On Aug. 2, right above a story about a new and cheaper generic version of Prozac, The New York Times announced the Internet of tomorrow. If the prediction comes true, network managers may be glad Prozac will be getting very cheap in a few months. But, sorry Prozac makers, it will not come true.
The headline read “IBM Making A Commitment To Next Phase Of the Internet.” The article described “grid computing” and said, “The grid vision is that everyone at a desktop machine or handheld computer could eventually have the power of a supercomputer at his or her fingertips, by amassing the processing power and information resources attached to networks.”
The basic idea here is to reach out over the network and use the idle power of computers all over the world. A widely used working example of this is seti@home (www.seti.org), which has been installed on a few million computers in an attempt to find signs of intelligent life elsewhere in the universe. The grid folks (www.globus.org) see a wide application of this type of technology from weather prediction to earthquake simulations. It makes sense for that type of application, and there are more than a few applications that fit this picture.
OK, I’ll bite: Why do you and I want this? If the Times‘s headline is right and this is the next phase of the Internet, then it should be useful for ordinary mortals. It is true that some software vendors seem to have development projects to increase the size and complexity of their software to compensate for any increases in memory size and processing power, but even they will not be able to make use of this much power. And I don’t think I need my Palm forecasting the weather on its own.
IBM Corp., Microsoft Corp., the U.S. government, a number of European governments and other groups are putting a lot of money into this grid vision. But only some of this money is being well spent.
The grid technology will be quite useful for a number of applications, and the number of these applications may grow over time. But I do not think the grid will transform the Internet.
The grid vision depends on solving difficult problems and on the assumption that there will be general utility for the idea. One of these problems is figuring out how to compensate the owners of the computers for their use. Without this being solved, I fear that the technology will only be useful for feel-good projects, such as listening in on E.T.’s calls home.
I also fear that, like the Distributed Computing Environment (better known as DCE), the cost of management will be higher than the cost of replicating the resources for most applications.
I expect most of the Internet will be unaffected by the grid vision. Remember that one of the big strengths of the Internet is that it can support any number of limited-use visions such as the grid.
Bradner is a consultant with Harvard University’s University Information Systems. He can be reached at [email protected].
