Next Generation Internet One Year Later

By Louise Arnheim

In the year since the computing research community gathered for CRA Next Generation Internet (NGI) workshop, a substantial amount of behind-the-scenes policymaking has been taking place. As a result, there are signs the NGI initiative is gaining needed congressional support and capturing the attention of a wider audience. For example:

• NGI received a widely televised mention by President Clinton in his 1998 State of the Union Address;
  
• Both the House and Senate Science Committees have introduced NGI authorization bills;
  
• NGI applications were recently showcased for lawmakers at a Washington, DC, exhibit; and
  
• Internet2, the private, university-led initiative, formalized its activities by establishing UCAID; the University Corporation for Advanced Internet Development.

Much of this more visible activity is attributable to year-long efforts by all parties-the Administration, lead agencies, Congress, academia, and the private sector-to improve the overall NGI policy dialogue.

Congress and NGI Funding

The most significant development is the introduction of the two NGI bills: S. 1609, and H.R. 3332. As this article goes to press, these two authorization bills are just emerging out of Committee stage. Prior to the Easter recess, S. 1609 was reported out of the Committee on Commerce, Science, and Transportation without amendment and sent to the Senate floor. The House version was scheduled for markup by the full Science Committee after the Easter break.

Both measures amend the High Performance Computing Communications Act of 1991 to authorize NGI funds at the following levels:

Senate:

FY '99: $102 million
FY '00: $115 million

FY '99: $110 million
FY '00: $115 million

(The House bill actually meets the President's budget request for FY '99).

Rep. James Sensenbrenner (R-Wisconsin), who chairs the House Science Committee, recently called NGI "the logical next step for Internet development." Given the Chairman's cool reception to NGI last year, his co-sponsorship marks a turning point in congressional acceptance of the initiative.

As previously reported by Computing Research News, March 1998, Sensenbrenner had pressed the Administration for additional information on NGI. At this time last year, the House Science Committee had decided against including NGI language in its NSF authorization bill. Months later, during hearings on the NGI, Sensenbrenner seemed less than satisfied with the level of detail provided by Administration witnesses.

Ultimately, NGI moved through the congressional appropriations process with key departments and agencies such as DARPA, NIST, NASA, and NSF receiving a collective total of $85 million for FY '98 only (in contrast to President Clinton's original request for $300 million over three years.)

Apparently, Rep. Sensenbrenner was not alone in his quest for more details. Sources tell CRN that the Administration's failure to provide sufficient programmatic and planning information was a major reason behind congressional reticence to further fund NGI. It now appears, though, that revised versions of both the Administration's NGI concept paper as well as its implementation plan have satisfied lawmakers' requests.

Another concern of lawmakers was the geographic distribution of federal funds; that is, would NGI resources benefit institutions in rural and low-population-density areas? Here, the oversight Committees were apparently satisfied enough to give the go-ahead for funding, but still wary enough to charge the presidential advisory committee on NGI (the Presidential Advisory Committee on High Performance Computing and Communications, Information Technology and the Next Generation Internet) with keeping watch. Specifically, each bill requires the advisory committee to see whether Internet users in such areas incur a "geographic penalty"; that is, costs that are "disproportionately higher" than those paid by users in urban areas.

When asked what had transpired between last fall and the present to change lawmakers' views, NSF's George Strawn, Division Director, Division of Advanced Networking Infrastructure and Research (who also co-chairs the Large Scale Networking Group) pointed to the revised implementation plan. Another matter which helped, said Strawn, was "the cooperation of NSF's EPSCoR (Experimental Program to Stimulate Competitive Research) with the NSF NGI effort." According to Strawn, this cooperation "helped spur broad geographic participation by the nation's universities."

UCAID's President, Doug Van Houweling, attributed the change to the nature of the policy process. "There's been a lot of communication since last fall," he said. "Part of this is just a matter of people talking to people and getting comfortable." Van Houweling also noted several outreach activities that have taken place in recent months. One such effort was a video jointly produced by UCAID and the National Coordination Office for Computing, Information and Communications. Another effort was a daylong exhibition of federal agency NGI activities called "Netamorphosis." Hosted by the Office of Science and Technology Policy and the National Economic Council, Netamorphosis was expressly designed to familiarize lawmakers with applications made possible by NGI.

Presidential Advisory Committee on NGI

While Congress is now giving the NGI initiative more serious consideration, it is also eyeing the NGI presidential advisory committee with greater scrutiny. Both the House and Senate bills require the advisory committee to make a report to Congress (and the President) six months after enactment, then yearly until September 30, 2000.

NGI Agency Activities

While some agencies, such as DARPA, NASA, and NIH are humming with NGI activity, at least two federal agencies may be stuck in second gear. The Department of Energy, for example, has yet to receive NGI funding, and NSF is currently barred from spending the funds allotted by Congress from domain name registration.

DARPA - Last fall, the agency issued a BAA (broad agency announcement) for work in three inter-related areas:

1. Network Engineering - DARPA's effort to achieve growth in networks by a factor of 100 (or more).
2. Quorum - The agency's new approach to network-based computing, which features "adaptive quality of service management" as its basic organizing principle.
3. Supernet - A program designed to 1) develop "ultra high-speed multiplexing and transmission technologies with advanced configuration management and control capabilities"; and 2) demonstrate "end-to-end network connectivity involving tens of sites (nodes) and applications."

DOE - In contrast, researchers at the Department of Energy can only contemplate NGI-related work. The Appropriations Subcommittee for DOE not only refused to provide funding for NGI activities, it expressly prohibited the Department from using any current funds to conduct NGI projects.

NSF - In funding NSF-NGI activities for FY '98, Congress directed the Foundation to use $23 million from the Intellectual Infrastructure Fund (the interest-bearing account derived from domain name registration fees that is run by NSF's subcontractor, Network Solutions Inc). The fund, however, is currently untouchable, pending outcome of a class action suit (William Thomas et al v. NSI and NSF) filed last year. In February, the judge overseeing that case issued a preliminary injunction, thereby barring use of the funds in any fashion.

How will this action impact NSF's participation in the NGI? NSF's George Strawn is "hopeful" that both the President's FY '99 request will be approved and that the FY '98 funds will become available. "Our FY '98 challenge," he says "is to be prepared for both possibilities of NGI funding, or no NGI funding. And I believe that we are prepared."

Other Agency Efforts - The NGI applications featured at Netamorphosis were representative of ongoing R&D at several federal agencies. These applications include NASA's work on a telemedicine project in echocardiography; NIST's octahedral hexapod (a remote, software-controlled metal-cutting machine); and NLM's drug design collaboratory, a project that enables researchers at different sites to analyze and come up with new drugs.

Internet2 and UCAID

Closely related, and in some cases paralleling NGI, is Internet2, the university-driven effort to ensure continuing Internet access in the now-commercial and often congested Internet environment.

Until recently, Internet2 was an informal collaboration of mostly university researchers meeting casually to discuss mutual concerns. To provide the group with an ongoing legal framework for future activities, UCAID was incorporated last fall. Among UCAID's 120-primarily university-members are organizations such as NYSERNet and corporations such as 3Com.

Though Internet2's interests are similar to NGI, they involve separate sets of activities. For example, while NGI and Internet2 both focus NGI's promoting advanced applications, explains UCAID President Van Houweling, Internet2 is more concerned with instructional applications, whereas NGI is more concerned with research applications.

Of particular interest to CRA, Van Houweling says, is UCAID's Network Research Liaison Council (CRA's Board Chair, Ed Lazowska, serves on the Council). According to Van Houweling, the Council will ensure collaboration with the computer science community on advanced networking issues.

Van Houweling also places special emphasis on his organization's efforts to promote technology transfer. "We are working closely with industry so that the technology that is developed and proven will in fact find its way into the commercial Internet," he says.

The CRA report produced from the NGI workshop, Research Challenges for the Next Generation Internet, is still available from the CRA office. Contact info@cra.org or call 202-234-2111 for copies.