Buried among the flotsam and jetsam of the Web, among the sites offering sex and soap and family pictures, are some amazing examples of how the Internet can link the world.
And perhaps change the future of computing.
Today, people use personal computers and the Internet to send pictures and words to other humans. But what if all those computers started talking to each other?
The old adage about "putting our heads together" could take on new meaning if personal computers pooled their power.
Think about it: Tens of millions of personal computers around the world spend a large portion of each day idle. Collecting and using all that computer time could allow humans to solve some of the world's greatest puzzles.
Sound ridiculous? Three young computer experts have done it.
Our story begins with RSA Data Security, a company that few have heard of but most computer users have encountered. RSA provides the algorithms (computer "formulas") that provide security on most of the computers and software we all use.
Since RSA is in the business of locking up computer data, it needs to build impenetrable locks. The company knows that while it develops better computer locks, the folks who want to pick those locks are getting smarter, too. It's like a game of one-upmanship.
The problem is the bad guys - the real bad guys who electronically crack into banks or Pentagon computers - don't advertise their successes. So RSA created a series of hacking challenges and put up money to attract the best.
The latest "Secret-Key Challenge" was a 56-bit encryption code that many thought was uncrackable. RSA offered a prize of $10,000. Not a lot, considering that the winner had to search through 72 quadrillion (that's 72,057,594,037,927,936) possible answers for the right one.
The code was cracked in October.
That's where we pick up the second part of our story.
RSA knew that cracking its code would require a computer to work very hard for a very long time to crack the code. But what if lots of computers worked together on the answer?
With tens of millions of PCs sitting idle in homes, the idea of linking them to handle massive computer chores has been bandied about for years. But the three computer experts decided to try it to win the RSA challenge.
Using the Internet to find others interested in the project, the three were able to draw in more than 4,000 teams of computer users.
The computers churned away for 250 days, testing and discarding 7 billion codes a second. After testing 34 quadrillion codes, the project ended when a Pentium Pro PC in a Belgium university found the winning code.
The effort was coordinated by Distributed.Net - http://www.distributed.net -
run by Adam L. Beberg, a Chicago software engineer; Jeff Lawson, a student at Harvey Mudd College in California; and David McNett, an Alabama computer programmer.
While Distributed.Net can't accurately estimate how many computers worked on the project, it says the project had the computing equivalent of 14,685 Pentium Pros - or 116,326 Intel 486s.
So while this story might be interesting, what does it mean?
First, RSA made the challenge to prove that the 56-bit code (the strongest computer encryption code that the United States allows to be sold in other countries) only offers a "modest level of security." It wants permission to sell stronger codes abroad.
Second, Distributed.Net will tell you that we don't need supercomputers. We already have them. All over the world. We just need to link them up.
The group's goal was to prove that it can be done, that "distributed computing will be the 'killer app' that will make the Internet a necessity rather than a diversion," Mr. McNett said.
But don't expect to start selling time on your home PC.
Distributed computing is "probably several years away from being viable as a rapid solution to attacking a specific problem," Mr. McNett added.
Distributed.Net isn't the only group using the power of idle computers. Others are discovering new Mersenne prime numbers and analyzing random noise from outer space to see if any of it comes from intelligent life.
And what was the 56-bit code worth $10,000? It was 0x532B744CC20999.
E-mail Charles Brewer with questions, comments and suggestions at CBrewer@enquirer.com
BREWER ARCHIVE