De Man On the Need to Merge Unconnected Techno-cultures
This article is a brief summary of some
thoughts presented by the author to the Long-Range Planning Committee
of the IEEE Solid-State Circuits Society. Hopefully, these thoughts
can help start initiatives in and stimulate thinking about the future
of our discipline in the decades to come.
Where Are We Heading?
We are gradually leaving the boom and bust of the PC and the dot-com
eras. Both were driven mainly by Moore's law, leading to an exponential
increase in computational power at ever-lower cost. However, mostly
it has been business as usual, building more of the same only faster
and at a lower cost. But what is next?
On the one hand, all too often we have considered Moore's law to belong
exclusively to the semiconductor or the solid-state circuit world and
we may have overlooked that exponential growth also occurs in many other
areas such as sensor networks, nano-electronics, network bandwidth,
MEMS and NEMS, biotechnology, medicine, and drug discovery. This creates
a wealth of new opportunities.
On the other hand, we are living in times of uncertainty as we evolve
from a technology-driven world to one filled with user-centric applications
where the modeling of human behavior becomes as important as the modeling
of technology. This is a time when new system applications, created
for the well-being of society, should emerge from many visionary R&D
projects, combining a wealth of technological disciplines. We must take
into account that economic growth becomes a global issue and that in
western society future health care for the elderly, security, adequate
energy provision, and mobility will cry for technological solutions
invented by the best engineering teams we can find. Needless to say,
the best technology also must become affordable for those who have no
access to it today.
All of this is happening at a time when venture capitalists have become
afraid to invest in new ideas. We may have entered a period similar
to the sixties where ambitious space and military research provided
the foundation for the innovations that made the PC and dot-com boom
happen. Progress in individual disciplines occurred by combining these
disciplines in designing challenging global systems. So maybe we should
embark on new endeavors in multi-disciplinary system-driven research
and development involving international collaboration of academia, industry,
Merging Presently Unconnected Techno-cultures?
Since the laws of physics in the coming ten to fifteen years will make
further scaling of our traditional CMOS electronics a bit meaningless,
we must reflect on where to look next. Already today, MOS devices leak
energy, signal integrity becomes a problem, interconnect limits performance,
and dynamic power grows beyond reasonable bounds. Designing around these
effects takes all creativity out of the application engineer. This should
not happen at the point when technology is badly needed to solve societal
challenges of the future. Engineers should design around complexity
and into usefulness instead.
Around the year 2010, at the 30-nm node, the scales of nano- and micro-devices
will meet. But even if further CMOS scaling becomes problematic, the
system opportunities for future societal information systems will be
virtually limitless. From here on we must look at what we can do with
this technology at the system level by combining it with the rapid progress
in other domains. We believe that the merger of different techno-cultures
offers ample system opportunities to make Marc Weiser's dreams of pervasive
computing and communication or ambient intelligence come true.
However, this will not be business as usual by just doubling transistor
count or clock speed. Most likely, progress will be the result of merging
presently unconnected domains and techno-cultures to create the ultra-low
power, low-cost, always-connected devices operating on novel energy
sources for as long as the system's lifetime. In contrast to the past,
and contrary to the usual techniques of managing complexity, worlds
ranging from advanced software such as artificial intelligence down
to molecular electronics will have to merge to build these systems.
So, more than ever before, we must bring design-oriented people together
to brainstorm on how to organize a dialogue on facilitating design methodology
over the discipline boundaries. A great result could be that they learn
each other's terminology and way of thinking, as well as open the road
to innovation to solve societal challenges. The best way of getting
there is to bring the best engineers together in open, long-term, team-based
projects with a clear goal to create innovation by merging their expertise.
Examples of such experiments are:
Organic electronics: throwaway electronics in medicine,
smart pills, smart security tags, smart walls, flexible organic displays,
organic solar cells, and chemo-sensors
Silicon sensing of bio-molecules and intelligent
interpretation of it (designing in four layers: biotech, molecular,
sensing, and intelligence)
Compute devices for bio-informatics
Designing novel e-grain architectures in ultra-thin
silicon in 3D silicon structures and interconnect architectures as
needed for ambient intelligence networks (layers of antennae, sensors,
RF, baseband memory, intelligence, and energy provision)
Ultra-low-power intelligent memory architectures
for wearable data-mining and artificial intelligence search systems
to be used in context-sensitive pervasive computing
Creating artificial senses.
The biggest challenge will be to bring together visionary people who
are willing to start a dialogue on what future micro-systems will look
like and how we will join forces to make progress in this new field
of "super-circuit" design. Of course the need will remain
for traditional conferences such as ISSCC, DAC, DATE, and ICCAD where,
every year, specialists move the boundaries in their domains by another
five percent. However, we are approaching a time when we will owe it
to Gordon Moore to take much larger steps while still focussing on the
nano-electronic systems of ten years from now. The exponential nature
of Moore's law indeed makes the next ten years equivalent to fifty years
of the past! No wonder that progress will have to be made in a different
way, urgently and by joining forces. Why shouldn't the "old"
microelectronic society take the lead in this act in trying to set up
a system-level dialogue between top system-oriented researchers in the
Bio-electronics, health, and comfort electronics
Nano-electronics (both top-down CMOS and bottom-up
CMOS RF and analog systems
Technology around CMOS (MEMS, NEMS, organics, packaging
(interconnect, 3D silicon)
Configurable architectures and intelligent networks
on a chip
Networking and 4G wireless
Advanced (embedded) programming and human interfacing
Medicine and drug delivery
Societal engineering, sociology, and psychology.
Notice that I deliberately use the word "system-level," meaning
that we want to understand what we can do together to build meaningful
future systems. This may seem a bit like dreaming, but I believe that
such a group of people recruited from a number of convinced visionaries
(à la Carver Mead), managed carefully, could break through barriers
and establish new opportunities and projects. These new projects may
have a great impact, not only by creating new industrial activity, but
also by contributing to a number of growing societal problems such as
health care (aging, disease prevention), mobility, security, energy,
environment, education, and poverty.
The first thing to do is to bring multidisciplinary visionary people
together in international workshops with the goal to set up a "Book-of-Visions"
much in the same way as the wireless research forum is doing (http://www.
wireless-world-research.org/Bookofvisions/Bov.html). There are many
models for doing this (Gordon conferences, Aspen meeting, International
Workshop on Future Information Processing Technology, etc.). SSCS should
take the role of leading this from the solid-state perspective. It could
fundamentally change the Society and its impact on society. We could
devote at least one session in the traditional conferences (such as
ISSCC, DAC, DATE, and ASP-DAC) to the most valuable visionary outcomes.
Also, tutorials could be organized to make engineers familiar with merging
technologies. It would be inspirational to the industry as well as the
research and educational worlds. Engineers could pick up the ideas and
contribute to a new boost to the economy, to society, and to the well-being
of all citizens. What are we waiting for?
Hugo De Man