NSF very high speed backbone network service (vBNS) and NLANR

The vBNS represents a testbed for the emerging broadband Internet service infrastructure in which all parts of the network will be experimented with: switches, protocols, software, etc., as well as applications. It will be a unique resource for network and application researchers nationwide to explore performance issues with the new technologies (e.g., how host systems and interfaces interact with ATM components of the wide area network). We have submitted a proposal for a National Laboratory for Applied Network Research (NLANR) to NSF, as a vehicle for providing additional facility and research support, necessary to ensure the success of the vBNS.

The NLANR cooperative agreement will focus on research and engineering support for the vBNS, including issues related to facility maintenance and optimization, resource management, and research coordination.

The NLANR cooperative agreement will provide NSF with several important strategic functions:

1. local support at the vBNS sites to facilitate maintenance and operation of the vBNS facility. A 0.25 FTE technician will be responsible for technical support such as maintaining the link between the vBNS and the local site LAN, coordinating installation and maintenance of vBNS equipment, integrating equipment needed to support research efforts, maintaining routing tables, and monitoring network availability.

2. local support at the vBNS sites for research and engineering activities, in particular for vBNS workload characterization and performance evaluation based on detailed usage data. Collaborating with other parties, this is targeted to result in metrics and definitions of system service specifications. A 0.75 FTE researcher/engineer will support usage of the vBNS and facilitate NLANR research. NLANR participants will focus on specific components of vBNS usage:
   • performance and scalability of vBNS usage for high end data movement

   • usage and performance of vBNS for interactive collaboration technologies.

   • usage and performance of TCP/IP over high speed ATM infrastructure.

   • flow characterization of both individual vBNS flows as well as the aggregate workload.

   • work with the vBNS service providers and the NSF to determine what metrics can and should be tracked on a regular basis, and coordinate placement of the statistics into a database for making it broadly available in both standard and customized reports.

3. participation in the vBNS Resource Allocation Committee (RAC).

   The objective of these meetings will be to track extensions to the vBNS, coordinate tests of new equipment, discuss usage statistics, and coordinate access by new research projects. The NLANR, as a collaborator of the vBNS project, will assist in developing strategies for the implementation of new projects on the vBNS.

4. resource management for the vBNS. It is expected that NSF will grant use of the vBNS to research projects that occasionally will require significant or even dedicated use of the network. NLANR participants will work with the RAC on availability schedules for their local site, and will help establish schedules for when dedicated access to the entire network is feasible.

5. research coordination. A 0.5-FTE Research Coordinator (RC), Dr. K Claffy at SDSC, will coordinate research usage of the vBNS, and communicate and collaborate as need be with vBNS researchers. The RC will work with the RAC for the day-to-day scheduling of the vBNS as a resource. A schedule will be communicated to vBNS researchers to ensure that research project staff using the vBNS are aware of scheduled usage times for other projects. The NLANR research coordinator will serve as a central coordinator for resource management, performance studies, and vBNS activities, in collaboration with staff of the vBNS service provider, and the NSF.

6. information services to disseminate results to the community. This will facilitate information distribution both among the vBNS and NLANR collaborators as well as to the networking community.

7. overall project coordination and management by the Principal Investigators, and the research coordinator. They will serve as an interface to the NSF and the vBNS researchers.

8. an NLANR Executive Committee (EC) to work collaboratively with NSF/DNCRI on furthering goals and objectives for the national networking agenda, with specific focus on the requirements of the NSF and the vBNS. The EC will consist of the principal investigators of all the sites, the research coordinator, and NSF officials. The EC will provide a vehicle for addressing implementation issues and coordination with the NLANR sites. The Principal Investigators for the NLANR sites are Doug Carlson (CTC), Marla Meehl (NCAR), Randy Butler (NCSA), Matt Mathis (PSC), and Bilal Chinoy (SDSC).

These support functions will enable network researchers to focus on their particular research while a vBNS research support vehicle is in place.

We have proposed the NLANR project in support of the networking agenda of NSF DNCRI and related divisions for the next several years, particularly in concert with the DNCRI vBNS activity. Both NSF and the NLANR participants understand that this proposal in its current form only constructs a framework for supporting coordinated research on the national vBNS network research facility, but does not include the resources to carry out significant research itself.