Editor's Note: Minutes received 8/8 CURRENT_MEETING_REPORT_ Reported by Bob Stewart/Xyplex Minutes of the Uninterruptible Power Supply BOF (UPSMIB) Agenda o To identify the scope of the problem of monitoring and controlling uninterruptible power supplies; o To discuss an Internet Draft containing an initial proposal for such a MIB, and o To assess the interest and commitment towards ongoing work, including the possibility of creating a Working Group to prepare and advance proposals for standardization in this area. If there is sufficient interest and commitment, the Working Group Charter and timetable will be discussed. For this meeting, Jeff Case presided and Bob Stewart recorded. The meeting was well-attended, about 30 people, with representatives of about 10 UPS vendors, many becoming involved in the Internet and the IETF for the first time. After considerable discussion and review of a proposal, the meeting decided to request startup of a Working Group, with most of the work being done via a mailing list. Goal Efforts to make uninterruptible power supplies to be monitorable and controllable via the Internet Standard Management Framework have already begun. In the past, when MIB standardization has trailed product development, as it did for terminal servers, intelligent repeaters, and MAC bridges, users have been faced with the difficulties associated with the unnecessary proliferation of similar, but different, enterprise-specific (vendor) MIBs. As a result, it is desirable to begin standardization efforts as soon as possible. A draft document has been prepared as an introduction to the problem. It states: This memo defines an experimental portion of the Management Information Base (MIB) for use with network management protocols in TCP/IP-based internets. In particular, it defines objects for managing uninterruptible power supply (UPS) systems. 1 The document which is in the Internet-Drafts Directory, as previously announced, is a new version of the memo which contains incorporates the suggestions received by the authors since the initial document was published, plus one new group, the upsTruck group. Jeff presented the organizations and procedures that oversee Internet Standard Development. o Mailing list discussion a bit of a problem due to lack of mail access by UPS vendors. o Stressed structure above IAB, current and changing. o All-volunteer lower structure. Jeff stated goals. o Decide whether to proceed. o Identify sufficiently interesting set of common attributes with no optional objects. The answer to a concern over response to unimplemented objects was that groups are the unit of conformance for interoperability. Jeff presented highlights of the strawman proposal, which caused several points of discussion. o Volts and Minutes are too coarse, tenths of volts and seconds better. o Concern over debating individual objects deferred for later detailed assessment of proposal. o Traps are to bin in a separate document to ease passage of MIB. o A UPS need not perform multiple tests simultaneously but may if it can. o It is implementation specific whether configuration options cause changes. o SNMP proxy mechanism preferred for handling multiple UPSes, rather than table with index. o A community string or party defines an agent. o Fielded systems are basic and advanced, MIB represents advanced, suggest organizing MIB accordingly. Agreed too much mandatory for less expensive devices. Agree with option by Group for predictable 2 functions. This is a marketing issue. This discussion should be deferred for formal Working Group. We all want one standard. We discussed whether we do indeed want to form a Working Group. o Charter 1 or 2 documents to monitor and possibly control UPS, low or high end, existing and future UPS technology. o Consider application to similar embedded systems such as power systems or power conditioners. o Prefer not to encumber UPS needs. o There was considerable concern about voting and influence. Process is by consensus without company dominance or simple votes. The major need is to define the problem and rally around a Standard. SNMP itself was such a compromise. o We need an editor and email communication. MCI, Sprint, Compuserve, etc., provide mail service, and some do not charge for reading mail. o Arbitration is informal, by consensus and compromise. o Credit is author's on front and sometimes individuals in acknowledgements section, with name and affiliation. o Mail is easier than news group. o Publicity is acceptable as long as not claiming standard before complete. Internet Drafts should not be referenced in procurement or product literature, but RFCs may. o Mailing list administrative address is upsmib-request@cs.utk.edu. o We plan to have a document by the next IETF, final by following. The next IETF conflicts with Comdex, a big problem. Suggested Las Vegas meeting that vendors attend is a problem for Chairs. One of the vendors (APC) presented an alternative proposal. o MIB being implemented but needs to consider strawman proposal. o Proprietary features were removed for presentation. Remainder was divided into basic and advanced to maximize compatibility with past and future systems. 3 o Extension objects point to further MIB, assuming it is similarly structure. A single object is preferred and sufficient. Several general issues were discussed before adjournment at 10:20. o Someone suggested a breaker group. That varies considerable across implementations. It could be handled by alarm group in strawman. o We looked at several objects in strawman and general consensus was they are implementable. o Although one of the proposers, Emerson, does not implement everything in the strawman, it was proposed for value to the industry. o On the issue of credit for draft and RFC the suggestion was to limit it to SNMP Research. Suggested that Emerson receive consideration as catalyst. Deferred to private discussions and mailing list. o Suggested September meeting central in U.S. deferred to mailing list. Attendees Richard Baxter Tom Brennan brennan@exide.com Jeffrey Case case@cs.utk.edu James Davin davin@thumper.bellcore.com Michael Davison davison@cs.utk.edu Roger Draper rdraper@cerf.net Bill Elliot David Engel david@ods.com David Fencl Owen Gallagher oweng@jjmhome.uucp Theodore Greene Kenneth Key key@cs.utk.edu Sharon Lewis lewis@cs.utk.edu Les Matheson matheson@cerf.net Paul Moran Paul_Moran@3com.com David Perkins dperkins@synoptics.com Marshall Rose mrose@dbc.mtview.ca.us Koichiro Seto seto@hitachi-cable.co.jp Houman Shafiezadeh houman@exide.com Timon Sloane peernet!timon@uunet.uu.net Einar Stefferud stef@nma.com Bob Stewart rlstewart@eng.xyplex.com Adam Stolinski Ray Wasson 4 Brian Young 5