Researchers advance to middle

NSF's Division of Advanced Networking Infrastructure and Research

The National Science Foundation's Division of Advanced Networking Infrastructure and Research is once again preparing to live up to its name.

ANIR has been the home of high-speed, high-performance networks for researchers for 15 years. It exceeded its Next Generation Internet goal of connecting 100 universities to high- performance networks by 87. One network worked so well that WorldCom Inc. has turned it into a commercial high-speed offering.

But ANIR has seen an almost complete turnover in staff since last fall, and the new team is viewing the changes as a cue to move in fresh directions.

"It's with a little bit of nervousness that we're coming into this office, which has been highly successful," said Thomas Greene, ANIR's senior program director for advanced networking infrastructure. "This office is constantly redefining itself because of the word "advanced' in our name."

Staff members are taking the next step in the evolution of networking by trying to solve the problem of middleware, with the goal of improving the user's network experiences for most applications, said Alan Blatecky, ANIR's new program director for advanced networking infrastructure.

Middleware is software that connects otherwise separate applications. In the networking arena, middleware refers to the software that is common to multiple applications and builds on network transport services to enable development of new applications and network services.

Proposals are due May 10 for the $10 million-per-year Network Centric Middleware Services program, the flagship in a series of three new ANIR programs. The solicitation asks members of the advanced networking community — which includes universities, government agencies and industry organizations — to collaborate in assembling a common base of knowledge about middleware.

"That's the big hole for the community," Blatecky said. If a user has trouble taking advantage of an application's capabilities, middleware is a major part of the problem. But because middleware is the glue between networks and applications, no one has really taken responsibility for solving those problems, he said.

NSF is looking for an integrator to design middleware services for high-performance networks and for ideas on how to make middleware available for common applications. The last piece of the puzzle is to figure out how to spur new research on middleware.

"This is a background set of problems that everyone in the world has," Greene said, "but we're attempting to ad-dress it first with high-performance networks."

The High Performance Network Connections for Science and Engineering Research project, which will connect more institutions to NSF's advanced networks, is accepting proposals for 30 grants, worth a total of $4 million in fiscal 2001, until May 18.

The other new program is Strate-gic Technologies for the Internet, a $10 million-per-year program that seeks research on unresolved issues about the operational and functional capabilities of the Internet. Proposals are due June 21.

"We can let our programs and funded research be a little less safe," Greene said. "A lot of in-cre-mental research is being done by the commercial sector. This allows us to have a place with a small amount of resources."