Visit Your Local PBS Station PBS Home PBS Home Programs A-Z TV Schedules Watch Video Donate Shop PBS Search PBS
I, Cringely - The Survival of the Nerdiest with Robert X. Cringely
Search I,Cringely:

The Pulpit
The Pulpit

<< [ Chase Cringely ]   |  Chase 2.0  |   [ Breaking the Bank ] >>

Weekly Column

Chase 2.0: Is that a supercomputer in your jammies?

Status: [CLOSED]
By Robert X. Cringely
bob@cringely.com

Last week's column about the death of my son Chase from Sudden Infant Death Syndrome set off something of a chain reaction across the Net. The column went out a bit earlier than usual, and by the time I awoke on Thursday, words of support were already in from readers in Japan, Australia, India, and Kuwait. By breakfast I'd heard from Europe, and by 24 hours later I'd heard from more than 500 friends around the world. A week later, the message total stands well above a thousand, with many of those � at least 30 percent � offering to help personally in Chase's SIDS monitoring project.

And the quality of those volunteering has been remarkable. Engineers with deep experience in medical devices and sensors stand out, of course, but the very breadth of talents offered is staggering. Programmers and poets, big idea guys and assembly coders, they just keep coming in. And we can use them all. There are even some names you might have heard before, like John Gage, chief scientist at Sun Microsystems, and Dan Bricklin, inventor of the spreadsheet. Slowly I came to understand that this little idea of mine, an idea that sprang from my grief over Chase's tragic, otherwise useless death, was taking on a life of its own. And it just might do some real good.

This is the Internet equivalent, I suppose, of a barn raising. People come together and volunteer their talents toward a common and laudable cause. And this type of volunteerism, more than e-mail and chat rooms and Web cams and Web logs and heck, the very Web itself, this volunteerism is the real essence of the Internet. It is something that literally couldn't happen any other way or through any other medium. There is very little in the world that is truly new,but this is new, and it gives me hope.

Growing from the Internet Engineering Task Force's Request for Proposal process and the Free Software movement, this idea of giving a few hours of deep thinking came into its own with the birth of programs like Linux, Sendmail, and Apache. In each of these cases, a hundred or more experts and enthusiasts built software not only good enough to compete with Microsoft, it is software generally BETTERthan Microsoft. Throws the concept of Big Business right on its head, doesn't it? The model is always being extended and Chase's monitor project is just the latest permutation.

Such collaboration simply wouldn't work without the Internet. When some engineer offers Chase two hours of labor per week, which is about the norm, the only way to get anything done is to eliminate meetings, eliminate travel, eliminate the effects of time zones, eliminate as much overhead and friction from the process as possible. And what's left over is the work, itself, which is made easier bythe absence of either a profit motive or a marketing department. Oh, it is possible to market these projects, but not until they are pretty much finished.

And it works beautifully. Very little money is required, but when it is required, somebody generally comes up with some. The only other currency is ideas, and all participants benefit from the experience. If they don't benefit, they drop out, just as they should. Unhappy workers just don't stay involved.

When Bill Gates first came to understand this process he said it sounded like Communism, but it is not. This is no way to run an economy, but one of the only ways to make major technical progress on projects that are otherwise noncommercial. And it works far better than begging for research funds. We'll have Chase's monitor in the field in less time than it would take a government agency to even acknowledge receipt of a grant application.

So here is what's happening with Chase's project. In a few days there will be a web site at http://www.chasecringely.org, where you'll find a picture of Chase and a collaboration software package called TWiki, which will be the only place most of Chase's volunteers will ever meet. Come meet us there and join up, or just follow our progress.

The monitor is already coming along. We have tentatively selected Eleven Engineering's XInC processor for the monitor's brain. XInC is a multithreaded RISC processor designed by Canada's best supercomputer engineer and it is a remarkable chip. XInC runs at around 50 MIPS, which is nothing special today, but the chip is designed to be uniquely tolerant of latency, which is to say itscore logic can run on silicon that would never work for another processor. Processors are expensive to make because they are big and because their big circuits have to be carefully timed to work together. XInC works differently, with the result that it can be manufactured with the CMOS manufacturing process normally used only for memory chips, not processors. This means that memory can be built into XInC, or XInC can be made as part of a memory chip (intelligent memory?). Whatever you call it, XInC is cheap. We hope to build the monitor as a system-on-chip including analog-to-digital conversion, processor, and memory, all for a couple bucks in volume.

We're exploring new sensor materials, too, including some that can be used once and thrown away, though the greatest benefit may come from making the monitor into a wearable garment. Our goal is still to build a $10 device.

The primary use of the monitor, of course, is to gather data to use in SIDS research, hoping to find ways of better identifying babies who are at risk and of predicting when an attack is imminent. For help with the data analysis, we'll be relying on the good people of the SAS Institute and the Santa Fe Institute, among others. Chase has lots of friends.

I told you this project has attracted some big idea guys, and the big ideas are flowing freely. What else can we do with this monitor? Let me share with you the current brainstorming. There are approximately three million babies born each year in the USA. What would happen if we could get a million of those babies to wear Chase's monitor to bed as part of their jammies? It could changepublic health as we know it and save thousands of additional lives.

Sure we'd put a dent in SIDS, but actively monitoring a third of all babies would let us easily detect and map infectious childhood diseases, which could be treated earlier and more effectively. By knowing exactly where kids were getting sick, we could keep other kids away, keeping them healthy. Where there are sick kids, there are often sick parents, too. Lives could be saved.

But how do we pay for all this? Part of the answer is to build it cheaply, as we are already trying to do; another part is gaining the assistance of those who would benefit from healthier babies and families. You can bet we'll ask Wal-Mart to throw its purchasing heft behind those high-tech jammies. My 78-year-old mother lives in Bentonville, Arkansas, where Wal-Mart is based, so Chase is putting her on that job. Watch out Wal-Mart!

But this program will still cost money, so we need to find a clever way to make some so more babies can participate. Here is the best idea I've heard to date to make that happen, thanks to a Harvard PhD candidate who wants to make Chase's monitor the subject of her dissertation on pervasive computing. We'll turn all those monitors into the world's largest supercomputer, and let people pay to runjobs on our big jammie machine.

Here is how the system would work. The monitors would function independently during the day, watching for serious health conditions, crunching their own data, and triggering alarms. The XInC processor has eight threads, and no more than two of them would be required for basic monitoring. The other six threads could be used for pre-processing gathered data before uploading it to the network, orperhaps for running other highly-distributed jobs. At night, the monitors would be connected to a power supply that also contains a modem chipset connected to the home phone line. This is how the monitors could communicate with each other while the babies sleep.

It doesn't really require much communication. The monitors can log on, send and receive data, then log off in perhaps a minute or less. At least 60 monitors could share a single Internet dial-up account, for which wholesale ISPs typically charge $6 per month or 10 cents per monitor per month. That is $100,000 per month OR LESS to network a million baby monitors.

One million monitors, running at 50 million instructions per second each, is an aggregate computing power of 50 trillion instructions per second. Figuring only 10 percent availability or efficiency, that's still a widely distributed supercomputer capable of five trillion instructions per second. Want to decode a secret message or render Toy Story 4? The jammie computer can do it while saving baby lives at the same time. And with a capital cost of around $10 million and a monthly cost of around $100,000, by selling excess cycles, it should be possible to save babies at no cost to society at all.

Comments from the Tribe

Status: [CLOSED] read all comments (0)