To: frost@geology, bpozos@sciences, wright@typhoon, rswiniar@sdsu, doug@sunstroke, sbarlow@sunstroke, jkatz@mail, dsharpe@mail Subject: Update from Friday (1/16/98) CENIC meeting at UCLA and requests Cc: short@sciences, pfrick@mail, paolini@kahuna, whitney@cs, beck@cs, elwin@math, varnell@sunstroke, morris@sunstroke, tlau@mail, zaslavsk@rohan, rledgerw@mail, kent@sdsu.edu, twest@calstate.edu, garye@csu.net, dave@csu.net Date: Fri, 23 Jan 98 13:19:49 -0800 From: "STEWART@SATURN.SDSU.EDU" Hello participants in the Aug. 1998, CalREN2 proposal (and friends in cc: list), I was able to attend the first meeting of the Applications Advisory Board meeting for CENIC [1/16/98] (you've received previous email on this) and I wanted to share this DRAFT summary from the UCLA sponsor. The negotiations between NSF and the CSU Chancellor's office are still ongoing. I hope you have contacted Kent McKelvey to make sure your campus network connection is adequate for your proposed research activities. I enjoyed meeting and spending the afternoon with Elhami Ibrahim of Cal Poly Pomona. No one from CSU San Bernardino attended. The notes below don't reflect the ending statement that the next meeting with likely be in May. I strongly feel that the SDSU part of this proposal needs to have applications that I can point to that exist now and would be better with more bandwidth. Can you provide me with either online-information (URLs are excellent) or hard copy of your proposed research? I would anticipate that when "push comes to shove", we will have a stronger argument if we can show actual activities, what do you think? I'll also post this to www.edcenter.sdsu.edu/projects/vnbs_summary.html kris stewart@sdsu.edu 619.594.0491 SDSU/LLA #73 www.edcenter.sdsu.edu NPACI/CSU Education Center on Computational Science & Engineering ------- Forwarded Message Date: Thu, 22 Jan 1998 11:22:12 -0800 To: "Doug Cummings, OAC" From: "Doug Cummings, OAC" Subject: DRAFT notes from 1/16/98 CENIC AAC meeting Thank you for attending the first meeting of the CENIC Academic Applications Council last Friday. We hope that the meeting was a useful exchange of information and will provide the basis for us to begin launching application partnerships in the near future. Below are my notes from the meeting and a list of the attendees. Please let me know of any changes, additions, contributions, etc. you would like to make by Tuesday, January 27th so that I can publish a final version to the full AAC list on Wednesday. Doug DRAFT Notes from CENIC Academic Applications Council Meeting UCLA -- January 16th, 1998 Morning sessions Gayle Byock opened the meeting with a review of CENIC's structure, goals, and objectives. Success of the enterprise will require close working partnerships between the campuses. As a first example of an NGI application, Bill Jepson provided an overview of the World Server application, including a video and a description of the system environment and operating characteristics. Dan Valentino followed with an overview of current and planned network based radiological treatment and diagnostic applications using slides and video to give insight to the scope and scale of the projects. Both application areas involve high bandwidth and quality of service requirements. Both applications emphasize the requirement for support of partners at distant locations. Mario Gerla lead a group discussion of NGI architecture and infrastructure in relation to application requirements. This discussion raised the following requirements and issues: 1. Bandwidth -- is important for aggregation. Two issues need to be addressed: average bandwidth versus burst levels above the guaranteed level. 2. Latency -- in some applications, additional data which allows for prediction can moderate latency requirements. 3. Delay jitter -- is an important consideration for realtime audio and and video. Applications vary widely in the level they can tolerate. 4. Costs (prices, rates) for customers must be understood. 5. Quality versus cost tradeoffs are important and vary with applications. 6. Connection quality (e.g., packet loss) requirements vary with applications. 7. Heterogeneous network adaptivity (e.g., wireless) relates to quality of service issues. 8. Multicasting (video conferencing, distance learning) unicast integration. 9. Security 10. Connectivity between national networks and quality of peering between networks is necessary to support supercomputing applications. Interaction between small peer groups (e.g., partnerships) versus broadbased usage (e.g., web based servers). 11. Virtual networks for computer science network research. 12. Connection and connectionless services have differing requirements. 13. Application scheduling maybe required to live within the available bandwidth. Scheduling may be a centralized function. 14. Opportunities may exist for a centralized organization to provide "middleware" services (e.g., caching, reconciliation, replication, network storage, etc.). Some of the PACI and NPACI participants are or will be involved in these activities. (Sid Karin is on the CENIC board and provides some coordination between CENIC and PACI. UCLA, USC, and Cal Tech are participants in NPACI.) 15. There is a need for a customer accessible network monitoring service service which can be queried about the health and current operating characteristics of the network. This information has been described as a "network weather report". Some of this information is available in the routers. The issues are largely policy and procedure related. There will be a CENIC network operations center. Questions include what data will be available and when and how the data will be provided. Those concerned with these issues need to communicate their needs to the appropriate councils. An SDSC based group, the Cooperative Association for Internet Analysis, is involved in this area. Afternoon sessions Russ Hobby provided an overview of the activities of the CENIC technical council and the currently planned structure for the network. There will be four OC-12 rings in each of the northern and southern California segments. Each group of four will consist of a pair of ATM rings and a pair of SONET rings. The current implementation schedule calls for the base network to be available in March, 1998; the statewide links later in 1998; multicast in June, 1998; policy routing in September, 1998 (this depends on some modifications requested from equipment vendors); and quality of service available in 1999. Russ will make his presentation available on the web. The gigapop operators are discussing banding together to form a national network in parallel to the vBNS network by creating direct connections between the gigapop sites. The description of the proposed network implementation appears to match quite closely many of the application requirements identified by the group. Applications could be divided into four categories: realtime, delay impacted and realtime, delay sensitive; realtime, non-delay impacted; large data transfer; and common best effort. While an order has been placed for the ATM equipment, the final vendor is still open to discussion. Ideas and comments on options should be directed to the Technical Council. The current pricing model calls for a flat rate charged to the campus with the campus being responsible for any allocation to local users. The availability of quality of service will eventually allow for direct charging to individual users. Each campus will be responsible for its gigapop. There will be a central network operations center which can provide advise, but local hands will be required to do the fixes. An RFP will be issued for establishment of the NOC. Council Organization and Structure Council participation is to be inclusive rather than exclusive. The designated campus representatives should include two application oriented and one computer science/network oriented faculty. Members of the Technical and Business Councils will be kept informed of the Applications Council activities and actions. In addition, any other interested parties are encouraged to participate in AAC activities and communications. Suggestions and comments from the attendees: 1. It would be useful to have a single campus contact point, in particular on campuses where computing and network responsibilities are dispersed among many organizations. There should be, at least, a well defined channel for communication concerning CENIC activities. 2. There needs to be a better picture of the capabilities available before campuses can effectively "inventory" potential CENIC applications. 3. Applications may see considerable development on a campus before they need CENIC level capabilities. 4. A "private" place (an "idea chat space") where people could float developing ideas for discussion could help foster application development. This could lead to affinity group formation. 5. The Applications Council could evolve into an association of smaller special interest groups. 6. Make descriptive information regarding CENIC and CalREN2 capabilities and structure more readily available. The CENIC web site (to be relocated to Cal Tech and updated soon) could host this information. (www.npaci.edu provides information on NPACI.) 7. Consideration should be given to a structure for allocation of resources to large research users if there is an expectation of attracting and keeping these users. 8. There needs to be a well defined interface to NPACI. 9. We need to have a suite of applications ready to run when the initial network configuration is launched in March. Stuart Lynn would like to have a joint meeting of the Applications, Technical, and Business councils in the Spring. The next Applications Council could be held jointly and then we could decide when to have the following meeting. Attendees at the Academic Applications Council Meeting January 16th, 1998 California Institute of Technology Harvey Newman Roy Williams CSU Pomona Elhami Ibrahim San Diego State University Kris Stewart Information Sciences Institute Clifford Neuman Stanford Dale A. Harris Jay Kohn UC Davis Russ Hobby UC Irvine Stephen Franklin UC Los Angeles Gayle Byock Mario Gerla Ben Goldstein Paul Hoffman Louis Hook Lu Huang Bill Jepson Luis Llambias Janette Miller Marsha Smith Daniel Valentino Mike Van Norman Doug Cummings UC Riverside Krassimir Bozhilov UC San Diego Jim Madden UC San Francisco Thomas Ferrin Peter Brantley Kyung Hwang UC Santa Barbara Kevin Barron University of Southern California Fred George Ulrich Neumann - ------------------------------------- Doug Cummings Grant Development Support Coordinator UCLA Office of Academic Computing (310) 825-7402 ------- End of Forwarded Message