The Napster story is distributed computing. Net Ventures hed: Napster’s workshop dek: The Napster story is distributed computing. By Sean M. Dugan
I swore on a stack of Bibles I wouldn’t write a column on Napster. Writing articles about Napster is too easy–like making jokes about George W.’s hat size or Bill Clinton’s pants. So, this is not going to be a column about Napster. At this point, what hasn’t been said about the music-swapping service? I’m not going to paint a picture of noble upstarts trying to revolutionize a corrupt industry–or alternately, deplore the evil perpetrated by a band of thieves bent on robbing people of their intellectual property.
I refuse to hype peer-to-peer networking as the solution to all life’s problems, or declare how even if Napster falls, there’s Gnutella, FreeNet, OpenNap or a dozen other efforts waiting in the wings to accomplish the same thing. I won’t opine on whether there’s a technology solution to filtering copyrighted music on Napster. I’ll avoid sharing how thrilled I am about the prospect of a hard drive from IBM that will fink on me if I have copyrighted material. I won’t invoke Courtney Love’s low opinion of the record industry. Neither will I blithely place my faith in the record industry’s ability to come up with a workable, equitable technology solution.
I’m not going to say all this, because the battle the music industry insists on fighting (soon to followed by film, broadcast, and publishing) is doomed. Losing this particular fight is inevitable.
Let me state for the record that computers are all about copying. Your computer copies data from a hard drive to RAM. E-mail messages are copied across the Internet. Data is cached all across the Net to speed things along. This is the potency of the Internet. Computers are all about copying digital data–to make them incapable of copying is an uphill battle. It goes against their fundamental nature, like making George W. ponder deeply or Bill Clinton keep his trousers on.
In fact, I’m going to break my own rule and say that it’s ultimately an unwinnable battle.
Why? Because Napster is a testament to the power of idle resources. Napster and the whole peer-to-peer sharing concept doesn’t work unless you have a lot of people, with a lot of spare hard disk and bandwidth resources, who are willing to share it in exchange for access to other people’s music. Since nobody ever has enough groovin’ tunes, Napster is a triumph of untapped resources being harnessed to “solve” a problem. And here’s the kicker: There are a lot more resources–and problems–where that came from.
If you’ve been paying attention to Moore’s Law and its corollaries in data storage and bandwidth, you know we’re getting more and more processor speed, hard-disk space, and bandwidth. More than anybody will ever use, which means there are vast computing resources to be harvested out there.
Distributed computing has been around almost as long as PCs. Napster is simply the first commercially successful distributed-computing application. Distributed computing is a slippery concept, but one worth getting your head around. In short, it’s the idea that many hands make light work. If a math problem is too tricky for one computer to solve in a reasonable amount of time, why not break it up into discrete chunks and have dozens, hundreds, or thousands of computers work on it? This is exactly what the SETI@Home program is doing. You download a screensaver that uses a small amount of your computer’s power to search space for extra-terrestrial signals.
But searching for alien life is just one of the practical uses of distributed computing. In fact, it’s the probable future of computing in general. Napster Inc. doesn’t need to buy 6 terabytes of hard drives when every user contributes a sliver of space. This also applies to bandwidth and processor power.
Imagine an Internet based on Napster-like distributed computing. A Web site that doesn’t “live” on one server, but is cached in any number of locations, including your personal hard drive. If you need to store some files, does it really matter where they’re physically located, as long as they’re secure and you have speedy access?
The Microsoft.NET initiative is Redmond’s first foray into this arena, but there are also things like Sun’s Jini, and Lucent’s Inferno. The whole model of the Internet is one of distributed, decentralized resources. Now apply the idea of distributed computing to business.
Nature abhors an empty hard drive, but idle resources are the devil’s workshop. So are there good ways to use those idle resources? Plenty, I’m sure.
Sean M. Dugan welcomes all your good ideas–see if he’s got any at www.seandugan.com.