Present: Bill Doering, Glenn Davis, John Doner, Rick Johnson, Shea Lovan, Elise Meyer, Alan Moses, Joan Murdoch, Larry Murdock, Stan Nicholson, Joel Rising, Deborah Scott, Jamie Sonsini, Bob Sugar, John Vasi, Karen Whitney
Not Present: Debbie Anglin, Kevin Barron, Art Battson, Ken Bowers, George Gregg, Phil Handley, Sonia Johnston, Bill Koseluk, Tom Lawton, Pam Lombardo, Tom Marazita, Bill McTague, Ed Mehlschau, Kevin Schmidt, Vince Sefcik
The group reviewed and adjusted the format for submission of proposals for spending Network Augmentation & Maintenance funding. Here is the result (re-numbered and slightly re-ordered):
ITPG Proposal Format:
- Project name
- Project summary
- How this project supports the academic mission of UCSB
- Funding source and how this project relates to it
- Costs: initial-year and recurring (attach spreadsheet)
- Matching opportunities (if available)
- Staff support required
- Existing resources to be used
- Project timeline
- Life cycle of result
The group also considered the following steps for the process of submitting and evaluating proposals:
- Preliminary discussions with ITPG regarding scope and coordination
- Submission of proposal
- Review by ITPG
- Mid-point review
- Final review (including cost spreadsheet)
Funds available for allocation in 1999/2000 include:
|Network Augmentation & Maintenance Funds for FY 99/00
|Network Augmentation & Maintenance Funds left over from FY 98/99
|Remaining from initial allocation of $120,000 by Communications Services
|Remaining from 98/99 allocation of $31,520 for FDDI Maintenance
|Total Funds Available
|CalREN-2 maintenance & research group connections
| FDDI maintenance
|Total Funds Committed
After discussion the group decided not to consider funds made available to Communications Services by Residential Services for use of the CalREN-2 network and the related ISP, and, similarly, not to consider the costs of upgrading the modem pool. As proposed by Vince, these items nearly offset each other and both are deemed to be in the domain of Communications Services.
In response to a question regarding the scope of ITPG recommendations in the face of the funds available, Bob suggested that we consider proposals for all projects, not just those possible with the funding we have now. Different projects might have different funding strategies, some incremental and some lump-sum. We should not ignore projects because ITPG doesn't currently have the funding.
Some feel, however, that it is not a good tactic to propose what is essentially one piece of incremental implementation of NGBB without campus commitment to the entire project. Nevertheless, we need a proposal for the NGBB with options for phasing. One strategy might then be to use the money available now for the initial phase.
Elise framed the strategy issue by noting that the NGBB could be broken down into three "horizontal layers" that must be provided to all participating locations: 1) single-mode fiber, 2) electronics and, 3) staff support. One possible strategy is to use the available funding to pull single-mode to as many locations as possible. Alternatively, we could choose to provide fiber, electronics and support for some initial locations and grow the network in "vertical columns" rather than horizontal layers.
It was suggested that decision makers will first want an idea of the total commitment and will then want to know where the pressing needs for the NGBB are in order to decide on phasing.
In the discussion that followed, BEG members expressed a need for guidance on the goals of the NGBB. For example, if the goal is infrastructure to support all buildings, the proposal might be to pull single-mode fiber. If, alternatively, the goal is to ease the migration from 10 Mb/sec to 100 Mb/sec for units who need to do that, then existing equipment must be replaced. If, however, the goal is expanding CalREN-2 to those who wish to join, additional staff support might be needed. While BEG is trying to stay focused on technical aspects of the NGBB, an expression of the priority of goals would help. Trying to develop technical plans for all possible scenarios would be difficult.
It was suggested that BEG narrow the range of strategies to two or three such as providing gigabit service to all buildings or providing "just in time" bandwidth to those who need it most. It was also noted that Physics, Geology, Geography and GSE have requested CalREN-2 connections. Further, it was suggested the NOC might have additional information on which units have been requesting greater bandwidth.
It was then proposed that we use some of the $155K to connect those units that are ready to the CalREN-2 network.
A difficulty, however, with growing CalREN-2 is the need to support two different network structures (FDDI and NGBB) at the same time with the same staff. Security issues also consume the current staff for weeks at a time. BEG members are concerned that the people who now support CalREN-2 could not "scale up" to support a production network. The NGBB centralized routing or VLAN (Virtual Local Area Network) concept is new and will require a staff support structure that does not now exist.
Some questioned the proposition that support for VLAN would require significant additional staff. If, however, frequent manipulation of such things as Access Control Lists (ACLs) became popular, staff support might be required. Others inquired whether the centralization of routing might free up staff currently needed to support the existing decentralized model. It was noted that one big box might be less costly to support than lots of little boxes. It was also noted that the original CalREN-2 plan was to operate the network with the first nine research units moving data for three months to measure the degree of support required. Unfortunately since migration of research groups to CalREN-2 has occurred more slowly than anticipated, we have not yet been able to make these measurements.
It was expressed that more staff time needs to be allocated to the network and some staff needs to be dedicated to dealing with security issues. The following alternatives were suggested: 1) add one person split 50/50 between security and CalREN-2 support or, 2) hire one person to deal only with security issues thus freeing up existing NOC staff to re-focus on network issues. The possibility of converting some one-time funding to ongoing staff support was suggested. The Context Paper also shows the need for an additional person to work on security.
It was noted that other campuses have extended CalREN-2 to those who were ready. It is, however, not known whether the additional units at UCSB would be comfortable attaching to an "unsupported" network. Nevertheless, it is worrisome to limit access to an expensive infrastructure. If staffing is a limitation, we should have a proposal for additional staff to consider. Bob agreed to carry the message that we will have difficulty supporting any additional networking without some increment in staffing back to ITB.
It was suggested that information regarding "what" should be done and how it might be funded, should flow back and forth between BEG, ITPG and ITB while the new OIT would be the organization responsible for providing the central support and deciding "how" to do it. In this model ITPG would help define the phasing.
Based on the discussion outlined above, ITPG members felt that we had provided "pre-proposal guidance" to the BEG and that they could now write up some alternatives to bring back for discussion. The proposal would include an overall NGBB architecture along with a couple of first-step alternatives to consider.
The question of entertaining proposals for other items such as intrabuilding wiring and classroom wiring was raised.
We speculated that intrabuilding wiring might become a campus issue because it was currently before the ITB. Advantages of moving responsibility for intrabuilding wiring to a campus unit include the flexibility to reconfigure offices in the face of a space shortage and the ease of providing funding to a campus plan as opposed to proceeding piece-meal. In the campus model, "standard" wiring to the wallplate (without electronics) would be provided and units with need for superior service would be allowed to upgrade using their own funding. Although a vote was not taken, ITPG members seemed to support this proposed change in policy.
It was decided that proposals regarding the security item and the authentication item should come from the subcommittees formed around these issues. Alan volunteered to submit proposals for CBT and EDUCAUSE and Glenn volunteered to do the same for Gartner Group. We decided to leave the software licensing issue for the OIT (because the issue of mass licensing for Microsoft products is not yet on the table). Elise volunteered to champion web-based instruction (along with Larry and Art?).
The group agreed to extend the deadline for proposals to October 31.
Returning to old business, it was noted that the review of the proposed Electronic Communications Policy is, in fact, the "real campus review." Previous thinking was that we would get another chance later, but subsequent information from UCOP made it clear that the final review was only a formality and that we should convey any substantive comments now. Comments may be transmitted to Meta Clow until November 1. Her address is: Meta.Clow@vcadmin.ucsb.edu .
The meeting wound down with re
ports on GTE SMPT filtering (Jamie sent a subsequent note with details) and experiences with cox@home. After the meeting ended, Elise and Jamie conducted experiments with competing wireless connectivity options for their Palm Pilots. (Note from Elise: Jamie’s connection was slower, but it actually worked. Elise’s cell phone kept disconnecting, but in subsequent tests from home it was faster.)
The next meeting was planned for the October 21-22 period. We will also try to find a regular time for the ITPG to meet so that people can plan in advance.
Back to ITPG Meeting Schedule