Open Source Is the Loom for Fabrics
In the Fall of 1998, Erick and Linda von Schweber, founders of the Infomaniacs think tank, published an article in eWeek predicting the evolution of "a new wave in computingone that we believe promises within the next five years to deliver almost limitless cheap computing power and to change the balance of power among technology vendors."
They called their idea "computing fabrics" and described it as "a new architecture" that would "erase the distinctions between network and computer" by linking "thousands of processors and storage devices into a single system."
Today, only three years into their five-year time frame, the concept is called "grid computing," and, increasingly, its hailed as the Next Big Thing. Fabric? Grid? The only difference is the density of the weave.
Recently, as I reread the von Schwebers prescient article, I realized that the one major phenomenon they failed to anticipate proved as important as all the things they got right.
Although Linux was already gaining popularity on campuses and in research centers in 1998, the phenomenon they underestimated was the open-source movement. As it turns out, it is the open-source Globus Project that is making grid computing possible in a wayand at a speednone of the proprietary initiatives they cited could have hoped to achieve.
In 20/20 hindsight, it now seems to have been inevitable. Open-source development is the human version of distributed processing, and it boasts many of the same efficiencies that make grid computing attractive.
The problem is that we tend not to take free stuff seriously. At about the same time Bill Gates was writing a version of BASIC destined to seed the worlds largest software company, Ward Christensen, an IBM systems engineer from Dolton, Ill., coded a little program he called modem, which enabled computers to send binary code to one another over phone lines. The difference was that Christensen released his code to the pubic domain, enabling other programmers to enhance and perfect a system of checksums and protocols that became the foundation of all computer communications.
Similarly, many of us tended not to take Napster seriously, and we smiled a little condescendingly at the thousands of volunteers donating computing cycles to SETI@Homes search for extraterrestrial intelligence. Both contributed to peer-to-peer networks and the distributed architecture of grid computing.
There is a lesson here. The next time some wizened vendor on the streets of New York counsels, "For free, takefor buy, waste time," take a moment to consider whether he might be right.
Are the best things in IT free? Let me know at firstname.lastname@example.org.