Gelsinger pointed to , an experimental network that sits on top of the Internet, as a step in the right direction. Hewlett-Packard and Intel have begun work trying to commercialize the project, which was started in 2002, in order to overlay the Internet with intelligence and adaptability. And the Public Broadcasting Service will use Planet Lab to broadcast high-definition TV shows, Gelsinger said in a speech here at the Intel Developer Forum.
Dual-core chips are a highlight
of the three-day gathering. For
all stories from the event, plus
video, click here.
"We think the work we're doing today is laying the foundation for the Internet of tomorrow," Gelsinger said, dubbing the fruits of PlanetLab work "the new Net."
Servers embedded in the network provide PlanetLab with the new layer of services, Gelsinger said. Those services include event processing to monitor what's happening; network mapping to arrange connections between computers; content distribution to optimize where information is stored; and Webcasting to make broadcasting efficient.
Gelsinger described two current research efforts that he said demonstrate the potential of PlanetLab to improve the Internet. The first was a project based at the University of California at Berkeley to combat malicious computer attacks. The so-called Public Health for the Internet project has developed a system for monitoring network attacks, and it can determine their source, Gelsinger said. Such information could be sent to firewall applications at companies or Internet service providers to bolster security, he suggested.
Another example of PlanetLab in action is work at Carnegie Mellon University to improve broadcasting over the Web. By setting up media proxy servers, the project takes the strain off the primary computer streaming out content and allows for a better-quality transmission, Gelsinger said.
Gelsinger provided a first-hand demonstration of the technology, dubbed End System Multicast. A display showed that a live Webcast of Gelsinger's speech deteriorated when several new computers connected to the main machine broadcasting it. But when a PlanetLab "node" was added to the network, it took over broadcasting duties for those computers, and the overall picture quality rebounded.
Gelsinger likened the Webcasting system to a more adaptive version of Internet infrastructure provider. "Think of this as PlanetLab-based Akamai of the future," he said.
Gelsinger applauded efforts by Cisco Systems and other networking companies to improve the existing Internet plumbing, but he argued that such work is not sufficient to deal with the complexity of, the shift to and other issues.
To lend weight to his opinions of the Internet's limits, Gelsinger shared the stage with, who helped invent the Internet more than 30 years ago.
"I still think it's pretty primitive," Cerf said of the Net. "I think we're still in the Stone Age, when it comes to serious networking."
Cerf detailed a litany of Internet issues:
"There is a capacity problem," with tens of billions of devices expected to be connected, he said: "How is the Net going to support all these billions of additional devices and users?"
"Users and applications are going to demand much more bandwidth than the current infrastructure currently delivers," an issue that will be complicated by localized traffic jams that Cerf called "flash crowds."
With the growing breadth of the Internet, there will be increased variations in network service response times. That will mean that quality of service will be less predictable, Cerf said.
And while the private sector will be able to fix some problems, tax and crime policy "have to be tackled by governmental or intergovernmental framework(s)," he said.
PlanetLab today involves 150 universities, along with Intel, HP, AT&T Labs, France Telecom and NEC Laboratories America: Now the project is poised to grow beyond those roots, Gelsinger said. "We believe it's time to begin commercializing services built on that network."