Peer-to-peer computing promises to revamp the software industry but could also help jump-start some of the plodding segments of the PC industry, Intel's chief technology officer said Wednesday.
The corporate PC market could become one of the indirect beneficiaries of the movement toward P2P, or distributed computing, which allows computer owners to string together ordinarily disparate computing resources, Intel CTO Pat Gelsinger said in an interview Wednesday at the Intel Developer Forum in San Jose, Calif.
For the past few years, corporate IT buyers haven't had many compelling reasons to buy their employees top-of-the-line desktops. Voice recognition, graphic-intensive applications and improvements to office application suites have been tossed at corporations in recent years, but with limited success, he said.
That could change if P2P takes off. With P2P, thousands of desktops could be strung together to run applications that would ordinarily require an expensive mainframe.
"You are now literally offsetting mammoth supercomputers. Now you have a compelling purchasing argument for the IT environment," he said. "It is the kind of argument that has been lacking in the business computing world."
Of course, corporate managers would have to buy new, upgraded desktops to ensure that workers would have the sort of spare computing power required to participate on a P2P network.
"Hallelujah," Gelsinger joked.
While all this sounds fine on paper, a number of analysts have said the prospective market for distributed applications is small. In the broadest application of distributed computing, after all, companies must be willing to loan spare computing power to outsiders. Security, protocol and other issues loom.
Gelsinger acknowledges that acceptance will be gradual. Companies will use these applications internally, then gradually let close customers into the network.
|Meta Group says that for individual consumers, P2P promises more control over their systems and a return to more "personal" computing. For businesses, however, it's a whole different ball game.
expect corporate IT managers anytime soon are going let anybody use their spare MIPS," he said, referring to their unused computing power. "But how about their supplier or dealer network?"
Distributed computing has the potential of advancing a number of technologies, such as Universal Plug and Play, which simplifies how devices plug into networks. Applications developed out of the P2P effort could serve as "glueware" for plugging gaps in the Universal Plug and Play infrastructure.
"I think this will be a facilitator for a lot of those applications," Gelsinger said. For example, "it would be pretty trivial to create a network across my home."
How and when these P2P applications emerge, however, depends on a number of variables. One of the first issues that needs to be ironed out is that the industry needs to come up with a coherent set of policies and practices for creating or maintaining distributed networks, he said.
A consortium of companies called the Peer-to-Peer Working Group recently anointed a structure and technical subcommittees. The group will try to come up with a framework for its work during the second quarter and define specific architectural requirements for P2P networks by the third quarter.
"The purpose of the Peer-to-Peer Working Group is to arrive at an agreement
on what the (P2P) framework will be," he said. Ideally, he added, applications will start to emerge by the fourth quarter.
The emerging P2P industry could also benefit from the establishment of standards. A number of start-ups have been founded in the past year hoping to capitalize on the P2P trend.
But for now, about 70 percent to 80 percent of these companies are spending their time resolving technical issues. But once network architectural standards are established, these companies can move on to develop more ornate applications and services.