 |
|
|
|
|
| |||||||
 |
August 28, 2001 | |||||||||||
 | ||||||||||||
 | ||||||||||||
 
|
|
|
||||||||||
Embedded projects take a share of Intel's research dollarsBy David Lammers
Self-assembling "ad hoc" networks, sensing "motes" dropped from airplanes, even a four-node IP network adapted for use on distant Mars are all on Intel's research agenda.
David Tennenhouse, a vice president and director of research at Intel, spoke Monday (Aug. 27) in a kickoff presentation at the Intel Developer Forum (IDF) here. Intel's goal is to evolve from a company that dominates the PC processor space, which consumes about 200 million units annually, to a major player in the embedded processor space, which accounts for 8.5 billion units a year, he said.
Some of these research efforts are well known and are well on their way to becoming real businesses. Four processors have already been developed for third-generation cellular phones, and represent a huge opportunity, Tennenhouse said. The digital products within the home are "dying to be networked," he said, with wireless home networking "ready to pop" in terms of market acceptance.
Other efforts are slightly further out, such as the sensors and actuators that Intel plans to fabricate as part of its microelectromechanical systems (MEMS) effort. Biochips are being studied as a way to "sense things that are wet."
Many of the projects are being conducted jointly with researchers at a variety of blue-chip research centers. David Culler, a University of California at Berkeley researcher who also heads up Intel's Berkeley research lab, mounted an ambitious mini-event during Tennenhouse's presentation.
About 800 tiny networking modules, each about the size of a quarter, were hidden underneath the chairs at the San Jose Civic Center. Culler asked attendees to find the modules, turn them on and hold them aloft so that they could "self assemble" into an ad hoc network. A majority of the modules — which have a tiny processor, memory, sensor and wireless transmitter/receiver — "woke up" and formed links, causing red LEDs on the thumbnail-sized modules to light up.
Culler asked the somewhat surprised IDF participants to keep the modules with the expectation that some of would recognize each other and light up through Thursday as they are carried about the main IDF venue.
"This is the biggest ad hoc network ever to be demonstrated," Culler said, who noted the challenge of assembling a network in "a noisy, uncontrolled environment."
Motion monitors
Other Intel-funded projects target more clearly-defined problems. A network of widely-spaced sensing instruments is being built offshore by the University of Washington to detect changes in tectonic plates. Another group is putting sensors in San Francisco buildings to report damage from any future earthquakes.
Another project conceivably could help to identify missing persons by sensing motion within a network. In the first phase of that project, a research team dropped a net of "motes" from an airplane into a secluded area in northern California.
Tennenhouse said that one of the most important applications of such sensing networks could help search for life on other planets.
Intel is already working with Internet pioneer Vin Cerf and staffers at the Jet Propulsion Laboratory (Pasadena, Calif.) with a goal of putting a four-node Internet Protocol network on Mars, Tennenhouse said. The nodes would assemble bundles of information using TCP/IP protocols. The project must take into account the "very long delays of sending IP packets over the long distances of the universe," Tennenhouse said.
|
|
|||||||||||
