Next Generation Internet Will Rely On Oregon Research

The Defense Advanced Research Projects Agency, developers of the first Internet, have awarded UO computer scientists $1.2 million to conduct basic research. The UO researchers will provide a key element in the $300 million government-sponsored research effort. Negotiations continue around additional funding at a similar level for developing applications of the UO group’s work.

“These grants put us at the forefront of developing the Next Generation Internet,” says Zary Segall, head of the UO Department of Computer and Information Science. “The government is putting a significant amount of resources into this effort. Much of the money will be spent on building new hardware. Of the money devoted to basic research, we received a significant proportion.”

President Bill Clinton and Vice President Al Gore on Oct. 10, 1996, in Knoxville, Tenn., announced their commitment to improving and expanding the Internet. The Next Generation Internet (NGI) initiative carries out that commitment. The initiative is a five-year collaborative effort of numerous government agencies, universities and the private sector. The Clinton administration has made an initial three-year funding commitment of $100 million per year.

Unlike the current Internet, which was created as a loosely connected network of networks, the NGI is being conceived of and designed as one huge, integrated system, explains Segall.

“The NGI does not represent an incremental improvement over the current Internet, but a whole new approach,” he says.

Such an advanced system will be capable of performing much more complex tasks than today’s Internet. For example, it could facilitate surgery where a surgeon is in one city or country and the patient is in another. Or, similarly, engineering experts at the headquarters of a high-tech company could use real-time audio and video links to monitor and supervise a difficult on-site repair of the company’s product that is installed thousands of miles away.

In these kinds of situations, users depending on the Internet cannot afford to receive the ‘Net equivalent of a busy signal or “out of order” message. The NGI would solve that problem.

As society raises its level of trust in and reliance on the Internet to such high levels, Segall notes, corresponding increases in the ‘Net’s level of performance, security and availability will be required.

UO scientists will create improvements in these areas, known collectively as Quality of Service (QoS), by developing a whole new kind of “smart” software.

Members of the Oregon group—computer science Professors Segall and Steve Fickas, Assistant Professor Amr Sabry and Associate Professor Stuart Faulk and several graduate students—have experience in creating software for important and complex computer systems, such as air traffic control systems, NASA space station programs and AT&T telephone switching systems. But all these systems are, in a sense, static in comparison with NGI, which is envisioned as an evolving system.

“In the world of the NGI, software needs to be ‘self adaptive’,” says Steve Fickas. “A software application, such as telesurgery, can encounter a wildly varying runtime environment, changing not only over days, but over seconds. Our task is to make software that can respond to these changes and adapt itself on the fly. Again, after the software controlling a remote scalpel has begun its work, a ‘runtime error 3, aborting’ simply will not do.”

Segall points out that practical use of the new NGI-related research will begin in two or three years during the application phase of the initiative.
------------