Bill Blass' Comments
Comments regarding the NGI initiative to the Presidential Advisory Committee

 

Bill Blass, Professor of Physics,
University of Tennessee, Knoxville

 

Thank you for allowing me to address the group this afternoon. I am Bill Blass, Professor of Physics at the University of Tennessee, Knoxville. I also serve as Senior Consultant to the Director of Telecommunications and Network Services, Division of Information Infrastructure. In the past I also served as the Associate Director for Research of the University Computing Center. 

 

The first issue involves AUP's (Acceptable Use Policies). We are all aware that AUP's are necessary for orderly utilization of networks and I have no argument with AUP's . I do encourage the Advisory Council to consider recommending some way to encourage and enable University - Federal Agency cooperation and experimentation by, perhaps, enabling easing of AUP restrictions on a case by case basis. In order to facilitate and encourage research and experimentation between Universities and Federal agencies, we must change the mindset which simply discounts the possibility of shared network connections because of AUP restrictions. In the context of NGI, it should not take heroic efforts to achieve rather simple cooperative activities in pursuit of creative synergy.

 

A second issue is also related to AUP's but less directly. Interoperability between I2, ESnet, NREN, etc. must be an early goal of NGI. This interoperability is needed as early as possible in order to facilitate the NGI experimental research in R&E networking. Most of my peers need high bandwidth connectivity to ESnet and NREN primary sites and if they are to participate in the networking research by providing the application drivers, then the time for interoperability is early rather than late. Once again, narrowly drawn AUP's or insular network attitudes will preclude successful synergistic University- Federal agency activities in support of NGI. Once again, narrowly drawn AUP's can exacerbate the situation.

 

Returning to the first issue, I'd like to relate what we are doing at Tennessee in collaboration with ORNL. The University of Tennessee at Knoxville and Oak Ridge National Laboratory have a long history of cooperation. At present UTK and ORNL are OC3 connected and with the expected implementation of an appropriate BGP at the ORNL end of the link, all of the UTK - ESnet primary site traffic will flow over the OC3 pipe. In this same vein, we are discussing with SEPSCOR the possibility of cross connecting UTK with SEPSCOR at Knoxville. A primary benefit of this cross linking would be making ESnet access available to SEPSCOR sites over the UTK-ORNL OCS3 link. Further synergistic sharing of resource such as the sharing of an OC3 or even OC12 Knoxville - Atlanta link carrying ESnet, vBNS, and I2 traffic is complicated on one hand by the broadly drawn ESnet AUP. We know that there are ways to breech the obstacles but it is my position that NGI supporting research activity should not require heroic efforts on the part of the players.

 

Ladies and Gentlemen, thank you for your interest and attention.