ICL COMPUTERS FOR THE USSR

Created: 2/22/1971

OCR scan of the original document, errors are possible

Briefing Materials on ICL Computers for the USSH

Page-

Back

CIA Intelligence

Becommendstions to the NSC Undersecretaries Committee k

Comments on the OST

Attachment A Operating Characteristics of Selected Computers

Attachment B Business Week Article on Serpukhov

Attechment C Proposed Safeguards for High Performance Computers

Attachment D Recent Soviet Efforts to Obtain Large Western Computers

Attachment E Computer Capability of the USSR

Attachment F "Pros" and "Cons" of US Approval

Background

In

the British computer firm. Internationalas considering the export of large COCOM-embargoed computer systems to the Institute of High Energy Physics at Serpukhov, USSR. In addition to being embargoed, the computers in question contain US peripherals end components and therefore require US licenses. In0 the British forwarded er, side memolre to the US outlining its proposals to export the the above-mentioned computers end asked for US agreement prior to submission to COCOM. The UK formally submitted the proposal to COCOM in September without waitingesponsefrom the US.

The computers proposed for export to Serpukhov include the following: two interconnected computer complexes each consistingA computerA computer linked3 communicationsreeA; and various peripherals. The value of the transaction isillion. COCOM currently limits thehe processing data ratef computers that con be exported to Communist destinations without COCOM approval. The current cut-offillion bits per secondAER of kk mbs; the PDRA Is lessbs. (Seeor characteristics of selected computers). The PDR for each of the proposed interconnected systems would be in excessbs. The cabinet-level Export Administration Review Board (EARB) decided in0 on US policy for handling exporta of computers that exceed COCOM cut-offs. The EARB determined that the risk would be low if one computer per year in the rangebs were exported to the USSR, but that the risk would be

* The processing dBts reteeasure of the ceoabilityomputer topecific calculationiven time period. Peripherals and software are Important determinants of the rate.

"signifleant" if more then one of that size wereomputerDR of more thenbs vould require the endorsement of the EARBblue ribbon" case.

The proposed consignee, the Institute of High Energy Physics at Serpukhov, reportedly vould use these high-power computer systems to opgrade Its capabilities in basic research. The British claim that the only computers now evsileble at Serpukhove PDR of aboutbs If operating et designed capacity)ew small Soviet snd Western computers and that no large Soviet computers will be available in the foreseeable future. Additional computer power is said to be needed now for use with theillion electron volt (Gev) particle accelerator, currently the largest in the World.* In defending its proposal, the UK cited the open nature of the institute's work, the institute's international standing, the Soviet need for large scsle computer power to further its research, snd

The US response to the British side memoIre in0 vas negative. The principal resson cited was that the computers would be easy to divert to stretegic us<

lile admitting

tne peaceful nature ot tne worK at serpuKhov and its Internationa standing, the US emphasized that it vould be extremely difflcuLt with current technology to detect diversion without full time, expert, on-site monitoring and that proper monitoring of the work of the computer would require

* Theev particle accelerator at Batsvis, Illinois will be the largest in the world when it becomes operational this year. ecent article In Business Week discusses the work at Serpukhov and its cooperation with Western Institutes

SE&iltl

constant sompling of the sub-routines for evaluation by an expert.* Such an evaluation would be further complicated by the fact thet high energy physics programs are similar to thoee for weapons development and very few experts are capable of distinguishing between the two. Bubble chamber experiments could resemble experiments on nuclear veapons effects, for example. US authorities slso expressed concern st the implicstlons which the export of these high performance computers vould have on maintaining COCOM controls over advanced computers in the future.

The US formally objected In COCOM in earlywo weeks lster the British forwarded another aide memolre stating that lt could not accept US statements particularly with respect to the danger of diversion. In aid-December British newspapers sired the US-British dispute on the computers end indicated thst Prime Minister Heath would raise the matter with Prosident Nixon, rxirlng his visit to the United States later in December, Mr. Heath expressed his interest in getting US approval of the transaction.

ICL'b use of important British government officials to Influence tbe US to approve transactions with the "JSSR, es notedime-honored tradition. TCL'a proposalto sell third-genorstion computer technology to thebu formally launched in this country by high-level Board of Trade officisla. Similarly, the intercession of the Minister of Technology, Anthony Wcdgcwood-Benn,0 vith Deputy Secretary of Defense David Pocksrd was instrumental In securing US approval of en0 computeroviet research institute having military affiliations.

* Earlyesult of Soviet interestor Serpukhov, US authorities tentativelyet of operational conditionsigh performance computer could be instslled it Serpukhov without prejudice to national securityhese conditions apparently form the basis for US objections relating to diversion. An earlier report by the Rational Academy, of Sciences concluded that it would be impossible to0 against the possibility of diversion.

** Recent Soviot effortp to obtain large Western computers arc summarized in Attachnent D.

British government support Is not surprising because ICL Is the only major British-owned computer manufacturer snd it hes had difficulty in meeting foreign competition. Although this transaction is valued atillion end ICL total ssles in the entire East European market0 were only about $L0 million, ICL believessst market can be tapped In the USSR and Eastern Europe. Consequently, it isajor sales effort in the sreo and is the first end only Western computer firm to be permitted toales office In rfoscow. ICL also plans to help the USSR to developn area in which the USSR is particularly weakew series of Soviet-designed computers.

Tho ICL computers are not the first choice of the USSR for the Serpukhov Installation. The Soviets would prefer theespite the fact that the throughput of the two ICL systems reportedly would exceed thathe large amount of occumulated Western experience with the latter tn similar applications would prejudice the Soviets towardoreover, the use of interconnected computershe63atheringle machine could introduce additional complications. The CDC would also cost the Soviets less than half vhot they are prepared to pay for the ICL computers-The negative US response to'informal Soviet inquiries concerning the installationO at Serpukhov evidently led the USSR to seek the ICL computers.

CIA Intelligence Contributions

CIA has not been called on toorms! Intelligence input in this case, but hasumber of them on Soviet computer capabilities In connection with T

in recent years. ummary staieEent ui curmnu Soviet capabilities in the computer field.

Recommendation to the HSC Undersecretaries Committee

The Under Secretaries Committee was asked. Kissinger onsnuary to review the question of the export of the ICL computers to Serpukhov. Mr- Nathaniel Samuels, Deputy Undersecretary of State end acting chairman of the comrittee,ecommendstion of spproval for consideration of the

The "pros'" end "cons" of lifting the US objection to the export of the computers to Serpukhov listed in Mr. Samuels' drsft are in Attachment F. It is believed that en important "con" is omitted end namely that the upgrading of high-energy physics progrsms et Serpukhov uould elso enhance the international prestige of the USSR .in this importsnt area of research.

CCsssents on the OST Report

The recommendation for approval vas based largelyechnical report submitted by the President's Office of Science end Technology (OST). The rationale for approval Is contsinedentence in the OST report: n technical grounds the degree of potential security.risk to the US represented by this transaction ia extremely On the whole the technical analysis contained in the report is good, but certain considerations appear to have been overlooked. The OST report lists seven conclusions:

first conclusion states that the Soviets could not extract equipment technology from the machines. Our technical experts agree.

tates that the effort required to effect clandestine diversion could not be justified by the Soviets unless use of the orderf the machines were obtained-This Bssumesenefit will accrue only if available machine time is adequate for the complete design of one warhead.

diversion

tates that the diversion of one-quarter of Serpukhov'Ns computing time Is In the range where the loss of computer time for legitimate high energy physics needs would be noticeable by Western scientists working and familiar with the research program. Wo believe that it is doubtfulcientist

could detect oi aizefloLe parts ol the computer's capability.

There is greater merit to OST's contention, however, that the Soviets would be reluctant to tBke the risk of revealing Soviet weapons codes to Westerners.

tates that If diversion attempts were being planned, provision for core dumps on demand and examination of the printout by US specialists would Increase the chances for detection and further reduce the risk that the Soviets would make the attempt. We believe that the dump provision canontrol onlyump can be madeS monitor wishes. It does not follow thet an ICL employee could be trusted to perform security functions. The OST report implicitlyeed for US participationontractual basis. The British, however, state that "they would not wish to extend these contrsctuol rights to persons other then ICLnless there IsS monitor who Is competent toarge dump, this provision has little value.

tstes thet the software systems for the machines should be readily adaptable to use of Western high energy physics programs and would be inappropriate for Western weapons codes compiled in machine languageO). The very tight security controls on US weapons codes, in general, would preclude their falling Into Soviet hands. It is believed that further consideration should be given to the possibility of Soviet sceess to British-designed codes; the extent ofA compatibility with computers used for British weapons codes la not known.

We concur withhat further explanation is needed from the British on why such large storage capacity is needed.

We have no comment onhich notes that Western scientists could benefit from cooperative projects at Serpukhov.

Operating Characteristics of Selected Computers

Model

Bus Bate (mil. bits/sec) a/

Processing Data Bate (mil. bits/sec) b/

0

5

5

o

a.Central Processing Unit (CPU) bus rate is the rate that data can be exchanged from the CPU to the internal memory. Computersus rBte ofr more exceed COCOM guidelines, b- COCOM Processing Data Bate (PDR)easure of the capabilityomputer topecific calculationertain time period. Peripherals and software ere major deterninants of the rate. ComputersDRr more exceed COCOM guidelines.

of the architecture of the0 series,cannot be calculated- The rankings are thepositions basedumber of factors includingfees.

USSRDR slmller to that of therobably does not perform at maximum designof failure to provide adequate memory end softwareof provisions for input-output.

{Pooling braoimshe alom

scienlisls are participatingrojecl at Russia's reactor center

The soaring oral of research ia high energy physics is activatinginternational cooperalloa-

Next month, frre Americsn UsU will travel lo Protvino, nearGO mi. south of Moscow and site of the world's biggest atorn smasher, or "particlehey will wort withix-month projectIh* piiny subatomic par-lickhought to contribute to the forces that hold the atom's nucleus together. The Americans willartrouporeign scientists at Protvino, where the Russians have cur.science city" out of birchto house the giant accelerator.

Meanwhile, In Switzerland, thehare been working with theitfd ceis (Conseil Europeeo pour la Recherche Nudeaire)nd using cew's atom smasher. Andall over the world are eagerly awaiting the completion of another new slom smasher in thel Ba-tavia. III. next yearwo Russian scientists spent several weeks at Ba-taria this summer, planning passible experiments on the accelerator.

Because each of the new machines willmillion or more to build, the giant atom smasher is nol Ihc kind of research equipment every country can have. In facl. the high cost has made high-energy physics the mostof all fields of scientificThus the cooperation. Method. The basic concept of atomic particle research is simple. To learn more aboul the nature of matter,have to understand the particles that make up the atom. Moreew particles have been discovered In the pastears, ranging fromto heavy biryons. Their sites are measured In billionths of inches, their lives in billionths of seconds.

Because they are so tiny andthe subatomic particles are on the borderline between energy and matter. The only way to produce them is, in effect, for the physicist to hurlbigger, more manageable pur-tides, such as protons, at close to the speed of light. Energy that comesout of the crash is converted briefly into the subatomic particles.

Tbe physicist achieves the speeds necessary for this proceis bythe protonsircular tunnel, which is lined with magnets to hold the proton* away from its walls and spood them up. He detects the subatomic par-tides by the -tracks" they leave as they travelas-filled or"bubble chamber- To push protons up to tbe speed of lighthuge equipment, however.ring al Protvino is nearly adiameter and the facility cost anto build.all over the world covet thisenormous energy ofvolta (orev.ut they insistneed an even moresmasher, one withev or more lo reallyparticles. And applying a of thumbarticle accel- costs upachine would cost

Acfuevemnta. The precedents for cooperation that achieve this goal have already set at Protvino and CEBN. At- the accelerator atonly throe yeara old,have achieved some imprti-

For one thing, they have dis-covered matter calledoughly comparable to ordinary helium, but with opposite electrical charges. And they have cast serioui doubts on accepted theories of howchrrgod particles interact with other particle* at high energies.

Dr. Romanyear-old dep uly director of Ihe Center of High Energy Physics st Protvino, is proud of the scientific successes at hisBut he says his colleague* are at-

U.S. will acquire the world's snost powerful atom imaahcr-sooner. cheaper, andigher energy level than its planners origi--nally expected. Gonstructiin at Ba-lav Is,s going "exceedinglyays Edwin L. Goldwasser, depuly director of the installation, known as the National Accelerator Laboratory.

The atom smasher, being built far. Atomic Energyis due to start operations nextear ahead of schedule Its cost* are running below Iheoriginally estimated. lisanticipate that It willenergy levels upillion Geneva electron voltsay over the W> Gev they had looted for ia early planning.

Lock, good management, snd ad-Taocing technology have all helped to brighten an otherwise bleak scene in U. S. physics, hard-hitudget cutbacks. Several largecontracts were let for the accelerator just before President Nixon's order last September lo cut back federal constructionnther contracts have cone In below estimates. Says Gold-wasser.ecession, contractors are hungry Moreover, they are ready to move in with noconomies. So far. about SS9-million has been spent on the sal, almost entirely on the accelerator. No money has even been allocaleda. ft main researchwhich has beenower priority by sal Director Dr. Robort

Rathbun Wilson. Labs archoused in other buildings.

Also to promote economy, Wilson is encouraging competition between his contractors. As one example, he has ordered only two-thirds of theagnets needed for themain ring, from twocompanies. The company that does the best production job will get the order for the remaining third.

Technical advances account for the boost in the accelerator's ratingev Designers found that circuit-switching devices called thy. ristors could lake higher electrical loads than anticipated, permitting use of much greater electric power.

The nal will be operated forC by Universities Researchonsortiumniversities While some experiments will be car-ried out by the UL staff,e performed by scientists from these and othereito ie*

- iSjavBTat. wo

ready working to eipand lU capacity. They ha reinitiated:

A series of international nuclearnder which both men and supplementary ref arch equipment arc beingto Protvino from Europe and the U. S.

A program to explore whether super, conducting magnets, nowcan be used to boost its rating. Cooperation- The closeness of thecooperation isby the projects under way atozen Preach scientists there areiter bubble chamber calledrenchmen will Join them when components ofmfllion sphere, built at the French Atomic Energy Commissariat's SacluyCenter outside Paris, arelo Russia and assembled there. MirarxUc will remain French property, but will be used by the French and Russian*ive-year agreement.

Sovlet-cns cooperation it even closer. High energy physicists shuttle "back and forth between Russia und Switucrlnnd, where cersev atomormal cerv-Soviet agreement has been in effect for three yean.oaen eras scientist* at Protvino Save launched the secondcnes of projects, begunS, to study particles called heavier neutral nx-rani Last April, the CT.HS groupons of equipment to Trot vino.

including an0 computer, all to be returned to CESS.

At the same time,ovietIn Geneva are working with highly sophisticated systems that have been developed atfast ejector system to switch particles out of the accelerator and radio frequencyaoparatoni- These systems will be shipped lo Protvino inear for permanent installation there. Payorf? Soviet ties with cern could pay off for the Russians if ander European accelerator is duHl. Years of argument over financing and locating this proposed atom (masher caused seven of CERN'sartners, in-

Suporconducting magnets, would sharply boost atom smashers' energy

eluding England, to opt out of the

Last Jane, however, project director Dr.riton, urged cms toadically new proposal. Under its terms the new accelerator would straddle the Swiss-Frenchal Mcjrin andGev atom smasher that is, already there at part of its system

The Adams plan looks cheaper and quicker than earlier proposals. It could mean only eight years'ost ofTherefore. CERN officials hope that it will lure the

Bonpartidpating countries back Into the fold, and they lookernby this Christmas on whether to go ahead.

A vital part of the Adams plan is that, initially, the ceax accelerator would get only half the number of nag nets it is capable of using. This would giveating ofev. But space would be left for installation later of much more powerful magnets-cryogenic superconducting magnets.

These magnets arc atill in thestage but could permit energy level*ovov. If they do not work out, Uie accelerator would get another sol of iron-core magnets,aximum ratingev.

Russia, too Superconducting magnets, which are under intensive research at Britain's Rutherford High Energy Laboratory, among other places, are at the heart of Soviet research lo boost the rating of the Protvino accelerator, too. According to Sulayev, use ofmagnets would boost the atom-smasher's energy level three to eight times-that is,aximum of moreev.ot of workto be done in thi* field-but itefinitee says.

Officially, the u. S. has madeto cooperate withThi* winter's visit of theteam, headed by Dr. Darrellot the University ofLos Angeles. i> based on anof letters betweenchairman ofommission, and Achairman of Ihe Sovieten the UtilisationEnergy. Bji scientists inare hoping forA wider, morevould be

Proposed Safeguards for High Perfornoncc Computers

In0 US computer experts agreed thst installationomputer of thelsss vould proceed without threat to national security provided the conditions noted below prevailed. These safeguards ere describedeport dated lU9 by

the President's Office of Sclenco and

Technology!

Closed shop, with systems programming done by US personnel only, adequate surveillanceurance of non-use during any off hours.

Batch computing only (no terminals).

3- Fortran programs only, with adequate documentation required to ensure efficient operation ond for verification -purposes.

k. Complete recording of input and sampling of output.

5- Sampling of internal computer executions.

6. Creation in the United Statesart-time group of high-energy physicists, computer center managers, ond weapons designers tomall semple of the recorded data.

Recent Soviet Efforts to Obtain

In8 the UK Informed the US that it was considering the sale of technology to the USSR to produce the third-generationeries of computers initially0 series ands. Theachines sre produced under license from RCA. They are compatible with theeries and canull range of peripherals. The US opposed the sale on the grounds that the acquisition of Western state-of-the-art computer production technology would facilitate Soviet production of computers which could have both military and civilian applications. The British subsequently withdrew the proposal.

In the Spring9 Control Data Corporation appliedS license to0 scientific computer for use at Yerevan, USSR to process data generated from experimentsev synchrotron. Consideration of the case was delayedtudy of the risks of diversion involving the proposed Installation0 at Serpukhov where the USSR6 Gev particle accelerator. The USSR had Inquired informally whether the US would exportomputer to Serpukhov if application were made. US officials concluded that It would be extremely difficult to'detect diversion without full-time, expert, on-site monitoring and such monitoring would be impractical- No application was made to export the0 and the0 application was denied in the summer

9 the UK asked for COCOM and US approval to export twos to the USSR, one to the USSR State Planning Committee (Gosplan) and the other to the Institute of Automation and Telemechanics (now called the Institute of Controlhe Gosplan case was approved by the US end COCOM In relatively short order, but because of evidence that the Institute of Control Problems was engeged in military-related research, the US refused to approve the case. Following the intercession of the Britiah Minister of Technology with Mr. Packard at Defense, however, the case was approved in

By the end0 the USSR had producedigital computers. This is leas then one-tenth of the number of computers lnstslled in the United States- About half of the computers in the USSR arc small, the other half being In the medium range vith fewer than one percent being classified as large computers- Of the0 computers Installed in the United States, about two-thire's sre smell (in the0 range) and most of the remainder are medium-size. The United States also hes0 large computers.

"These comparisons are not entirely meaningful, however-The majority of the computers installed in the United States arc third generstion machines equippedariety of input/output devices vhlch permit relatively full use of the central processor's capabilities. Soviet computers are almost entirely second and first generation machines with inadequate peripherals and software. Largely because of Inadequateequipment and software, but also becuase ot obsolete design and limited memoryoviet computer hes "Significantly lees capability'to' perform useful work thenS computer with virtually the some operating characteristics, bus rate, or processing dsta rate.

In terms of the COCOM processing data rate, the USSR Is known to have produced only limited numbers of one model which exceedsegabit cutoff. That computer is thef which fever thannits sre thoughtto have beer, produced. Even these few machines ere unable to perform at maximum design levels because of the failure of the Soviets to provide them vith adequate memory end software.

The major shortcoming In Soviet computer production Is tholr inability to develop satisfactory production technology, as opposed to design technology. esult, the USSR1 is able to produce only limited numbers of obsolete machines vith poor reliability. They are in no position to begin producing large numbers of third generation computers comparable to those produced in the US5ithout substantial Free World assistance, the USSR probably vill not produce large numbers of third generstion computers of acceptable quollty and performance before then.

"Pros" and "Cons" of US Approval

"Pros"

3-

be responsive to Prime Minister Heath's personal approach to -the Iresldent.

The on-site inspection arrangements couldossibly useful precedent for future arms control negotiations-

Would provide increased opportunities for Western cooperation In high energy physics at Serpukhov end might cause some general improvement in the Soviet attitude toward cooperative scientific ventures.

Would ease pressures for sales of computer production technology to Communist countries by demonstrating that occasional seles of advanced hardware can be approved when adequate safeguards ere available-

Would incur the security risk referred to above.

Would. end other firms to attempt

seles of large computers to research institutes in the

USSR and Eastern Europe.

Would_ cause some increased pressure principally by American manufacturers for relaxation of the embargo

Original document.

Comment about this article or add new information about this topic:

CAPTCHA