Regional Advantage: Part 1- Pitting California's U.C. Berkeley against Belgium's IMEC
Rabaey described an environment of ambient intelligence that will provide energy management, environment controls, security, health care monitoring, advanced user interfaces, and entertainment. Accomplishing this requires novel wireless devices, which are compatible and reconfigurable. This amount of flexible computing will require teraflops of data, and high degrees of portability and scalability.
The answers will come from dual core, multi core, FPGAs, and a variety of operating systems.
* Dr. John Canny and Dr. Marc Davis speaking about "The Future of Mobile Applications"
Canny and Davis said there are 6 billion people on earth, and in the next year 1 in 8 will buy a cell phone. That's 4 times the number of PCs and TVs combined. Today's phones are fairly simple, voice-based SMS devices. But programmable, media-rich smart phones will bypass those earlier models to serve not just as communication devices, but as the most common compute device on the planet, while providing an ever deeper feature set including video, gaming, camera, storage, and that final holy grail feature - to know where you are at all times and what time it is.
Canny and Davis want to penetrate the Third World and overcome the lack of access to digital technology by using the phone as a CPU in conjunction with an inexpensive keyboard.
* Dr. David Wagner speaking on "The Future of Software Security"
Wagner said the Bad Guys are finding bugs faster than we can. He asked the audience three questions:
* Do you use Microsoft Outlook for your e-mail application?
* When was the last time you backed up your computer?
* When was the last time you downloaded the latest patches to your security hardware?
He said networks and firewalls are getting clobbered by applications purposefully tunneling through and over HTTP - and virus detectors are a total kludge. "It's just patch and pray."
Developers need to try harder, develop hardened languages, tools to find bugs in legacy code, and exercise better programming discipline. Wagner is working on a C model checker that's showing promising results.
* Dr. Chenming Calvin Hu speaking on "The Future of CMOS and Memory"
Hu spoke at length about CMOS. He said despite critical problems with power and quantum tunneling at small process nodes, there are ways to get around these things and CMOS still has a long future ahead. He said there are a variety of strategies that will ensure that future, not the least being a personal favorite of Hu - the finFET.
Hu said the world continues to want CMOS, and won't be turning to self-assembled systems anytime soon because the manufacturing infrastructure worldwide is already in place for CMOS. It's too expensive to re-tool that global infrastructure and the world can't afford to gamble on unknown technologies. It's for good reason that Dr. Hu is the TSMC Distinguished Chair Professor of Microelectonics in EECS at U.C. Berkeley.
Per Hu, "CMOS will be a platform for deploying innovations in materials, devices, circuits, and nanotechnology. CMOS is a mutant monster. You can knock it down, but it will only absorb your energy and grow stronger. I'm not discouraging alternative research, but our budgets need to not be based on winning the lottery."
* Dr. David Patterson speaking on "The Future of Computer Architecture"
Patterson said we desperately need a new architecture that embraces parallelism. Also, the old paradigm has changed. It used to be that power was free, but transistors were expensive. Now power is expensive and transistors are free. It used to be that chips were reliable internally, and the errors were at the pins. Now there's a huge increase in hard and soft errors on-chip. The old wisdom was that innovation came through compiler optimization. Now that innovation takes more than 10 years.
We need a renaissance of performance, which can be had through using multiple processors. Looking at recent product announcements in industry, now the processor is the new transistor. We need to design complex systems with hundreds or thousands of processors on chip. The future is multicore and everything's going to be parallel.
Patterson and U.C. Berkeley are part of an ad-hoc consortium of players from MIT, the University of Texas, CMU, Stanford, Intel, and the University of Washington who are assembling a design protocol for these massively parallel systems. They've had success assembling a 1000-processor system on 40 FPGAs, and they hope to announce more over the coming year. The group is calling itself The Watering Hole, and I think they're serving up more than just well drinks. They're serving up algorithms, programming languages, compilers, operating systems, architectures, and libraries.
Patterson invited those in the audience to get involved. "How would you like to get a preview of the computer of the future to do software development on today?" He didn't ask for a show of hands, but it was obvious that many in the audience were tempted.
All of the speakers presenting their research on February 23rd at U.C. Berkeley were completely captivating in their energy and enthusiasm. It's a great faculty with lots of exciting things going on. If these are the Best and Brightest in The Bay Area Corporation, we've got a great future ahead of us!
Epilogue: How ideas are born
At the core of everything we do, as technologist and human beings, is the question of how ideas are born. How does innovation happen? And once achieved, how is innovation implemented to both improve society and enhance the financial circumstances of a company, although not necessarily in that order. Of course, the question is far more complex and n-dimensional that these brief words would suggest - even the definitions of "society" and "company" are open to debate - but you get what I'm saying.
If you're reading this column, you're probably interested - in fact, involved 24x7 - with the question of how ideas are born, in high-tech in particular. Well, giving birth to ideas in high-tech is very tough, and it's tough for two reasons.
You've got to be really, really smart to come up with new ideas - in general, not always, but in general you need to be formally trained in engineering, or computer science, or one of the basic sciences - and you've also got to know how to collaborate with other smart people, whether they're annoying, unkempt, tidy, proud, self-effacing, wealthy, poor, optimistic, pessimistic, foreign-born (a relative term), mono-lingual, or mono-syllabic.
That collaboration thing is a difficult nut to crack, but without sorting out the process - how to work as a team to come up with new products and services that are based on the ideas that come out of the individual brains of the individual team members - nothing that we commonly refer to as "high-tech" can, or will, move forward to that next great sign post on the horizon. Absolutely nothing.
Now, along with the critical dynamics and skill sets associated with teams in high-tech, there's also the problem of protecting intellectual property - those ideas that you yourself give birth to at 3 o'clock in the morning while everybody else is sleeping but you aren't, because a problem you're working on won't leave you alone - and the ideas that are contributed to the team effort from other equally obsessed problem solvers, which when put together with your unique ideas creates a whole that's bigger than the sum of the parts and therefore constitutes the intellectual property of the team. Intellectual property that, for the sake of the group, needs protection from idea-raiders.
These are not easy things to do in a corporate setting - the generating of new ideas, "team think," and protecting intellectual property - but they're even harder to do in academia.
It's true that academia is about being smart, and working with other smart people in a collaborative manner to birth new ideas and concepts - not to mention, teaching and mentoring the young - but that part about protecting your intellectual property and that of your team? That's a much "dice-ier" problem in academia than it is in industry.
How new ideas are born in academia - and protected from voracious carnivores in other ivory towers, not to mention industry - is a problem that all major academic research institutions deal with. Because academia is, in principle, about sharing. It's about promoting smart people and getting their ideas out into a public forum, so those ideas can instigate ideas from yet other people who are also interested in exploring the same solution set or subset. Academia is about cooperation, but it's also about protecting the institution's intellectual property when there are carnivores around.