California Takes the CENIC Route

The network will serve as the California branch of Internet 2, the high-end network that will link research universities across the country. Internet 2 will combine the forces of a number of regional networks, each of which will be known as a "gigapop," or gigabit capacity point of presence. These points of presence, routing centers maintained by regional Internet 2 members, will perform at speeds hundreds of times faster than those generally available today.

Information Technology's Russ Hobby, chair of the CENIC technical planning group and one of ten engineers designing Internet 2, helped write the proposal for funding submitted to the National Science Foundation on January 1st. The majority of funds will come from the university participants: UC and CSU have each pledged $100,000 to create the CENIC organization, and Stanford, Cal Tech and USC have each promised $30,000.

"The idea is to create a California network with Internet 2 attributes," says Hobby. The best-effort level of service provided by today's Internet isn't sufficient for research purposes, particularly with the recent surge in Internet traffic. Lack of reliability may inhibit researchers and instructors from taking advantage of available technologies. Engineers building Internet 2 plan to provide several levels of service, allowing users to reserve bandwidth for specific projects at certain times.

Imagine an astronomer on campus who would like to view the sky via telescope at the same time each week. Rather than make a weekly road trip to Lick Observatory, the astronomer could link remotely to a telescope at Lick from a computer on campus. Today's network might not provide the reliable service that would make this weekly viewing a certainty. But with Internet 2 in place, the astronomer would be able to reserve bandwidth in advance, and the viewing time would be guaranteed.

"Internet 2 is application-driven. We aim to do things we haven't been able to do before because of low bandwidth or less than adequate reliability," Hobby says. Creating new services based on the current and anticipated needs of academics mirrors the evolution of the original Internet. The development of Internet 2, like that of its predecessor, provides a forum for experimentation, exploration, and exchange of information. Protocols in development include Ipv6, a packet-delivery system that allows a user to assign priority to data transmitted via the Internet, and multicast, which duplicates data being sent to multiple recipients at a branch point in the network, making more efficient use of both network capability and desktop memory.

As new protocols become tenable, their use will expand into the commercial sector.

"It's not cost effective for commercial vendors to build equipment exclusively for universities. If the technology has a use in the private sector, vendors are more interested," Hobby says. Partnerships with the private sector provide academic engineers with the support they need, and later allow companies to market new technologies to a wider public.

Production on a grand scale also makes equipment more affordable to universities. A statewide academic network was discussed in the late 1980s, Hobby says, but universities opted instead for autonomy. Now that networking expenses are rising dramatically, collaboration has become more attractive. The partnership will bring costs down across the board by creating leverage with vendors and sharing the costs of building the Internet 2 infrastructure.