 16 August was read and forwarded in five hours During the retreat after 18 Aug 1944 most of the decrypts were too late to keep up with the fluid battle situation Decrypts showed a temporary lapse in German discipline and morale 7 Entire paragraph Allied use of COMINT changed radically after the crossing of the Seine River because only events in the North were covered by the arrangements between GC CS Goverument Code and Cipher School and 21st Army Group The American forces in France handled ULTRA COMINT differently from all other intelligence Despite security rules ULTRA was discussed at all time on the Signal Corps telephone circuits in France 6 Entire paragraph Only a small fraction of the decrypts produced were used actively to conduct operations Most COMINT was used passively to confirm other intelligence or for information COMINT was found of less use to the attacker who had the initiative than to the defender One common active use of COMINT was to direct reconnaissance at specific targets This confirmed and covered the COMINT In general active use of COMINT required much more than timely decrypts but the results were significant 6 Entire paragraph CRYPTOLOG Page 6 'SP SHeRe l JIfBlb DOCID 4011953 By September as their Army was retreating German COMINT became very good while their OOMSEC tightened and Allied COMINT quickly fell off 10 German SIGINT gave 90 percent of their Intelligence about Allied O B 8 The Allied advance came to a halt at about the same rate that Allied COMINT dried up After another six-month buildup and several major calamities such as Arnhem and the Battle of the Bulge the Allies were finally able to penetrate Germany and occupy it but the Intelligence situation was generally very poor and COMINT had serious outages for months 8J The German Army fell back onto landlines inside Germany and the gradual spread of Reflector D on Luftwaffe Enigma circuits cut down on decrypts 7 Three German Army keys were heavily read for the last three months of the war but Tunny was the main source although it took longer to decrypt and therefore was not as timely 7 10 At Remagen COMINT gave the German plan to bomb the bridge 7 Allied aircraft bombed the German airfields and COMINT also gave damage assessment from the raid 7J This protected the bridgehead After the Allied armies broke into Germany knowledge of German Order of Battle and capabilities was critical to the speed at which the Allied forces moved East 7 Revelations of German intentions were confused by the disparity between orders given and actions taken 7J Although COMINT fell off after September 1944 because the Allies had extraordinarily good COMINT during the critical months before and after the invasion they were able to apply their local superiority in airpower and sea power to hold down casualties and keep the German ground forces from splitting their armies apart or driving them back to the sea COMINT gave an unparalleled insight at the operational level to what the Germans were doing and even though the beachhead expanded much more slowly than the planning called for the Allies avoided disaster until they were strong enough to achieve a breakout References 1 Keegan John Six Armies in Normandy New York Viking 1982 2 Bennett Ralph Ultra in the West New York Scribners 1980 3J Bauer Cornelius The Battle of Arnhem Kensington N Y 1979 4 Gabriel Richard and Savage Paul Crisis in Command New York Hill and Wang 1978 5 G C C S Army and Air Force Sigint Vol XIII p 196 228-30 1952 TSC 6 G C C S Army and Air Force Sigint Vol XVIII p 193 TSC 7J G C C S Air Military History Vol VIII Ch I V 1952 TSC Air Military History Vol 8J G C C S Xl pp 220-5 TSC 9J G C C S Army and Air Force Sigint Vol XVIII pp 125-7 TSC IHistory Lesson Cryptolog 10 1 May 1982 11 1 Der Fall Wicher NSA Technical Journal Vol XX No 2 Spring 1975 TSC 12 Cassidy G L Warpath Toronto Ryerson 1948 Milton Defeat in the West 13 Shulman Seeker Warburg t948 14 G C C S Armyand Air Force Sigint Vol XIII pp 14 2 TSC 15 U S Army in WW II Army Ground Forces Procurement and Traini g of Ground Combat Troops GPO 1948 pp 12-85 16 U S Army in WW II Brea out and Pursuit GPO 1950 pp 692ff 17 U S Army in WW II Lorraine Campaign GPO 1950 p 1 18 Mallory K and Ottar A The Architecture of War Pantheon 1973 p 123 Ch 10 P L 86-36 Q c -- -- o U Summing up Six Armies in Normandy is a very readable and interesting review of a campaign that in hindsight can be seen as an outstanding military success for the side that had the best COMINT and COMSEC German tactical effectiveness generally offset Allied material advantages The small size of the Allied force and the low casualties compared to the German losses and collapse suggests that COMINT and COMSEC combined had a much higher tactical value than is generally recognized Sep 82 CRYPTOLOG Page 7 DOCID SESRrl'f 4011953 l @_ t ' I ---- ---l t I l 11 P his paper will convey some of the lessons that have been learned about the effect that Terminal Sub Systems TSSs have had on NSA This paper was presented Conference 25 May 1982 at the CISI U TSS REVOLUTION ieT I can sum up the major effect that TSS has had on the Agency in one word REVOLUTION sudden complete change The technology is revolutionary because it is a friendly but machine human interface the PLATFORM network remote data and o not provides lovable access to TECHNICAL REVOLUTION the software facilitates flexi ble a adaptable system that can satisfy many active analytic requirements The T55s have brought about cultural changes that are revolutionary for those analysts software developers and Agency managers who have access to or deal with the TSSs Therefore I will technical cultural revolution that has been going on in the Agency for the last five or six years I believe that we now stand on the brink of yet another revolution at this agency that of the powerful personal computers local area networks and true distributed processing Some of the lessons we learned during the TSS Revolution can give us insight into the next revolution that lies ahead speak about Sep 82 8 eea First what's revolutionary about the TSS technology The PDPll and UNIX version 6 are mid-70s technology--certainly not state-of-the-art in 19821 But if you step back and consider what a TSS gives you it might be viewed in a different light It's a mini-computer that is not overly powerful in terms of performance but quite powerful in terms of the CRYPTOLOG SES 'f Page 8 IhI ffBhB 'VIA eallIlfI' elhl ffNBhS eUhY L 86-36 DOCID sseRST 4011953 software the connection with PLATFORM that provides access to analytic systems and data all over the building and in spots around the world and the tempested smart terminal that can be loaded with several different software packages U What's so great you ask about the software After all it gives you the ability to sort edit compile and execute programs and do the normal things that most systems do Well UNIX has some features that set it apart from traditional operating systems especially in the interactive analytic environment SIGINT as an Art e eeo One point we must realize is that SIGINT analysis like programming is an art Individual analysts perform similar jobs in different ways reflecting their individual styles and approaches to problems For example o o o one analyst may have a collection of message traffic that he scans in several passes first very quickly to get an idea of the content then in several additional passes extracting relevant information from which he is generating his report The next analyst might make only one pass extracting all the information needed by carefully going through each message only once The next person may extract data and reformat it with say a permuted index or a matrix to present the data in a form which he can analyze Each is doing the same job but doing it in a different way In this dynamic multi-faceted Agency environment a system that forces an analyst to conform to the way the system wants him to work and that does not adapt to individual analytic approaches will not achieve widescale acceptance Flexibility and Adaptability U The key to the acceptance of UNIX by analysts has been its flexibility and adaptability The I O is set up so that instead of writing a special program that does all the Sep 82 functions you want you have a large set of programs that perform specialized functions and chain them together in the sequence you need to do the work The output of one program goes into the input of the next program until the series of functions is complete This gives a tremendous amount of flexibility in the sequencing of functions an analyst might need to perform a job U Add to this feature the ability to create INTELLIGENT command files--shells--and you have a powerful set of functions that can be configured easily into a tailored human interface that is analyst-friendly P L 86-36 Add to this the ability to acguiredata electronically from and other major and minor computer systems via PLATFORM and you have an Analyst Support system that helps an analyst I l 0 get the data he needs 0 format i t they way he wants it 0 manipulate analysis it for various 0 store and trends accumulate data 0 update data appropriate 0 access word processing functions erate reports o bases on o to major forwardl i t e e c troniCallY through types soon jUand for of monitor hosts to if gen- release P L interact with analysts on other systems using electronic mail and OPSCOMM-like functions U So even though TSS technology is mid70s and not state-of-the-art it does represent what I call a technical revolution at NSA CULTURAL REVOLUTION U Now what about the cultural revolution I spoke earlier about the cultural effect that TSSs are having on analysts software developers and managers I would like to briefly describe these effects CRYPTOLOG SSSRKT Page 9 Ik' UB158 VIA SQtHN r SIk NUSr S QtlbJPY 86-36 DOCID Sfl8Il fl'E 4011953 Software Developers U The TSS revolution has had a significant effect on software development We learned through experience which approach to software development seems best for this type of system Turn-Key Approach feije The first applications project that we tried to develop was a TSS-based transcription system Since this was a formal system acquisition we did a lot of work on requirements analysis and came up with a design which was given a formal review The review determined that the UNIX software which existed at the time would not produce the desired result so a UNIX-based development was not started and an alternate approach was taken This project is still in development and hasn't been operationally used by an analyst Analysts I think you can imagine the effect that TSS has had on the end-user from my previous description of how the system might be adapted to different analysts and analytic approaches The flexibility that I described not only permits each analyst to do a specific job in his own way it also provides a wide range of functions that allow each analyst to do many different types of jobs UNIX supports a wide range of cryptologic functions collection CA TA linguistics reporting data processing management and others It is thus a powerful tool iOUO Yet there are frustrations It has been estimated that less than five percent of the analysts in the Agency have access to TSSs People in A3 are literally lining up to use the terminals So in a sense we have not kept pace with the success of TSS in making the system available to analysts on a large scale Also the performance and reliability can be trying at times--hardware software and communications problems can interfere with the consistency of use that analysts should receive P6ij6 Some systems are overloaded to the point that performance has degraded to a serious degree I believe I ImT44 coined the term Tyranny of Success to describe a system that was so successful that it quickly bogged down because it was so popular Tyranny of success is what we're experiencing on many of our TSS systems These problems are being addressed but they do have a cultural effect on end-users Sep 82 Evolutionary Approach U In our next few projects we felt that it was necessary to take a less formal approach to development especially since we didn't really understand exactly what the analyst wanted and the analyst didn't really understand what the system could do So we prototyped--we experimented Our psychology changed from delivering a turn-key system to that of evolving into a system through iterations--build what we best understand is needed try it out in an analytic environment make changes try it out etc THIS APPROACH WORKED This brought about many cultural changes Software developers actually talked with SIGINT analysts They actually learned about some of the functions that SIGINT analysts performed They built small systems and some not so small that allowed the analyst to do a job This was a good motivator for the developers and the analysts REAL users were using the software that was developed on real targets some of them very high priority This stimulated and chalhmgedthesoftwaredeve16pers Analysts we re getting prompt attention to their needs on a small scale and getting a growing set of functions that helped them do their jobs Also their advice and guidance was being sought--they had great influence over the software functions that were developed CRYPTOLOG Page 10 Ik IQJ 6 IliA SBUHI'f SIW 18J S BilbY P L 86-36 DOCID 4011953 U The people developing the TSS utilities like Data Base Management Systems forms packages intelligent terminal software etc had some feedback on how useful the tools they built were and what other functions might be needed--and this was very important It caused an impressive set of utility functions to be built and made available based on feedback from REAL users and applications software developers Generalized Software 8 SSQ An important lesson that was learned during this development period was that the software could be generalized quite nicely in many cases so that the same program properly parameterized could be used on various targets by different organizations on a varietv of nroiects Analytic Approach to Evaluation S SSQ The analysts we worked with had two basic approaches to deciding what they needed One approach was to get onto a UNIX host go through all the programs available and try to do their job If they ran into a situation where there wasn't an existing function to do what they needed they would ask us to build one A good example of this is the sort rour The ASCII codes for characters didn t give the proper sort sequence so PI requested a version of 'sort' that could handle non-standard alphabets--and it was provided The other approach used by analysts to decide what they needed was to write a specification of some type--a concept of operations CONOP a requirements specification or some similar document--and we would develop the capability in an iterative fashion This approach worked well for the larger more com rehensive ro ects like L This was different from the transcription system experience I mentioned earlier because we understood from the beginning that it would be an evolutionary development rather than turn-key Users were generally pleased with this approach because they would get the opportunity to evaluate the development while it was going on and give feedback to the developers on changes they might need In some senses it was a rapid prototype environment __' ---- ----- I There are many instances where this generalizing if designed into the project from the beginning and developed iteratively has worked better t 1 lanwehad thought possible EO 1 4 c P L 86-36 Graphics Experience 6ij6 An important example of this is the work the Graphics Investigation Group T4 T3 R53 did on computer graphics which resulted in a highly generalized graphics capability available to any TSS In 1980 and 1981 the GIG sampled analytic organizations to find out what the general requirement for graphics was at the Agency We concluded that most of the organizations had similar graphics requirements mostly overlays on geography and management displays charts graphs etc The similarity of functions at the software level told us that we could satisfy a wide range of requirements with a comparatively small software effort by means of generalized graphics software U After this had gone on for a while we looked around and discovered that a significant stockpile of basic software components existed for various types of analytic work We basically had a box of analytic tools that had been developed and tested in operational areas on real problems Sep 82 P L CRYPTOLOG Sl BRGBPl'P 86-36 Page 11 DOCID 4011953 U To make a long story short a generalized graphics package was developed partly commercial partly in-house and installed on a TSS This gave us a powerful adaptable graphics software capability that operated on a TSS and could interface with the existing alphanumeric graphics programs 3 eee Hore and more TSS systems were deployed and more and more people learned about the ability of TSSs to assist them with their work we began to et a lot of re uests for software s stems The system required some sophisticated alphanumerics graphics and transaction processing that challenged us to deploy what is probably the most sophisticated software system I know of on a TSS today The development effort was successful the system is in operation today I would like to acknowledge the substantial software effort put forth by R53 in developing nearly all of the graphics software for OHS Hanagement pe e As you might suspect with approximately 50 TSS systems in existence the management problems in dealing with the many aspects of these systems are quite significant This is an entire revolution in itself we certainly don't have time here to even scratch the surface but I do want to emphasize one point there is a significant management problem throughout the Agency in dealing with this many systems 1 4 c P L 86-36 REQUIREMENTS One final note There is one substantial weakness in the software development process whether on TSSs or other systems that I feel is the cause of many failures and problems in system acquisition REQUIREMENTS SPECIFICATION Based on my experience we don't do it well It is a very difficult task It's the most critical stage in the development of a project yet it has the weakest set of tools of any phase of software development I feel very strongly that if we had an automated method that would generate the requirements documents--or a functional prototype that both users and technical people could understand and work with--it would dramatically increase the Agency's ability to develop systems that really satisfy user requirements U I've tried to give you some of my thoughts and observations on what I call the TSS Revolution at NSA Even though I've only touched on a small number of the issues involved with a technological and cultural revolution we must remember that we are now on the verge of a new revolution They are going to start coming more frequently now because of the rapid advances being made in technology We have to learn how to deal with them with as little negative effect and as much positive effect on the Agency's mission as possible We're learningl The TSS Configuration Control Board was established last year to gain control over TSS management They are currently dealing with many management issues that affect current operations and future architectures and policies The need to provide more direct customer support for a large number of interactive users and software support organizations may require significant changes in DDT and DDO organizations U The lessons we have learned from the current TSS Revolution--mixing operational prototyping experimental development evolutionary development generalized software etc may and SHOULD have an effect on the acquisition of interactive systems We can't afford on a corporate basis to get too informal yet the more formal route has been shown to be not very successful Sep 82 CRYPTOLOG SESRB' Page 12 1h'd'BbB 'IA 861II1'' '' 8HMflffibS ellb'l DOCID 4 a 11 9S3-----L---- L SIGINT 1990 u L 86-36 EO 1 4 c by Joseph Meyer P13 11 ajor developments in telecommunications technology and systems during the 1980's will have a profound effect on SIGINT by 1990 The main threads of this development are new U satellite systems optical fiber cable electronic switching the coalescence of computers and communication nets and the increasing complexity of telecommunications What new problems will SIGINT have to face by 1990 Wba t do the new trends in technology tell us about the not-so-distant future Th author has adapted this article presertted here in several monthly installment s from his presentation at a January 1982 session of CA-30S The overall classification of the series will be TQP SECRET UMBRA U In order to see what the communications environment of 1990 will be like and to analyze the impact on SIGINT we will begin by looking at current trends in technology networks and traffic growth costs and specific systems Sep 82 IiS'i' RGBPbekSABbS 'i'S eSli'i'R Ae'i'SRS CRYPTOLOG 'i'SF SSeRS'i' Page 13 ItAlfflbS 'f' A eSIIHi'i' eltAliliI'lbS SlibJPY iQP SSSRSi DOCID 4011953 SIGINT 90 TRENDS IMPACT CHOICES U In order to respond to the changes in the telecommunications environment it is necessary to make some choices and to consider some specific actions policies and projects that will be needed Most of these will be dealt with later in the paper however a fundamental choice is to begin with a suitable definition of SI INT viz SIGINT is the process of obtaining secret or unknown information from communication systems or signals U In this definition it is the process not the product that is important Also the desired information which may be intentionally kept secret or merely unknown to us is to be obtained from communications systems not merely from communications The system includes all the traffic but also includes things such as computer memories circuits switches software documents etc Any method of getting secret or unknown information from the communication systems or from signals falls under the authority of SIGINT U The importance of getting the definition of SIGINT right is that the classical passive model of SIGINT sitting back waiting for signals to reach it is no longer appropriate for the problems of 1990 The attacks against computer-communication nets and against systems such as optical fiber require operations based on physical and electronic penetration of the target links and nets tightly coordinated with monitoring and analysis to gauge how well the penetration is doing This cannot be handled by multiple agencies trying to coordinate a mission the authority and operations must be unified into SIGINT paper Why Pollyanas Prosper at the January 1982 AAAS meeting noted the meliorists discount the future at an extremely high rate in favor of confidence about the present Thus all the forecasts are rosy up to the very end THE TECHNOLOGY BASIS FOR TELECOMMUNICATIONS U Many of the major advances in telecommunications have stemmed from advances in materials science MATERIALS SCIENCE IN TELECOMMUNICATIONS Material ooooooo oooooooooooooooo o Application Copper o ooooooooooooo oo oooo o Telegraph lines Carbon Microphone ooooooo oooooooooooo Telephony Gutta Percha oooooo ooooooooooooooo Ocean Cables Tungsten oo ooooooooooooooooooooooo Vacuum Tubes Crystal Growing ooooooooooooo Stable Frequencies Semiconductors Transistors oooooooooo Computers Magnetic Oxide ooooo Core Memories Disks Tapes Cryogenics oo oooooooooooo Space Communications Solar Cel1s ooooooooooooooo Space Communications Glass oooooooooooooo o ooooo oooo Optical Fibers U The list above although not exhaustive indicates the importance of improvements in materials in the growth of telecommunications The purification of copper was a fundamental step for it made possible electrical engineering of all kinds and electrochemistry has been a basic technique for purification of other materials The advances in glass technology are currently bringing about a revolution in telecommunications by making lanclline F t 86-36 cheaper than radio relay or satE l1ites EO 1 4 c 1 1 - - - -_ _ U There is little basis at present for optimism about the future if SIGINT continues to operate in its present framework yet unoptimistic forecasts are rarely welcomed As a Sep 82 US ftflbShSABbS iQ SQUiRs 'zSiQRS U The effect of applying materials advances and various other inventions to telecommunications has been a steady improvement in technology so that much more information can be sent at a much lower cost CRYPTOLOG iQP SSSRBi Page 14 IMUBhB VIA SQIIIUi Slh IUShS QllbY DOCID 4011953 ep SHeM The Sequence of Inventions in Telecommunications 1840 2000 I_stment Cost Per Volee Ct1 nel Mile IDoIlsrs Equ' 'ent Bits Per Second 100 8 l11on Ii 13 Billion lOO OG 1 Billion lOG Million 10 MIllion 10 OG 1 MIllion IOG llOO 10 000 I OG 1 000 lOG 0 10 i ' ---- r---r---r--- T--r--r---r---r- T' --r--r-__ ---r- T' - -- __ - J llSO 1110 1110 1120 1130 I Souoce _ 11 ' 'ull of '- ow J P oL 86-36 EO LA c 1 SEQUENCE OF INVENTIONS IN TELECOMMUNICATIONS U From tna primit1 Ve lO words-per-minute wpm transmissions of land line Horse telegraph to the gigabaud satellites and optical fibers of the late 1980's there has been a steady advance in transmission capacity amounting to a tenfold capacity increase ev ry 20 years since the invention of the electrie telegraph At the same time the average capital cost of building these systems has dropped tenfold in the last 50 years i VOLUME FORECASTS ' U Both traffic and equipment have increased in volume ao that the 400- 111ion telephone plant of 1977 is expected to grow to a l500-mi11ion subscriber plant by 2000 Sep 82 lfe'F lU bBA 9ABbl' 'Fe eSII'Pftilre'FSR5 CRYPTOLOG 'Fep 5BSM'f Page 15 YIA SS lIN'f SHllflHBb5 BIlbY 'PElP SBeitB'f DOCID 4011953 with many of the technical improvements in SWitching transmission and security will make it harder for SIGINT to find and exploit the specific traffic that is worth working on To quantify this the U S with about 40% of the world's telephones generates about 170 billion calls per year of which about 20 billion are toll calls Non-U S toll traffic is about 20 billion calls each about 10 minutes long No agency can look at all of this traffic The problem of selection is paramount and selection will become much harder a6 5 j l' o g J 1960 1980 2000 Foncast growth in the number of telephones in use l the world 6 Asia and Oceania Europe d North America Central and South America f Mrica 2 FORECAST OF NUMBER OF TELEPHONES U The total book value of the world telephone plant at present is about $300 billion An A D Little study predicts that $640 billion will be spent over the next 10 years for a new telecommunications plant with expenditures rising to about $80 billion by 1990 This will be for a predominantly civil telecommunications plant S S O A second impact of the growth of plants is that the turnover time for equipment will decrease At present transmission systems have a cycle of about 15 years after which it becomes economical to introduce new transmission technology e g analog satellites of 1965 are being replaced by digital satellites of 1980 Switches have had 30 year cycles even though old switches often are operated for 50 years But these long service lives are no longer economical and equipment will be replaced at a hastier rate by more capable equipment This means that the lifespan of certain kinds of SIGINT systems and techniques will also be shortened and new SIGINT plants and methods will be needed to keep up with the target changes U In the developed world the growth of telephone stations has slowed as shown in curve d above but in Europe and Asia the growth is faster as high as 10-12% In Africa the telephone density is not only low but will remain low because of lack of money to install the plant INTElSAT II-A CAHTAl Z IHTElSAT II TAT S 'HTElSATm INTElSAT II fTElSAT I TAT 4 TAT 3 CAHTAT' TAT Z TAT I U The cost of adding a telephone station to a network is about $2000 which covers the cost of the subscriber loop local switch and long-distance transmission facilities Hence the expected cost of telephone plants from 400 to 900 million phones by 1990 will require an outlay of about $1000 billion reduced someeconomical switching and what by more transmission plants TAU 'a9740 I ' 1 96 O --1 - 91 7-0-l 98 0 TeI' Phone calls between the U K and North A enca and the introduction of cables and satelhtes 3 PHONE CALLS U S U K 5 ee6 The impact of this growth in telecommunications plants which will carry many services in addition to POTS plain old telephone service is that more and more of the political economic and military activities of the world will pass over telecommunications circuits and will be dependent upon those communications This means that the potential returns from SIGINT will increase at least as rapidly as the telecommunications traffic which is already growing at 20% annually However the vast physical plant along Sep 82 P1Q'F Rl ibBASMlbf3 'f8 e8N'flblte'f6itS U One of the most dramatic effects of telecommunications advances has been the increase in intercontinental traffic as cheaper and better systems have allowed economic and governmental activities to operate on a global scale In 1927 transoceanic telephony was initiated with calls averaging 2000 per year The first submarine telephone cable in 1956 brought about rapid growth in traffic which continued to grow at 20% annually The development of higher CRYPTOLOG tap S8SR-fit Page 16 N i 1' COMUTl CIUNIllF'S OUY DOCID 4011953 ep SBSllB'l' capacity satellites and cables has given a transoceanic capacity of about 12 000 circuits During the late 1980's this transAtlantic capacity will be more than quadrupled see undersea cables discussed below voice U In 1927 when there were only 2000 trans-Atlantic phone calls each of those calls cost about $400 but today with the increased volume and improved equipment a trans-Atlantic call costs only a few dollars The introduction of cable in 1956 improved call quality so much that it uncovered a demand which has justified expanding the capacity to the 12 000 two-way channels now in service 500 i u a0 200 100 e TIIICOlMfruric4tiofU obfrcti 2000 '0 l I 4 3 10LL _l--_ L-_ L _---J'--_ I- 1m 1980 1990 2000 2 Intercontinental traffic foredUll telephone caD fronl the U K 4 INTERCONTINENTAL TRAFFIC FORECASTS U Various authorities including Intelsat have estimated intercontinental traffic growth at 22% annually but more conservative estimates expect growth to taper down to 10% after 1990 The growth of telephony however does not express total traffic growth because owing to time differences the calls are generally placed in a short time window when people on both sides of an ocean are at their desks Thus late afternoon in Europe connects to early morning in the U S U The transoceanic networks are designed to pass this peak load traffic but so far have been comparatively idle during off hours with traffic dropping to 20% of trunk capacity outside the peak four-hour period The development of new kinds of traffic for international circuits particularly digital facsimile electronic mail and computer data traffic will tend to fill up the slack hours of the nets so that total traffic flow will grow much faster than 20% By 1990 the major telephone operating authorities expect digital facsimile to be the next major traffic after Sep 82 ue'f RBbSltSltBtiS 'fe eeU'fRAe'feRS ---'- P L 86-36 EO 1 4 c GBPvolulion du tnIlc ttltphonique loW 5 6% to 12% INCREASING TRAFFIC GROWTH U CNET Centre National d'Etudes des Telecommunications has projected increasing growth rates for French traffic from 7% in 1977 to 12% in 1990 The projection is apparently due to expected improvements in speed reliability and cost of the services as it becomes easier to use the network Penetration of the telephone into new use areas is also a factor for as more people are brought into the network telephony replaceS mail and travel CRYPTOLOG TOP SeeRB'f o Page 17 H 1tliBbB VIh Sellnl' Sll zNUShS QIIl JPY DOCID 4011953 'FSP SESRi i'F will be almost no aspect of the economic political social or security activities in any country that does not use and depend upon the telecommunications systems Hence SIGINT will have the highest potential growth rate of any intelligence service and will be capable of the deepest and most extensive penetration into the activities of any target country-providing it can be done successfully and is adequately funded only be a small part of the total in spite of the fact that the French have been developing Telecom satellites The main growth will be terrestial circuits which is logical since France has substantial business with neighbors The hidden element is that the French as well as the other Europeans intend to stay away from satellite circuits in favor of optical fiber interurban and international trunks Satellites will be used for mobile services and for backup and to establish new services temporarily until landline or optical cable can be installed Teticommunicalions objeclij 2000 milUers nombre de vol multiplexes mllrlOflS 'FS 88S he impact on SIGINT of the European preference for optical fiber cable over satellite is that their international traffic will be harder to intercept If the relations between the U S and Europe change during the 1990's e g by the dissolution of NATO or by a political shift toward the Soviet Bloc or to Eurocommunism those countries may become lA active targets but access to their traffic will be harder This access will be even harder if the U S presence is greatly reduced e g by withdrawal of U S military and governmental organizations from Europe d' uipemenls 150 100 50 PHONE DENSITY GNP RATIO quipements d'extremite U The correlation between national wealth and telephone density is well established see page opposite The rich industrial countries have a high telephone density as many as 70 phones per capita while poor nations such as India have only about 3 telephones per 1000 inhabitants In Africa the telephone plant is so undeveloped that only Kenya appears on the graph at all millilllS de citaJiIs 100 U In spite of the lack of civil telephone facilities the governmental telecommunications plants are often more advanced with satellite and microwave and long-distance shortwave nets The U S S R is a case of special interest because its telephone density is quite low yet it has a highly developed governmental telecommunications network 50 Circuits inlemationaux 6 CNET INTERNATIONAL CIRCUITS U The French forecasts of international circuits up to 1990 show several interesting characteristics Total circuit growth will be about sevenfold but satellite circuits will Sep 82 US'F ReE Ei GBP 'tB S 'FS SSN'Flb trS'FSRS U One of the striking features of civil telecommunications in Africa and Asia and other low-income areas is that there are practically no rural telephone plants at all At INTELCON 80 the Communications Minister of Nigeria asked his fellow Ministers of other African countries if any of them could name a single case of a rural telephone system and there was no reply The World Bank has been pursuing projects to introduce public telephone service into the rural areas of the CRYPTOLOG 'FSP SEeRE'F Page 18 trArmL VfA 8611Iff'F elblrlftffitS SlftJy --- - a DOClD 4011953 'PElP SBeRB'P THE RELATIONSHIP OF NUMBER OF TELEPHONES TO GROSS DOMESTIC PRODUCT IN VARIOUS COUNTRIES 100 S ' den S 'il rlalld N Zeolon JoponUnited Kil 9domAU lri IIaly_ - Canada O n 'o k o -1c aliQ -Nelhetland nlone o I' 'ar ay _ I Ger 'ony -B 19 i FrOllce 20 - Czechoslovakia Pottu9ol- A nli UrU9 1C1y_ Soul Il Africa- o C 25 0 -' E 4 Ko 3 Ecullllor o oo c 2 Q -I on 0 -' Q o Philippi o4 3 2 1 100 GOP per Capita U S Source Sep 82 N9T MUiM hBI 1 i '9 e811'lb ER8RB Dolla INTELTRADE CRYPTOLOG 'l SP SBeMl'l Page 19 Itldft'tB 'lI SElllIli'P elh dRlf lbS 8flbJPY DOCID 'fep SaeRS 4011953 Third World on the economic argument that the phone service will pay for itself but pro- 1OOmlllons gress is slow The basic problem is cost coupled with problems of equipment reliabil ity Satellite communications are not seen as 40 an answer to Third World telecommunications in general because the population is too sparse and the economic margins too slender to make two-way earth stations feasible even at the town level Even when nations get some cash 1 0 m l l l o n s l - - - - - t - - - - - - - - f - - - - - - - - t 8 flow e g from mineral exploitation the 8 money tends to stay in the cities where the 5 rate of return is highest 4 -------r---------r------- a' 3 -- 2 _ _ I S SGg The impact of this telephone GNP correlation on SIGINT is that for the rest of -- ---- the century the spread of telecommunications I mIlIlonl-----I---------b - ' services and the flow of traffic in the uq- 800000 600000 developed countries will be tied to the 500000 cities and to economic and governmental 400000 activities with the exception of broadcast- 300000 ing The demand of the Third World countries for robust long distance communications created a problem at WARe 79 over HF allocations and assignments The Third World nations will probably crowd into the HF spectrum as fast as they can buy equipment but there is not nearly enough HF spectrum to accommodate them The best alternative communications for remote area communications 20000 will be low-cost over-the-horizon systems such as meteor burst cited below thin route tropospheric scatter and low bandwidth mobile satellite services using high-powered satellites with big antennas At the same time the main telecommunications trunks serving cities and the industrialized countries will carry traffic that penetrates to every busi2000 ness household and government function In brief economic status will be the primary l000L-' '--_ __----- ' ' selector of telecommunications plant and 1977 1990 1985 1990 traffic The higher the relative cost of a Impact qUBnlllalir des services de 1 lkommunication message or an enciphered message in a coun pare de lerminaull try the greater its expected information value I En dehon du terminal annualre 9 eeS The SIGINT systems will cover increasing masses of traffic in which the expected value of any message continually decreases even though the total value of the information in the networks will continue to grow at least as fast as the GNP One of the most interesting SIGINT factors is that the third world nations being poor are still largely open markets for well designed heavily subsidized basic telecommunications plants which could with enough ingenuity create interesting SIGINT opportunities U While overall growth of POTS plain old telephone service will increase at about 6% annually a number of specialized and new serSep 82 tiS' Rri'bS SltB'bB 'fe SStI'fntt6'f8RS 8 CNET QUANTITATIVE GROWTH OF VARIOUS TELECOMMUNICATION SERVICES vices will grow at higher or lower rates The CNET graph above shows projections till 1990 for French services Traditional Telex which is already a saturated market will grow very slowly while new services such as private radio videotex teletex and telecopier will grow at much higher rates Private radio has been resisted by European PTT's but CNET shows an expected fivefold growth over the next ten years Teletex at 2400 bps is expected to have fifteenfold growth from 1980 to 1990 and more significantly is expected to become the lingua franca between various CRYPTOLOG 'SP SSSnB'f Page 20 IIMfB'bS 'JIlt SSHHfF SIWINB'bS StibY DOCID 4011953 P L 86-36 EO 1A c information systems computer data etc 'fElf Be6RE'f such as telex videotex U The growth potential of new specialized services is sometimes misleading as the case of SBS Satellite Business Systems shows Originally developed to sell high-speed data services to the top Fortune 100 companies it has only managed to attract about 25 major companies and they want POTS viz corporate voice traffic services Long-term growth potential of network computing services is expected to be good but over the short term the system and software costs may eliminate all but the richest companies L M Ericsson has bought up Datasaab to position itself in computer networks AT T is trying to develop a computer net service and IBM is looking to extend its computing capabilities to overseas markets by international operation of a domestic satellite SBS system Other domestic satellite Domsat operators are also looking for overseas outlets for their services Even the German telecommunications giant Siemens is struggling to develop and incorporate successful computer services even though its microelectronics VLSI technology is as good as that of the U S leader Intel Corp Business Week 1 Feb 82 p 87 The French government has undertaken a high technology program in telecommunications aimed at export markets and French industry is in a leading position in electronic switches and terminal equipment For the modern telecommunication services software development is a critical element and only the Japanese are in a position to challenge the U S seriously Sep 82 t1QlF R BJ 8t SABJ S 'F8 681l'PRit6'P8ftS u FACTORY SHIPMENTS OF ELECTRONICS V The steady growth of electronic sales see figure 9 in all categories shows a tripling in dollar value Electronics for information processing is the largest segment of the electronics industry but the demand for communications common carriers has been growing the fastest in the last half-decade Common carrier growth may be accelerated even more by future demands from home information systems P L 86-36 EO 1 4 c V S TRADE IN COMPUTERS V The continuing growth in computer sales and particularly the growth in imports is a manifestation of the spread of computer manufacturing technology The coalescence of computers andeommunications examined below will make thts computer base a major factor in telecommunications The strength of the foreign technology especially the Japanese is manifested in the imports Another factor of some importance is that the U S market for telecommunications equipment is now virtually unrestrained for foreign competitors who can subsidize sales to the U S to force an entry into the V S market As the U S telecommunications industry is fragmented in the name of competition more foreign computers communication equipment and services will take over parts of the U S market e g oo for PBX's interconrections satellite links data terminals and even basic transmission and switches CRYPTOLOG Page 21 'PEIP SeeftS'P P L 86-36 EO 1 4 c Hi Nflb8 VIA eeUHf'f ellMftil3bS ellLY DOCID TQP SE6RS'l' 4011953 Annuli Flctory Shipments Comntunlcallons Ind Electronic Equipment by Type U S 1970 1979 IIlIIon 01 Doll 10 u ' Communicillonoo looldelll Sludl lnllrcom M IVlllon Quldlncl ETC Equlpmlnl lor UII b CommunlcllIonl Common Clntlr 45 -J 35- CompUler and 'arlDlwrll Equlpmenl Olhlr 3lI- 5 0 I o 1170 _ Electronic _rIo __ 117 lin lIT4 lin 1171 75 1177 117 117' U S TRADE IN COMPUTERS AND RELATED EQUIPMENT BILLIONS 8 0 7 5 7 0 6 5 I 6 0 5 5 I I 5 0 4 5 4 0 3 5 --_ ' EXPORTS - 3 0 2 5 2 0 _ _ _ ill' 1 5 1 0 5 -- 1970 1979 CBEMA Source Sep 82 fleT Rl3bl3ABABbE 'fEl 6Elfl'flhl e'fElRS CRYPTOLOG Page 22 ft1l1ft L1 Ii fA eS I'Ell'f elhlrlfNl3bS QlH Y Tep Sl3eRl'l'f i _ JDOCID 4011953 'f9P SBSRB'f 86-36 EO 1 4 c 90 70 50 30 '-IllO 9 l 7 5 3 11 NY-LONDON PHONE CALL U In 1927 when there were only 2000 calls per year each of those calls cost about $400 but today such calls cost only a few dollars The introduction of cable in 1956 improved call quali ty so much that F t 86- 3 6 uncovered a demand which has justified expta- 1 4 c sion in capacity to the J2 o 000 two-way channels in service nOW D 80 Il 5 '0 COSTS j At the same time that the civil telecommunications plant has greatly increased in size and traffic capacity the cost of services and equipment has gone down drastically This has encouraged greater usage by a wider public giving a larger revenue base which in turn leads to faster expansion and greater cost reductions U U To some extent the economies in technology in the civil nets have been applied to governmental and military dedicated nets but at the same time the military nets have been given more difficult requirements e g anti-jam security transportability survivability etc so that the unit cost of military communications has not fallen much and the demand for much greater capability and higher information rates has made military telecommunications more expensive for the same mission Sep 82 I 60 microwave radio system 40 wave'lJidj 20 p c i nb 0 1945 1985 12 TRANSMISSION COSTS FOR DIFFERENT MEDIA U Technological advances in transmission have greatly reduced the costs of moving traffic For many years microwave radio relay was unsurpassed as the cheapest medium for trunk traffic Efforts to develop radio waveguide transmission were unpromising but optical fiber waveguide has developed so rapidly in the past ten years that even the major manufacturing companies have been surprised Over the next ten years optical fiber will be substituted for existing or new cable and radio relay on many thousands of kilometers of trunk routes particularly where traffic is heaviest CRYFTOLOG 'FQP SBeM T Page 23 IWi9loi HIt SQUINT SII MOINl ibS QNI JPY DOCID 4011953 500 H 'fep S'BeftB'f great deal of conservatism in telecommunications planning and a resistance to innovation By contrast computers and the associated operating systems and manufacturing plants have usually been written off after 7 or 8 years because of the rapid changes in technology Now as digital switching is introduced into telecommunications networks the capital cost of the switches is dropping to about 15 percent of the total network so that much less money is tied up in the switches o At the same time the switches are more capable The No 4 ESS of AT T has undergone a complete replacement in hardware technology while keeping the software and functions the same and this has reduced frame cost space and power requirements by 60 percent INTELSAT 1 200 100 oS 'I l i INTELSAT II INTELSAT IV INTELSAT IV A 'I l j INTELSATV 20 1 10 current bandwidth coat 5 i of current bandwidth --- cost 2 1 1970 1980 1990 Satelllte bandwidth 2000 COlt trend 13 SATELLITE BANDWIDTH COST TRENDS U Reductions in satellite costs have been even more rapid than in terrestrial systems The development of stabilized satellites which could keep solar panels and directional antennas pointed at sun and earth has given much increased power efficiency A combination of improvements in space and earth systems has produced more than twentyfold reductions in channel cost in little more than a decade space switch'ng analog digital switching transmission analog _ transmission - digital switching digital transmission r-' 14 TELEPHONE SWITCHING COSTS U Capital cost of switches used to be 50 percent of the total cost of the telephone plant This was one of the reasons that switches were designed for 3D-year financial lifetimes viz the cost of the switches was amortized over 30 years in calculating the rate base for subscriber costs This led to a Sep 82 eeO A major implication of the reductions in switch cost which is further affected by the deregulation of telecommunications in the U S and the opening of the U S market to foreign switch manufacturers is that the financial and functional lifetime for switches may be reduced from 30 years to less than 10 years over the next decade It will not be worthwhile to keep an old switch in service when a cheaper and better switch can do more and provide entirely new services This has an important implication for SIGINT cultivation discussed below 8 SSe The major impact on SIGINT of all these cost reductions in telecommunications systems is that the target costs to install and operate nets are going down while the costs of SIGINT are not going down SIGINT is inherently labor intensive and because it must extract unknowns from the target nets and traffic it can never match the level of automation and efficiency of ordinary telecommunication nets By analogy a computer can search text for words or do lengthy calculations far better than a human but a person is far better at writing an essay or deciding how to attack and solve a mathematical problem Most telecommunications is routine while the most critical parts of SIGINT are very nonroutine Even the routine parts of SIGINT are subject to continuous change so that economies of repeated large-scale operations are seldom realizable When SIGINT is highly efficient it is a result of chaining information in a highly non-routine way Inevitably the cost reductions in telecommunications will result in more and more plant and traffic which will be more and more overwhelming to a fixed level of SIGINT effort is CRYPTOLOG 'fep SBeRB'f to eSe One of the keys to reducing costs standardize the traffic so that all Page 24 th'dfBbB VIA S9UHI'F SlhI HNK '8 QNlsY DOCID 'fOF 4011953 kinds of traffic voice facsimile telex data etc can flow through the nets easily Another key is to concentrate traffic so that efficient wide band transmission can be used The immediate effect of these measures on SIGINT is that larger volumes of data have to be collected and scanned with less easily identified characteristics which means that the SIGINT costs per extracted message go up as the costs of transmitted messages come down This is an ineluctable consequence of technologic advance as long as SIGINT depends on a modus operandi and a physical plant which creates this cost exchange RADIO COMMUNICATIONS e eeO Because modern SIGINT has been predominantly based on radio interception it is worth looking at some of the main trends in future radio systems to see what effect these will have on targeting and collection and on the subsequent analysis of radio traffic 3 0 ---320 -- 2'0 220 '10 FUTURE OF SATELLITE COMMUNICATIONS U Saturation of the radio bandwidth has been a mounting problem in satellite communications figure 16 This is particularly crucial for orbital slots used for transoceanic communications At WARC 79 arrangements were made to increase satellite bandwidth allocation to nearly twice the pre-WARC amount During the 1980's the major satellite developments will be in the 12-14 GHz frequency range but by 1990 current projects show that demand for satellites will outrun the availability of those frequencies and various countries are now developing technology for the 20-30 GHz range for the 1990-2000 era For special purposes millimeter wave and laser beams will be used for up-down and for intersatellite links U These new satellites will be capable of significant traffic volumes because they will be able to reuse frequencies A typical transponder has a nominal bandwidth of 40 MHz and will carry about 40 Mbps If 1000 foreign transponders are in service this corresponds to a traffic capacity of about 40 gigabits In addition to the foreign and international satellites there is a proposal to allow domestic U S satellites to crossnet into foreign networks because many of the multina- INTELSAT Annual Report 1979 15 INTELSAT GROWTH TO 1980 Sep 82 fl8'1 RBI5EftSABhB '18 eeU'I'RJ Ite'feaS U The principal international satellite system INTELSAT began in the mid-1960's with one trans-Atlantic satellite and a few big earth stations and by 1980 had over 400 antennas at about 270 earth stations U The initial applications of satellites were to international circuits through the INTELSAT organization Now the area of rapid growth appears to be in domestic and regional satellites that operate outside the INTELSAT framework During the next decade NASA will launch over 100 payloads most of which will be communication satellites During the same period ARIANNE will launch about 50 payloads also mostly communication satellites Market estimates expect a demand for 1000 transponders for the Western Hemisphere and about the same number for the rest of the world added to existing satellite systems Broadcast satellites for direct-to-home or for broadcast distribution to fairly small antennas will also grow These broadcast satellites will also be capable of one-way message services Mobile communication satellite services particularly for shipping are already in operation and will extend their services to isolated fixed and mobile users EARTH ITATIOHI IVoo r IDI III Source SBeRB'I CRYPTOLOG 'Pep SESReE'P Page 25 Ib'triflhB ' fA eeUiIf'f elb'tlftffitS Orit'l DOCID 4011953 Application BroadQ In mo ionoI Domestic 01' Men FiJ '6 I Induotri at priwle c - Moblo Frequency S 4GHI I 14112 GHI 30 20 GHI ToW GIobII C C 5 T chniqu I Multiple ICC8a I _ E I I F _ O nd 5pmbe _lion 7'I On-_rd switching I Int relil link R_ratiw repe tw ItJl I P L 86-36 1 4 c Z I I FUTURE OF SATELLITE COMMUNICATIONS tional customers have foreign facilities and they want their dedicated corporate satellite nets to reach directly to the foreign sites The power of these companies is so great that they may cause the foreign PTT's to yield to their needs U Various new techniques will be used to increase the efficiency of satellite communications principally spot beam operation onboard switching and regenerators on the satellites By 1990 the transmission techniques will have advanced to the point where it will be necessary to measure the position of the satellite within a few millimeters and to transmit a timing signal accurate to a picosecond around the U S to allow the high bit rate 10 gigabits sec operation and onboard switching The spot beams will switch from ground point to ground point to direct satellite energy to time-switched users Mbps trunk encryption IS eeO The impact on SIGINT of this expansion of satellites and traffic and the introduction of new technology is that first the data volumes will be overwhelming Second the spot beam operations will require intercept earth stations to operate within the footprint of each targeted satellite and third the collection technology will have to be at least as sophisticated as the target satellite The fourth problem implied by the change to spot beams is that intercepted traffic must be relayed from outside the U S in great quantities U An example of the new trend in satellite communications is the French TELECOM I which combines the features of MARISAT and SBS and provides a low-speed wide-area service and spot beam high-speed data relaying with 25 Sep 82 liey Il 8bEh SABUJ ye ee'l fft1te'fOft S CRYPTOLOG 'FQP SI iSRI i'F Page 26 U1dffltP HA ee lfli'f elhlrlUlBbS eNbY DOCID 'fep BBSRGBP'f 4011953 millions dekm vole 500 P L 86-36 EO 1 4 c YS eee Obviously it would be better to select desired traffic before relaying the data but improvements in switching especially Common Channel Signaling and bulk encryption of satellite channels will defeat the existing systems of preselection and targeting 500 millions de km voies Syst mes sur clbles 18 GROWTH IN FRENCH CABLE U The French who had concentrated on cable have not abandoned this medium and CNET studies show a fourfold increase in cable capacity over the next decade Digital cable transmission is also a fast-growing subset of this new plant Syst mes sur faisceaull hertziens 17 GROWTH IN FRENCH RADIO RELAY U The advances in satellite systems has by no means canceled the growth of terrestial radio relay The French CNET studies show a fivefold increase in their radio relay over the next decade with digital transmission accounting for half the growth The digital microwave systems pioneered by the Japanese have shown robust performance in the presence of urban noise In addition they lend themselves to data communications and to bulk encryption Sep 82 He'f R Kbl h BA8U 'fe GeN'fR 'J riJ'feIl S e eeO The impact of these French developments on SIGINT derives from the fact that the French have an aggressive export policy in the area of telecommunications technology Only a small part of the $60-billion world market is actually accessible to competitive marketing but the French have studied this problem and have established a program to make equipment developed for the French PTT particularly suitable for the export market The French banks and government collaborate in getting the overseas contracts despite keen competition from the Japanese and other European manufacturers As a result technology developed for the French market including switches and terminal equipment will be sold on favorable terms in the Third World market to get the French manufacturers in on the ground floor The French are even selling new switches and terminal equipment in the lucrative U S market The result will be a steady flow into Third World countries of modern digital telecommunication equipment accompanied by the adroit national presence of the various supplier countries CRYPTOLOG Page 27 DOCID 'PBP SBSRB'P 4011953 PROJECTED GROWTH IN CELLULAR RADIO 1400 ADVANCED MOBILE PHONE SYSTEM lAMPS lJOO @tAU MDCIU IITI 1200 - OUND_IffOP 'eCftL UUIII - - CILL IItI T 1100 o DA L ' 0 0 0 1000 en p CILLI wn US too ' 1 uM e p ItF CH NELI U 100 ' en 100 00 ' 0 p 500 400 Switched Digital Capability lOO 200 19 CELLULAR RADIO 100 c Il U One of the large growth areas expected in new telecommunications equipment and services is mobile telephone which can connect directly into the civil telephone nets The current scheme for such mobile phone systems is to have a large number of frequency channels and many transmitters in an area e g a city The radio system will track each user vehicle and switch the call to a radio channel that can be heard by the mobile station The mobile receiver is switched at the same time For ordinary FM radio links the user is unaware of the switching As the vehicle moves it is assigned new frequency channels As a result the circuit for a single conversation may change frequencies many times as the vehicle moves A large electronic switch No lA ESS is used by the Bell System for their mobile telephone system Other proposed schemes would use spread-spectrum transmission so that many competing services could operate competitively in the same geographical area With the interest in privacy and security for voice digital mobile services are being studied but the synchronization and switching are much more difficult Q ee07 The interception threat is not trivial for a BEARCAT 300 scanner can monitor current mobile telephone channels easily However the problems of intercepting the switching circuits at 800 MHz is beyond a simple device such as a BEARCAT scanner 14 UB'f REriBA SA BbB 'fS es t IUlre sftS 15 I' 11 II It tl YEAR Source Motorola 20 GROWTH IN CELLULAR RADIO U Motorola has projected rapid growth in U S mobile radio equipment over ther next decade from less than 100 000 to over 1 4 million sets Since Motorola is the dominant company in U S mobile radio they are in a position to make the forecast materialize The technology will will be copied abroad in the countries with high telephone density IS CeO The impact on SIGINT of the growth of these switched public telephone mobile services is that very interesting information may become available in the UHF spectrum in major cities but it will present a complicated interception problem Usually a set of frequencies is assigned to each transmitter and in the Bell scheme a seven-cell cluster is used to provide geographic separation for frequency reuse SIGINT collection can thus be locked to the frequency plan as the radio switching signals are also transmitted However analog encryption will probably come into vogue since most conversations will probably occur in a single cell The mobile telephones are expensive so that the traffic will have a higher expected value than ordinary telephony This combined with the large market may encourage improved encryption systems that can operate over the switched radio links without loss of synchronization S eeS The Sep 82 0 CRYPTOLOG 'fBP SBSRfJ'f cellular scheme contains an Page 28 ItldUlL ' VfA eonIn eltA fftfi lbS BU-LY DOCID 4011953 'Pep sseRfJ'P indirect hint to SIGINT about how to do certain kinds of collection of mobile terrestial targets for the propagation and noise studies and the tracking systems could be adapted to SIGINT U In addition to the high-growth radio systems cited above there are a number of radio systems of special interest which will be mentioned later in the paper HORE FREE GOODIES oooo S sse Delayed somewhat by a war the third P14 How to working aid detailing UNIX PINSETTER techniques for the traffic analyst is in final draft and should be in distribution shortly This working aid titled How to Search details the mechanics of using the search keys and the binsrch grep and precog commands P L U Organizations and individuals already on the distribution list for these working aids should receive copies by mid-August If you or your organization is now or will be using TSS and UNIX PINSETTER and would like to receive additional copies or be added to distribution please contact I lin P14 on 3369s LUTION TO From lrm at geishg04 Subject Kudos To cryptolg at barlc05 cc lrm NSA-Crostic No 42 I ' C R' ' 'YPT O L ' O ' G- Ma y 19 8 - 2I NSA U Orchids again tol IWho e particular interests and insights never cease to interest and enlighten me As a frequent writer of documentation for several volunteer groups none of which have anything to do with computers and an NSA manager I found her June-July review to have almost universal applicability I spend a lot of time off the job documenting procedures ranging from the mechanics of writing a business letter to the sweeping problem of handling customers' complaints Mary's review has sent me scurrying to find a copy of the original article to see what other valid points the author might have on any area of instruction Most of the highlighted points are equally applicable to any personnel manager who is either training an employee or assigning tasks whether they be computer related or not How true the quote If you tell a user to do something he does not know how to do Substitute person for user and you've got a great management guide Information Desk Many times NSA employees have brought to our knowledge the fact that NSA has b e mentioned on the news or that they have read an article where the Agency has been mentioned We appreciate this ooo because it enables our ffice to keep I G 7 '1r---- 4 O O 7 S l S SSe I will be the speaker at the Autumn meeting of the KRYPTOS Society on 14 September 1982 j at ono in the Friedman Auditorium His talk entitled Twice Told Tale pertains to a description of the solution of a cryptosystem which had been solved before including analysis of indicators cIpher text and cipher alphabets as well as depth reading programming and historical research The talk is classified TOP SECRET CODEWORD Irm@geishg04 Sep 82 CRYPTOLOG 'Pep SHEiRH'P PI-Sop 8Z-S3-2 561 I the senior-level people better advise i - L 86-36 U I Icame to the Agency as a French linguist in 1952 The bulk of his career has been spent as a cryptanalyst in G He spent several years as an instrucGroup tor in the NCSch and is currently a member of the Cryptographic Skills Enhancement Program in P15 Keep 'em coming 86-36 Page 29 IhldffibB VIA sellm elh'dRIHbS e fi5Y D ' l ' J D o _ 4 CI 1 1 53 _- - --- IIIS BOEUR O H IHN eON AINS EUROBH wrORB MA HRIA' -teP-SEERa- 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