CHAPTER 11: CONCLUSION
Introduction
11.1 As we have found, the United Kingdom has
substantial but under-recognised strengths in microprocessing
part of one of the world's largest industries. Building
on these strengths, there are substantial and exciting opportunities
for the United Kingdom. Action is needed by Government and others
as identified in our various recommendations[106]
if those strengths are to be maintained let alone expanded
in the global market for computing.
11.2 Our detailed recommendations are made throughout
this Report, where they are highlighted in bold type. For convenience,
all the recommendations are collected together at the end of the
Executive Summary in Chapter 1. In this concluding Chapter, we
summarise the principal conclusions arising from our Inquiry.
The need for ever greater
computer performance
11.3 Demand for computing performance has increased
rapidly over the 50 year life of the computer. We see no prospect
of this abating. The need for more accurate modelling
for example, of the oceans and atmosphere for climate prediction[107]
or of the airflow through aero engines[108]
creates an insatiable demand at the top of the performance
range for supercomputers of the highest feasible computational
performance. As Fujitsu noted (p 196) it is, even in consumer
products, already possible to see how a hundred times the current
performance could readily be employed to deliver higher quality
moving and still images.
11.4 However, sheer speed is not the only factor
by which the performance of computers should be assessed. In many
of the latest consumer applications such as mobile telephony
low power and functional performance[109]
are equally important. (In passing, it is worth noting that mobile
telephony is not only a consumer application. As the Ministry
of Defence noted (p 207), secure communications in the military
domain rely on increasing mobile computing performance.)
The current UK position
11.5 There are significant UK industrial strengths
in the design of microprocessors and related components. This
seems not to be widely recognised, probably because the companies
(most of which have developed only recently) are generally small
and diverse in their activities. Academic strengths are in evidence
too, although the activity in universities is limited to a relatively
small number of groups that can claim to be world-class in their
research activities.
11.6 As far as mainstream CMOS semiconductor
manufacturing is concerned, the United Kingdom (like many other
countries) no longer has a significant presence, and there is
no realistic prospect of this position changing. However, as CMOS
technology matures over the next 15 years or so, we see room for
alternative technologies particularly in niche areas.
When new technologies are introduced, there is an opportunity
for new players to gain a foothold in the market, and the United
Kingdom has promising research activities in alternatives to CMOS.
These research activities should continue to be supported but
with better co-ordination and focus. As the winning technology
has still to emerge, resources should be allocated broadly in
response to promising lines of research.
Building on strengths
11.7 The existence of a thriving UK design and
architecture industry is a major strength upon which future policy
should aim to build. Maintaining and developing this industry
requires not only a robust base of appropriate academic research
but also depends on the availability of a suitably skilled workforce.
Both of these need attention if the industry is to maximise its
potential to contribute to the UK economy in the coming years.
11.8 One of our key recommendations is for the
establishment of a national research institute to provide an appropriate
focus for a new industry/university collaboration on research
and development. The brief for the institute to pursue SoC initiatives
might usefully be given greater impetus by pursuing particular
challenges for example, in ambient computing or biomedical
systems. This new institute should also be supported by a wider
networking programme, on the lines of the e-Science initiative.
11.9 It is also essential that more attention
should be paid to the supply of skills and IP into the UK design
and architecture industry. The expansion and better co-ordination
of academic research in this area (which will lead to the expansion
of postgraduate research training) and the expansion of Masters-level
training will both contribute directly to satisfying the needs
of industry.
The need to improve
exploitation
11.10 As the UK design and architecture industry
is characterised by a dynamic and diverse mix of small and medium-sized
companies, a major mechanism for the exploitation of research
is through start-up companies. Typically, these may be spun out
from other (larger) companies or from universities. Alternatively,
they may be started by independent individuals.
11.11 Whatever the origin of such start-ups,
however, there is a requirement for an entrepreneurial culture
where technologists, entrepreneurs and financiers mix and network
on a regular basis. The establishment of clusters of start-up
companies in a localised area, as is happening in the Cambridge
and Bristol areas in the United Kingdom, will enhance the opportunities
for economic benefit that we found so evident in Silicon Valley.
Encouraging the market
11.12 The chip technology and design matters
at the centre of our Inquiry are at a remove from the new high-technology
products that will capture the imagination of the mass market.
The role of chip technology and design in facilitating those products,
and the long lead times involved, are insufficiently understood.
Leaders of the UK computing industry and those who help fund the
ground-breaking developments should do more to inform and enthuse
the rest of industry and the finance sector about the possibilities
and, in a two-way dialogue, the way that they can be realised.
11.13 The public sector is an enormous user of
computing power in all its forms, and its procurements are significant
events in the market. We think the Government should develop a
more strategic view of the positive influence aggregate public
sector computing needs could have on the market and the R&D
on which the market depends.
The difficulty of making
projections
11.14 Our Inquiry has attempted to look ten to
twenty years into the future. The dominant CMOS chip technology
still has some years of development to come before either the
fundamental properties of matter will make further progress impossible
or, before then, manufacture becomes too expensive. On the technical
side, the global industry's Roadmap (the ITRS) makes projections
that extend out almost to the end of the twenty year period. In
the past, these projections have been met and, in some
cases, surpassed.
11.15 While it may be possible to project the
development of computer technology with some degree of reliability
ten or more years ahead, it is much more difficult to predict
the development of high-technology products not to mention
the evolution of the industries that manufacture them. This is
amply illustrated by the case study of mobile phones in Appendix
10. Who, in 1980, would have confidently predicted that:
(a) there would be European dominance of the
industry;
(b) a microprocessor designed by a UK company
would be integrated into over 75% of handsets; and
(c) UK mobile phone users would be sending each
other a billion text messages every month?
What the future holds
11.16 What is abundantly clear, however, is that
computing's integral place in every aspect of our lives will extend.
That will result from both evolutionary and radical developments.
Evolutionary changes
11.17 On the evolutionary front, the dominant
and already powerful CMOS chip technology will mature into a relatively
inexpensive workhorse. Some advances in raw computer speed (or,
for the same speed, improvements in power consumption) will flow
from continuing miniaturisation down to the ultimate CMOS transistor,
but most of the developments needed here are already in train.
Many more improvements in computing performance will flow from
better design and architecture, in which the United Kingdom has
considerable world-class strength.
11.18 There will also be evolutionary advances
in the applications and devices with embedded computing
and we cannot emphasise strongly enough that the consumer market
is driven by these rather than the enabling technology. Improved
applications and devices are not necessarily dependent on any
advances in either chip technology or design. Evolutionary advances
in those areas would doubtless encourage improvements in applications
and devices, and revolutionary changes (such as the effective
implementation of artificial intelligence) would seem bound to
lead to new products.
11.19 We have seen the potential of near-future
microelectronic technology to deliver pervasive computing devices
that are everywhere around us: unobtrusively monitoring the state
of buildings, vehicles and appliances; and quietly communicating
with each other so that we can access the information we want,
whenever and wherever we want it. Some of these devices may be
on or even inside us, monitoring us to improve the quality and
lower the cost of our health care.
Revolutionary change
11.20 In the longer term, such systems may be
able to anticipate our wishes, requiring far less explicit instruction
than today's machines require for even the simplest of tasks.
Computing would then become an information utility like
other household utilities, simply on tap when needed and used
with little or no thought about the delivery arrangements. Through
evolutionary change, the way we live may be revolutionised.
11.21 The prospects for revolutionary change
in chip technology seem less certain. In our view, the new technology
would have to offer significant advantages of one kind or another
over the established CMOS at least in niche areas
to stand a chance of success. At the moment, quantum computing
seems the most radical breakthrough area. However, merits are
not always easy to spot in the short term and may, indeed, be
different from those for which the research was devised. There
are several revolutionary possibilities that are worth pursuing.
Summary
11.22 Our lives are already heavily dependent
on computing and will become more so. There seems nothing unrealistic
in the aspirations we heard from a number of witnesses to develop
ambient or pervasive computing. Moreover, these could be achieved
within even the current chip technology.
11.23 The United Kingdom can either consolidate
and expand its contribution to this global growth industry or
watch others overtake us. We are clear that there are opportunities
here which the country would be foolish to miss.
106 As repeated in paragraph 1.24. Back
107
As described in the memorandum by the Meteorological Office (p 206). Back
108
As described in the memorandum by Rolls Royce (p 218). Back
109
The complete performance of the integrated system, as opposed
to just the microprocessor within it. Back
|