Present: Arlene Allen, Debbie Anglin, Kevin Barron, Art Battson, Glenn Davis, Rick Johnson, Elise Meyer, George Gregg, Phil Handley, Bill Koseluk, Alan Moses, Joan Murdoch, Larry Murdock, Stan Nicholson, Kevin Schmidt, Deborah Scott, Jamie Sonsini, Bob Sugar
Not Present: Ken Bowers, Bill Doering, Tom Lawton, Tom Marazita, Ed Mehlschau, John Vasi
The purpose of the meeting was to clarify and discuss the twelve proposals that have been drafted for ITPG funding and/or recommendation. The discussion centered on the estimates for and the strategy surrounding the NGB proposal. Ordering and interrelationships of other proposals were also considered.
Missing pieces in the estimate for Phase I of NGB include the following:
- Cost of pulling and terminating single-mode fiber to interconnect the buildings housing the distribution switches (estimate in progress by Communications Services).
- Cost of connections between distribution switches and building switches (some of which might require additional single-mode and some of which might be supported with intermediate repeaters on multi-mode).
In discussing item #1 it was suggested that we should be thinking in the $100K ballpark. Item 2 requires additional specification before detailed estimates can be developed. It was guessed that $10-20K might suffice to acquire two-port switches to use as repeaters in those places where single-mode doesn’t exist and the distances are too long for straight multi-mode connections. The draft NGB proposal includes $20k for wiring from current building router locations to new building switch locations. The Phase I proposal also includes estimates for cabinets, power supplies and media converters to support the switches. BEG and Communications Services will continue to fill in the blanks until we have an estimate for Phase I.
A long discussion of possible strategies for funding the NGB ensued. The current estimate of $533K plus another $100K (or so) to complete Phase I exceeds the $155K that ITPG has left to allocate during FY 99/00. Thus, what to do with the $155K that is currently available becomes an issue. During the discussion, two general points of view were expressed.
One group felt that Phase I was the smallest chunk of NGB that should be considered and that funding for all of Phase I should be "committed" before we launch the project. This group felt that starting without sufficient funding might lead to the installation of only the distribution switches. In this scenario CalREN-2 and other researchers who are self-funded might have access to the new network, but many other units would not. Some of those who would not be connected have compelling needs to support instructional initiatives that are on hold because of inadequate bandwidth. Moreover, the NOC would be left supporting two disparate networks, one of which still contained unreliable equipment and convoluted software. If no additional funding were to become available, this group argues that the existing funds would be better spent replacing the troublesome equipment and/or laying additional single-mode fiber in preparation for the day when funding the entire NGB (or all of Phase I) becomes possible.
The second group contended that we should use the funds now available to start in the direction we want to go and that additional funding would very likely be allocated as needed to complete Phase I. This group felt that networking has become so important to research and instructional activities that additional funding would be provided if we made good use of what we already have in hand and the demand continues to escalate as expected. Thus, the second group argued that we should break Phase I down into "sub phases" such as installing fiber, acquiring distribution switches and integrating the classroom network.
Determining where the current bottlenecks exist would help in defining the sub phases (i.e., what to do first). It was noted that the FDDI network is not saturated. In response, it was observed that some units know their connections or their LANs will not support the applications they want to develop so they don’t "over-drive" their networks and we don’t see the demand in our usage statistics.
Some felt that the politics of getting Phase I approved would be easier if we could spread the funding request over multiple years and request the funding in yearly increments. Other felt that determining who would gain access first would be a difficult issue. It’s not clear whether research, instruction or replacement of unreliable equipment would have priority.
Continuing the discussion, some felt that "Phase I-A" (the first chunk) would have to contain all of the distribution switches and the single-mode fiber to connect them. The next phase, connecting building switches, might require additional single-mode. Alternatives to single-mode such as wireless (MNDS) or schemes using two-port switches as repeaters on the existing multi-mode were explored. Some felt that an additional $20K would suffice to acquire two-port switch-repeaters on the 8 to 10 links that might need them in Phase I.
Phase II will almost certainly require additional single-mode because the repeater solution probably won’t scale up to support the entire campus on multi-mode. Furthermore, the existing multi-mode is in a ring configuration and the switched architecture that has been proposed requires a star configuration. This raised the issue of whether we should be proposing to replace the entire fiber cable plant with single-mode in a star configuration as a first step. It was noted that two projects have already required pulling single-mode and it is not optimal to build a new cable plant one project at a time.
The discussion then moved to the practicality of making good use of $500-700K if it came all at once. Some observed that the bottleneck would be people to do the implementation (both our own staff members and vendors such as ODI). It was suggested that we install fiber, implement distribution switches and prepare the sites for the building switches in FY 99/00 and schedule the integration of the classroom network and the acquisition of building switches over the following year. This again raised the issues of who was to go first and how long others might have to wait. In response it was suggested that ITPG determine which delays might be calamitous as opposed to inconvenient.
At this point we attempted to enumerate the strategies being considered as follows:
- Propose to do all of Phase I in one year and wait until funding is committed to start.
- Propose to spend what we have in hand and ask for the difference as we go.
- Propose to do Phase I in increments over multiple years.
Bob then reported that discussions with senior officers regarding the OIT, intrabuilding wiring and the augmentation of IUC funds were ongoing. There is some indication that additional funds for IT will become available (note that the chronic shortage of IT funding is mentioned once again in the recent UC budget submission to the governor’s office). Thus, it is important for us to determine how much funding we need to complete Phase I and to give some overview of Phase II.
Moving to the other proposals, Art noted that classroom network augmentation is coordinated through the Classroom Renovation Committee and that funding in the current year had been provided. Thus, the proposal for funding classroom networking was for FY 00/01 and beyond.
A brief discussion of the Course Web Support proposal illuminated the difference between course management and content production. The relationship of the Physics Classweb proposal to the Course Web Support proposal was questioned. More discussion might be required.
The dependence of the PKI infrastructure on adequate directory service was noted. Also expressed was the desire for PKI to eventually support digital signatures, although that is not planned in the initial rollout.
In conclusion, Bill Koseluk noted that he was planning to attend a meeting at UCOP to explore our participation in the update of the UC study on workstation costs and needs. He will present his findings at the upcoming meeting and will undoubtedly require assistance in producing the desired statistics and projections.
Back to ITPG Meeting Schedule
