HC Deb 14 February 1992 vol 203 cc1288-96

Motion made, and Question proposed, That this House do now adjourn.—[Mr. Boswell.]

2.50 pm
Sir Ian Lloyd (Havant)

I am glad that the hon. Member for Ashfield (Mr. Haynes) is leaving the Chamber. I fear that if he were to oppose to what I have to say in such stentorian tones, I would not get very far.

On the first occasion that I raised the question of computing in the House, I produced a microchip on which it was possible to see about four transistors—and on that I made some fairly far-fetched predictions. Today, it would not be possible to see the structure unless we had an electron microscope, and we all know that such visual aids are deplored by Mr. Speaker. I am saved the burden of elaborate descriptions by the timely publication of the Parliamentary Office of Science and Technology report just before Christmas. It is the implications for the nation and for Parliament that I want to discuss today, and I am glad to have the opportunity to do so.

It is not just the machine that I wish to discuss, prodigious though its capacity may be, but the institutions, processes and policies required for its effective use. It is not just an old mainframe, as they were called, multiplied by 1,000, 10,000 or even 100,000 times. The difference performance is both quantitative and qualitative. The problems that the new machines can address are those facing mankind, and they are of much the same order of magnitude. Some are known as the "grand challengers", and we all recognise their scale. For example, the data produced by a satellite measuring environmental change amounts to about 1 trillion signals per hour. That would require 40 Washington telephone books per day to store.

I shall give two more illustrations. The human heart consists of millions of cells that communicate electrically in a most complex pattern. To understand how it works or fails, scientists have to construct models. Professor Dennis Noble of Oxford university presented a fascinating paper in Paris this week, pointing out that, even with the fastest existing IBM machine, at present the modelling of a few seconds of the heart's activity requires several minutes. On such a machine, modelling 100,000 cells of what is known as the sinus, let alone the hundreds of millions in the ventricle, is extremely expensive and impractical.

The new type of machine, known as a massively parallel machine, makes that possible. Alas, Professor Noble has to use the United States army computing centre in Minnesota for his work. A new machine on the horizon enables him to envisage what he describes as a reconstruction of the electrical mechanisms of a heart attack. Such is the power of the teraflop machine which Columbia university aims to have developed by 1996.

It is hardly surprising that the machine has been described as the critical tool of the lifesciences in the 21st century". and an even more significant description is that it provides the human race with its third scientific investigative technique after theory and experiment. It has immense powers of simulation, which enable us to detect similarities and differences where size and complexity are simply too vast for the unaided human mind.

It is worth making one more point. The machine's technical rate of development is unprecedented for an artefact of such complexity. Its computational power feeds back rapidly into its own design, construction and manufacture.

Why is this important to Parliament and the nation? There are several reasons. Senator Gore of the United States Senate summarised it in a statement he made on introducing a Bill recently: The nation which most completely assimilates high performance computing into its economy will very likely emerge as the dominant intellectual, economic, and technological force in the next century. That is some claim, but it has evidently been fully accepted by Japan, the United States, Germany and France. I wish that I could say the same for ourselves.

The Japanese Government have formally stated that the development of massively parallel high-performance computing is that country's most important priority for the 90s and has allocated some £200 million for the development of what are known as 'neural networks'.

In 1985, we and the Japanese each had four so-called super-computers. In the brief space of five years, the Japanese have acquired 176, most of which they make themselves, outstripping the United States, which has 162, western Europe, which has 101, and the United Kingdom, which has 21. Even more alarming, perhaps, is that Mexico now has more super-computer capacity than we have. All significant indices of growth and expenditure are in the POST report, which is available to Parliament, so I shall not weary the House by repeating them.

Qualitative aspects are more significant. There is a virtuose circle of semiconductor capability and computing capability which the Japanese in particular have appreciated. They will design and make most things first, fastest and best, as the super-computer facilitates total quality control, among other things. That has particular significance for cars, aerospace, electronics of every kind, and therefore the worldwide pattern and distribution of employment. It is that realisation which has doubtless prompted the United States to set up a series of federally funded super-computer training centres, one of which, at Cornell university, is matched possibly only by that at Edinburgh. It has also prompted a series of United States policy responses, emanating primarily from Congress and the President. I shall describe them, as they are a template against which, I fear, we must judge our own parliamentary and Executive response.

Since 1989, the United States Congress has introduced at least five major Bills, starting with the National High-performance Computing Technology Act. Its committees have conducted seven hearings. It has established a national advisory committee on semiconductors, which concluded recently that, unless the United States spent $1.2 billion per annum on research and development in this area, it would not be a serious competitor in the coming information age. The President's Office of Science and Technology Policy proposed in 1989 that there should be a massive five-year programme for high-performance computing and networks. Expenditure over five years has doubled from $450 million to £1,000 million. Since 1976, as the OTA pointed out last year, federal funding has grown faster than for any other scientific discipline in the United States. Since their power is dependent on communication networks, the Administration has accepted a programme to raise the network capacity from the present 1.5 megabits to 45 megabits for 200 to 300 users this year and 1 to 3 gigabits, increasing to the order of 1,000 for major research facilities from 1996, when 45 megabits will he available to 1,000 separate institutions.

I now come to the western European and the United Kingdom response to the challenge. There is, I am sorry to say, only one word to describe our response in the United Kingdom. I had on my original draft the word "lamentable", but I thought that that was too harsh. It is a curate's egg—good in parts. Yesterday, I checked with the Library and discovered that, by comparison in the same period, we have had the subject mentioned in six questions in this House, in four in the Lords and in one early-day motion. In December 1990, there was a brief late-night debate on our contribution to the European framework, Jessi and Eureka programmes.

We have, it is true, a £34 million massively parallel computing programme, since this is an area in which we as a nation excel intellectually, as it were. After the POST report, I understand that the Department of Trade and Industry increased its funding for our network, known as JANET, by some £10 million to upgrade it to SuperJANET.

I further understand that the Cabinet Office has now set up a super-computing and networking team to consider the policy implications in the POST report. We have also agreed to take a third of a 25 per cent. share jointly with France and Germany in the European Community's megastream project. The other funds will be provided by the Community and Cable and Wireless.

Rather more depressing is the information that the Advisory Board for the Research Councils has recommended a bit more money for supercomputing over the next three years—the equivalent of £6 million, £7 million and £7.5 million. One must set that in the context of the £90 million spent per annum by Michigan university or any of the major Japanese producers or users. I fear that that will cause a powerful attack of dismay, which will overwhelm us and for which the rhetorical tranquillisers that we employ in this Chamber will prove no cure.

I fear that Parliament must take its full share of the blame for this state of affairs. In the past 18 months, we have screamed at each other across the Floor about unemployment and heavens knows what else, oblivious of the fact that we are witnessing events of far greater significance, which are far more compressed in time, than the industrial revolution.

The OECD, to which we all subscribe, is not unaware of those facts. In its report, "Interdependence and Co-operation in Tomorrow's World", there was a stark warning from Sir Ralf Dahrendorf, There are those who argue that employment is but a passing problem … Growth plus training, they say, will deal with all but the most resistant unemployment and in any case there is enough work for all. Such views overlook the deeper issue which is at stake … Work, in the sense of paid employment (jobs) has been the key to the entitlement structures of modern societies and to their growth potential. At the same time social, economic and technical factors have conspired to reduce the role of work in people's lives … work has become scarce … and as a result work based on entitlement structures can no longer be sustained … Growth and employment are decoupled. I agree with that.

The principal technical factor is information technology, which the worldwide motor car industry has been discovering for some time, as it responds to a competitor who made that discovery about 10 years before it. What is as clear as daylight is that, although there is every prospect of manufacturing output and world trade recovering, that recovery will not inevitably or necessarily produce a recovery in employment, especially not in the manufacturing industry. That industry is now following, rather more rapidly, the example set by agriculture, which, in the past few decades, has suffered a dramatic decline.

For that reason I accept Sir Ralf's final conclusion: Traditional big government cannot continue. It is financially non-viable and organisationally inefficient and fails to engender the degree of support that is a necessary condition of effectiveness … New entitlement issues are as much a result of old solutions as they are answered by them. Neither the new unemployment nor the new poverty can be remedied by the traditional instruments of the Welfare State. What are our policy responses? First the good news, as they say. I was informed only this morning that the United Kingdom has the largest single computerisation programme in western Europe. That is excellent, but with what is it concerned—wealth creation or wealth redistribution? Alas, it is the latter. Doubtless that is justifiable and necessary, but should it have come first?

I am advised that the Science and Engineering Research Council, acting as agent for the Advisory Board for the Research Councils is considering the purchase of a large Cray vector machine and that the ABRC has recommended a modest increase in expenditure on this of £6 million, £7 million and £7.5 million over the next three years. That is extremely modest when one realises that the world's most powerful machine, the SX3 made by NEC in Japan costs between £10 and £15 million for a single output. The bad news is that, although that particular Cray offers between a threefold and a fourfold improvement in performance, a major single university in the United States is taking delivery of a Cray which offers a sixteenfold improvement.

The main question for us is not whether it is too late for the United Kingdom or western Europe to compete with the Crays. It is worth remembering that Europe has not one single producer in this area. That round has been lost. The question is whether by appropriate policy decisions, we will be able to attain and maintain a satisfactory position in the massively parallel computers sector, based on the British transputer in which we and Europe have considerable strength and maintain the lead. Two of the five manufacturers in Europe are in the United Kingdom.

The answer is yes, but there are some ifs. First, we must recognise that this is what is known as an enabling technology, which is crucial to the economic and industrial performance of western Europe. Secondly, whatever our political prejudices—I know how deep they run in the House—we in the United Kingdom will not succeed unless we recognise that if our competitors' programmes are based on a massive combined effort of state and private resources, we alone will not be able to do it without an injection of public money on a sufficient scale.

Thirdly, we shall not succeed if we pitch the level of investment commitment just below that needed for success and withdraw support when the going gets rough. The Inmos story deserves a closer analysis than it has so far received, for yet again British innovation has formed the basis of a massive new industry and investment abroad. This morning's announcement by IBM that that large organisation was going for a massively huge parallel computer has great significance.

In that context, we should consider carefully the recommendations of the Office of Science and Technology Policy's report to President Bush, since they apply to us. One says, for example: Unless the nation acts immediately to promote its position in critical generic technologies, United States competitiveness will erode further with disastrous consequences for … jobs, economic growth and national security. Another says that the advanced technology programme is designed to provide support without picking winners and losers, and that that requires in turn a fundamental change in the way United States industry competes in the market place. The enabling Bill enjoys bipartisan support. Congress has provided $289 million this year and $347 million next year.

Fourthly, as a nation we need more thinking and analysis—with less ranting and raving—not least in this place. The whole profile of the science base, and science policy and its direction and priorities must achieve a much higher profile, and Parliament and Government together must do that. It cannot be done without an informed Parliament.

Fifthly, we must accept that the age of automation has arrived and will not go away. It is not a question of whether and how we should revive manufacturing employment, but whether and how to maintain the wealth-creating capacity of the manufacturing system, which is now following the pattern set by agriculture.

The concomitant question—how to restructure the whole of the entitlement apparatus previously attached to manufacturing employment—is far more difficult, and the sooner we start thinking about it the better. To do otherwise may afford political parties the illusion that they are shooting each other in the body, but in practice we shall all be shooting the nation in the foot.

Sir Herman Bondi, the former master of Churchill college, said recently, in a brilliant paper on Isaac Newton, that the problems faced by society could be classified into three parts—that 80 per cent. were trivial, 19.5 per cent. were insoluble and it required a genius to identify and solve the remaining 0.5 per cent. We have the tools and we know the 0.5 per cent. We must now finish the job.

3.7 pm

The Parliamentary Under-Secretary for Industry and Consumer Affairs (Mr. Edward Leigh)

I wish at the outset to welcome the report on high performance computing prepared by the Parliamentary Office of Science and Technology. As the House knows, my hon. Friend the Member for Havant (Sir I. Lloyd) is the distinguished chairman of that body. He is probably the leading advocate of science in the House, which is why we always listen with great care to what he has to say.

We are indebted to my hon. Friend for bringing to our attention the main findings of the report. He always speaks with great knowledge. Indeed, the House of Commons is at its best when an hon. Member with such deep interest and knowledge of a subject speaks on it. We in government do our best to address the problems that we face. I am grateful to my hon. Friend for the way in which he made his speech.

The Government share my hon. Friend's view that high-performance computing is a vitally important and rapidly developing area of information technology. It is right that the House should devote time to discussing that important aspect of technology, which has far-reaching implications for research and development activities, in academia and in industry and commerce.

There has been a great deal of discussion and activity in that area in many countries. The findings of the POST report will make a valuable contribution to the consideration that has been given by the Government to the priorities for the provision and application of high-performance computers in the United Kingdom.

My Department and the Department of Education and Science have a particular interest in that, but other Departments are also involved; for example, the Department of the Environment has important commitments to supercomputing, for climate modelling research.

My Department believes that it is important to emphasise the solutions which this high-performance can deliver and to promote the benefits that can be gained by industry and academia.

As for the development of high-performance computers, as the POST report identifies, the highest performance supercomputers are currently built exclusively in the United States and Japan. My hon. Friend brought out that point in his speech. But the manufacturers of parallel computers are much more widely spread, and this country is particularly well represented in both the design and the development of those machines.

The "Transputer"—a powerful "computer on a chip" especially suitable for parallel processing—was designed and built here, and a number of companies here have developed powerful machines based on that device.

The "distributed array processor"—an alternative approach to parallel processing—was designed and built here, and it received the Queen's award for innovation in 1991.

My Department and the Department of Education and Science, through the Science and Engineering Research Council, have given substantial support to research and development in the field of parallel architecture computers, through our joint framework for information technology; and I assure my hon. Friend that we shall continue to do so.

The really large research and development activities that are necessary today to produce the next generations of parallel machines are now being supported by the Commission of the European Communities, under the Esprit programme. We believe that that is the right way ahead for our companies and academics, rather than concentrating on solely national initiatives. Only through such an approach can our companies access the world-size markets needed for commercial success.

Our policy is—here as elsewhere—to ensure that the objectives of the European Commission's initiative are realistic and attainable, and that the expenditure plans are well judged and reflect appropriate priority for that among other areas within the Esprit framework.

The availability of high-performance computers has also been raised. The POST report considers two aspects of availability: the provision of machines of the highest performance to researchers in the academic community; and their availability to industrial research workers.

The provision of supercomputing facilities for academic research is the responsibility of the Department of Education and Science, through its Advisory Board for the Research Councils. I shall ensure that my right hon. and learned Friend the Secretary of State for Education and Science is made fully aware of the excellent points made by my hon. Friend today. That board now has responsibility for supporting the three academic super-computing centres at Manchester university, London university and the Rutherford Appleton laboratory of the Science and Engineering Research Council. My right hon. and learned Friend has recently made available the sum of £20 million over the next three years to enhance the provision of facilities for the research councils and the academic sector.

Many other university departments have both advanced computing resources and a pool of skills in high-performance computing with parallel machines. I believe that all that activity must be taken into account when assessing how well our academic community is provided for in this vitally important area.

To make the best use of the national high-performance computing resources, high-speed networks are required so that remote users can communicate with the machines and receive their results. To provide such facilities for the academic sector, I am pleased that the Department of Education and Science has been able to make additional capital funds available to the Universities Funding Council. I am sure that my hon. Friend will welcome that. This will allow them to begin the development of a new, high-speed academic network, SuperJANET, that will offer many benefits in addition to improving access to high-performance computers. I was grateful to my hon. Friend for mentioning SuperJANET in his speech.

My Department welcomes this activity and wishes to explore ways in which industrial users could gain access to such a high-speed network. Both academic and industrial users will also need access to efficient high-speed networking across Europe and worldwide. We are supporting current European network programmes and further discussions that aim to extend that capability.

Obviously, availability to industry is a matter of particular interest to me and, although my Department does not consider it appropriate to use public funds to make machines directly available for industry, we accept that there is a need for greater involvement by the industrial sector. We want to ensure that industry can make informed choices about the computing resources appropriate for particular applications.

We are aware that few organisations have yet installed their own high-performance computers, but some are exploring the possibilities of such machines, using the resources available from academia. Still more organisations are collaborating in our new programme to demonstrate the development of real industrial applications on high-performance machines. It is the Government's role to address sluggish development in the marketplace, and to provide encouragement to industry to address its own problems, but, as I am sure my hon. Friend agrees, the commercial decisions must always remain with industry. I hope that my hon. Friend will note the point that I wish to emphasise—that it is indeed our role to address sluggish development in the marketplace. Therefore, the debate is timely. As one of the constraints to market development, the POST report identifies lack of software and difficult programming as barriers to the industrial take-up of the machines.

My Department, with the Science and Engineering Research Council, has recently launched the parallel applications programme, to help to overcome those barriers. The programme has a total value of about £39 million, with £13 million of Government funding, and it is supporting four centres, chosen by competition, at Edinburgh, London, Oxford and—I am glad to say—Southampton. Those centres will collaborate with industrial partners to develop solutions to real industrial problems. Such applications will act as exemplars to other industrial organisations of what can be achieved with this technology. They should indeed then be able to offer the "one stop shop" support for parallel computing which the POST report recommends, based on the example of the university of Maryland super-computing centre.

I am sure that the parallel applications centres will make a major impact on industrial competence and take-up in this vital area. My Department is also considering whether it is appropriate to mount additional initiatives.

I have spoken about initiatives within academia, and the academic role in ensuring that academia has the skills and capabilities needed to prosecute its own research programmes. Academia has a role, too, in reducing the barriers to industrial take-up—by communicating new developments to industry in a way which is relevant and exciting, by seeking industrial involvement in appropriate projects, and especially by producing skilled graduates, who will move into their industrial careers and transfer that important technology with them.

In conclusion, I welcome the POST report, and confirm that the Government are sympathetic to many of the views expressed therein. We do not propose sweeping or dramatic measures in the short term, but I am confident that, with the initiatives already under way and planned, and with the technical developments on the horizon, the United Kingdom is on course to be in a much stronger position in high-performance computing over the next few years. The debate will have brought that process ever closer. I am grateful to my hon. Friend.

Question put and agreed to.

Adjourned accordingly at seventeen minutes past Three o'clock.