Posts from SICSE: Lazowska suggests green IT as next space race

Posted on April 3, 2008


Ed Lazowska is chair of the Computing Community Consortium. He gave the last keynote at SIGCSE (slides here).

He describes the 20 greatest engineering achievements of the 20th Century. Computing is listed, as are electronics and the internet. Perhaps even more important though is the role of computing in ALL of the other achievements from spacecraft to health technologies to agricultural mechanisation.

Lazowska rallies against people who think that nothing much has happened in computing in recent years, or who think that it is all solved. He describes how in 1995 a list was compiled of computing innovations that would spawn billion dollar industries. A relook at that list in 2003 showed just how many things they had overlooked (the web, broadband last mile, data mining etc). He sees that the “future ahead is full of opportunity”.

There are challenges ahead that will shape the intellectual future and must attract the best of the generation.

and as we shall see by the end of his talk – those challenges are much about computing and sustainability.

Lazowska talked widely about the difficulty in predicting the future of computing. He used the old quotes (Watson’s five computers, Popular Science 1.5 tonnes, Olsen’s no reason) and then looks at trends that have influenced computing (Moore’s law, an equivalent in software, exponential growth in internet hosts, 1billion PCs representing 2% of processors).

As we’ve said many times before, computing underpins almost everything we do.

Or as Adam Osborne puts it: ‘The future lies in designing and selling computers that people don’t realize are computers at all”.

The Computing Community Consortium was formed to

create compelling research visions and the mechanisms to realise those visions.

I like the questions they set themselves to envision long-range, more audacious research challenges; to build momentum around such visions; and, to state them in compelling ways.

Any formal body that has the word audacious as a stated goal has to be good. But I worry about the technology focus of the group. Is this an unfair characterisation?: “society has been fundamentally changed by the big science we’ve done in the past, let’s do some really really big stuff to change it some more”.

I can’t see much evidence of driving this the other way (what do we need to fix?) on the CCC website. Giving them the benefit of the doubt for the moment, let’s look at where Lazowska thinks we’re going. He uses illustrates a vision for the future of computing with the Grand Challenges for Engineering (incidentally, note the overwhelming dominance of providing more energy in the public voting). Lazowska highlights 5 of the challenges as being computing, and two more as heavily involving computing.

Lazowska goes on to discuss his 14 most important drivers for the next generation of computer scientists.

eScience – sensor driven science. Through programmes such as the Ocean Observing Initiative and the digital sky survey and at all scales from molecular to global, computing is driving the automatic discovery of knowledge in massive amounts of data, in the words of the 2020 Science group: the building blocks of a scientific revolution. Computing is enabling a transformation of all areas of science and nowhere, says Ed is this more critical than in the area of sustainability:

There is no more important problem than our environment – this is the space race for today’s generation

Empowering the developing world

3 billion people in the rural developing world need the same information we do

Lazowska talks about projects such as Tapan Parikh’s CAM: Managing Information from the Grassroots where “information systems are the key to scaling microfinance” and computing facilitates expert advice, remote organic certification and so on.

Harnessing parallelism

I was surprised by this one and the implications for computing for sustainability. Lazowska (using arguments from Mark Oskin) points out that for 50 years computing performance kept up with transistor density (ie Moore’s law). After 2004, however, performance has dropped off: the “failure of the traditional architecture to keep up”. Lazowska says we’re in for a “parallel revolution – ready or not”, the “new Moore’s law 2x processors per chip every two years”. So what’s the implication for sustainability? Lazowska poses the question – “what if IT goes from a growth industry to a replacement industry?”. Instead of buying new computers, we’re swapping in cores, but even then currently software can’t effectively make use of that speed – there will be no advantage to buy a new computer. Here’s one answer – Blevis won’t have to be looking to the marketing department to make it trendy to have a retro computer – the new business model for computing is going to be based on it.

“The algorithmic lens – a computational perspective transforms the sciences” (eg Cyber-Enabled Discovery Initiative). I only hope that the computational sciences are also open to being transformed.

Wreckless driving is the continuing potential for computer augmented driving. The big challenge is to move the successes of the DARPA desert successes to urban conditions, but includes, from our perspective, vastly improved fuel management and adaptive cruise control systems. There are similar benefits in neurobotics personalised health monitoring.

In passing, I also note and agree with his assertion in entertainment technology that there is considerable need for content creation tools. In Simpa we are using games technology to help re-engage indigenous storytelling. I couldn’t agree more with Lazaowska who argues that the “expectation of user interface far exceeds our ability to create”. I think that this though that this problem is not just restricted to the entertainment area.

Personalised education. Lazowska talks about the opportunities that education might bring the developing world. The challenge is providing that education (in conjunction with the emerging technical solutions such as one laptop) . The solution, he sees, is not technology but rather what technology can enable. We know that 1:1 expert, adaptive tutoring is hard to beat, but this doesn’t scale. The challenge, then is adaptive tutoring systems that scale.

Lazowska finishes with myth busting in computing (coming soon on CCC). Mostly these are about changing perceptions of careers in computing (for example: Myth: there’s no jobs in computing anymore. Reality: there are far more people employed in computing than the height of the .com boom). One important myth is that “computing lacks opportunity to making positive changes to society and the world”. This is where we come in, although, as someone commented at the end of the session “few people are starving for lack of a laptop”.