![[Currents header graphic]](/homeart/currents_header.gif)
October 20, 1997
|
| J. J. García-Luna
|
By Robert Irion
Aficionados of the Internet appreciate how quickly they can send and access information, virtually anywhere in the world. However, today's Internet also limits most of its users in one key way: They're tied down to hardwired offices and workstations, unable to get online on-the-go. Even wireless modems require users to hook up to the Internet via an established service provider, such as a university or America Online. If you're out of range, you're out of luck.
A research team at UCSC aims to change all of that. Led by associate professor of computer engineering J. J. García-Luna, the team has developed innovative tools known as WINGs--Wireless INternet Gateways--that allow users to create their own mobile networks. The portable devices, which communicate by radio, can spawn roving "Internets" at any time, anywhere, independent of wires, service providers, or other electronic middlemen. That flexibility makes WINGs ideal for emergency situations such as earthquakes or floods, for military applications, and for our increasingly nomadic working culture.
Now, the Defense Advanced Research Projects Agency (DARPA) has boosted García-Luna's efforts with two new three-year grants, totaling more than $2.5 million. One project will extend the team's work on WINGs by making the devices more secure, adaptive, and reliable under a wide variety of mobile conditions and possible attacks. In the second project, the team will work on a new Internet architecture using "active" packets, which can modify the behaviors of computers that receive them.
The bulk of the funding will support García-Luna's team of researchers and graduate students at UCSC's Jack Baskin School of Engineering. About $524,000 will go to García-Luna's collaborators at Rooftop Communications Corp. in Los Altos, the main industrial partner in the WINGs initiative.
At the heart of both projects is the goal of increasing the flexibility and adaptability of the Internet, which has spread in recent years like electronic kudzu--and with just about as much order. Existing networks usually can't handle the demands of a roving field operation, or even a business executive trying to run his office from his car, lunch engagement, or plane.
"We're trying to make networking more pervasive and transparent," García-Luna said. "It will be everywhere, but you won't know it's there."
WINGs have taken the team a big step in that direction. The handheld devices and the protocols that run them have attracted the attention of several industrial collaborators, most notably Hughes Electronics, SRI International, and Bolt Beranek Newman (a subsidiary of GTE). Each WING is capable of carrying out the functions of a normal Internet "node," such as routing and transferring e-mail and other information. It uses radio frequencies to talk to its brethren, creating a mobile community of communication. García-Luna's team envisions long-range, high-power WINGs that reside in vehicles or on rooftops, supplemented by handheld, low-power WINGs for users on the move.
The major new project is called SPARROW (Secure Protocols for Adaptive, Robust, Reliable, and Opportunistic WINGs). "WINGs give us one part of the bird," said García-Luna, explaining the enormous acronym. "We wanted the whole bird."
Security is a major aspect of the endeavor. A mobile network isn't much good, García-Luna said, if someone can compromise the information in some way. That's especially true for military operations, of paramount concern to the funding agency. So, the SPARROW versions of the portable devices will feature new protocols to guard against such intrusions.
"The Internet started out with the notion that all of its users were good citizens," García-Luna said. "But now it's everywhere, and more and more people have to put locks on their doors."
The team will create devices that can use any of multiple radio frequencies to communicate (as opposed to the single frequency used by current WINGs). The hardware also will adapt to different regimes of power levels, bandwidths, numbers of users, their travel speeds, and other conditions that can vary dramatically in emergency situations.
"What DARPA wants is a network that is really hard to kill," García-Luna said. "If someone jams one frequency, it could switch instantly to others. It could connect to land links, satellites, or wired networks. It must be very survivable and very mobile."
The civilian applications of such "Global Mobile Information Systems," as DARPA calls them, are clear. An instant communication infrastructure would help rescue teams in disaster areas, where it's vital to transmit information beyond just words over a two-way radio. Researchers, other workers, and the rest of us ultimately could access the Internet or each other through any number of mobile nodes, García-Luna said. "I could use something as small as my watch to stay connected to my work space," he noted with a smile.
The second DARPA-funded project is part of a broad effort to extend the Internet's base architecture beyond such traditional methods as IP (Internet Protocol) and ATM (Asynchronous Transfer Mode). In the Internet today, said García-Luna, changes in communications protocols can take years. Information packets sent from place to place are "passive," as in a document to open and read. But as technology evolves, network protocols will need to change within months or days to keep up.
To foster that environment, García-Luna's team will try to develop software for active packets--the equivalent, he said, of network-level "applets" used in Internet Java applications. As their name implies, active packets can play a role beyond simple information transfer. They also can contain instructions or programs that modify the behavior of the Internet node receiving the packet. The team will focus on how to do this securely, reliably, and under "multicasting" conditions when many users may communicate and exchange instructions at the same time.