CURRENT_MEETING_REPORT_ Reported by James Davin/MIT SNMP Minutes SIP MIB Tracy Cox of Bellcore presented the current MIB for the SMDS Interface Protocol. It was noted that the document largely exports those objects found in an SMDS switch that conforms to the Bellcore TA for SMDS. Numerous minor editorial comments were made and accepted. Once this editing is complete, the Working Group recommends this document to the IESG for consideration as a Proposed Standard. Ether-Like MIB Chuck Davin, the Area Director, reported that the IAB had concerns regarding the ether-like MIB which had been produced by the Working Group and recommended by the IESG for entry onto the standards track. These concerns were in the area of 14 mandatory objects which might not be available on a chipset that is minimally 802.3-compliant. The Working Group met in an ad hoc fashion the next day to edit the document to include the objects as optional. A liaison statement for the IAB, expressing the concern of the Working Group was unanimously adopted. (The minutes of this ad hoc group are below). At the open plenary, the liaison statement was read. Although the ensuing discussion lasted nearly 1-1/2 hours, some points were crystalized: the IAB Chair, Vint Cerf, noted that the ``IAB had consulted it's 802.3 expert to consider the matter''. Vint continued by saying that the IAB's expert, IAB member Tony Lauck of DEC, had been on vacation, and this delayed the disposition of the MIB. Jeffrey Case, a member of the SNMP Working Group, disagreed with Cerf, noting that the IAB's 802.3 expert was not on vacation--as ``the IAB's expert in this matter was not a member of the IAB nor any of his employees, but rather the IESG and the Working Group which produced the MIB.'' Another SNMP Working Group member, Karl Auerbach, observed that to a mean-spirited observer it might appear that one vendor, i.e., DEC, had purposely interfered with the MIB out of self-interest. Auerbach prefaced his remarks by indicating that he wasn't advocating that perspective. Nonetheless this raised several issues concerning the propriety of the IAB's actions and their reporting structure, which was subsequently discussed ad nauseum. List of MIBs Dave Perkins of SynOptics presented a comprehensive list of all known MIBs. This document will be revised on a regular basis and likely made available via ftp/mail. 1 Implementation Reports The Chair queried the group of their implementation of various management technologies on the standards track: Manager Agent Concise Definitions high high MIB-II high high Token Bus ~15 1/2 Token Ring ~15 4 Interface Extensions ~10 5 DS1 ~5 1-1/2 DS3 ~5 0 Because of the near-universal implementation of Concise/MIB-II, these documents are being recommended by the Working Group to the IESG for advancement. On the remaining documents, Manager vendors complained of a lack of agents to test with. This issue will be raised again after the Interop SNMP Solutions Showcase. Hopefully this will provide better implementation experience. Minutes of the Ad Hoc Meeting Chair: Frank Kastenholz Date: Thursday, August 1, 1991 An ad hoc meeting of the SNMP Working Group convened the morning of August 1, 1991 to revise the Ethernet MIB according to feedback from the IAB. The IAB, after reviewing the Ethernet MIB, expressed concern about several of the variables of the MIB and their mandatory status. Specifically, the IAB felt that implementation of the following variables not be required in order for an implementation to claim compliance with the MIB: dot3MacSublayerStatus dot3MulticastReceiveStatus dot3TxEnabled dot3StatsSQETestErrors dot3StatsDeferredTransmissions dot3StatsLateCollisions dot3StatsExcessiveCollisions dot3StatsInternalMacTransmitErrors dot3StatsCarrierSenseErrors dot3StatsExcessiveDeferrals dot3StatsFrameTooLongs dot3StatsInRangeLengthErrors dot3StatsOutOfRangeLengthFields 2 dot3StatsInternalMacReceiveErrors The Working Group discussed several different strategies for revising the MIB per the IAB's comments. These included: o Splitting the dot3Table and dot3StatsTable each into two tables, giving a total of four tables. Each of the original tables would be split into a table containing the variables acceptable to the IAB and a table containing the variables with which the IAB had concerns. This option was rejected when it was realized that the dot3Table would be split into two tables, one containing two variables and the other containing three. Such small tables were considered to be very inefficient by the Working Group. o Assigning optional STATUS to the variables that the IAB had an issue with. This option was rejected when it was explained that the unit of conformance for MIBs is the mib group. The Working Group settled on the following strategy: o The Ethernet MIB will be divided into two tables. Each table will comprise a separate MIB Group. One table will contain the objects which the IAB found acceptable and implementation of that group will be mandatory. The second table will contain all variables with which the IAB had an issue and implementation of that table will be optional. The text on the MIB document that allows returning 0 for variables for which there is no underlying hardware support will be removed. The Working Group believes that this MIB structure is less useful than the structure as submitted to the IAB because the requirements for conformance have been reduced. Vendors can implement only the mandatory group and be able to claim compliance with the MIB. However, the Working Group believes that this group does not contain all variables that are necessary for effective management of Ethernet-like interfaces. The Working Group also believes that several extremely useful variables, as determined by implementation experience, will not be widely implemented because those variables have been moved to the optional 3 portion of the MIB. Specifically, implementation and operational experience have shown the following five variables to be extremely valuable in detecting and diagnosing network problems: 1. dot3StatsDeferredTransmissions 2. dot3StatsLateCollisions 3. dot3StatsExcessiveCollisions 4. dot3StatsExcessiveDeferrals 5. dot3StatsFrameTooLongs The Working Group then developed the following statement for including in the the Minutes of the meeting. The Working Group also directed the Chair to read this statement in the open IESG meeting on August 1, 1991. LIAISON STATEMENT The IETF meetings, unlike others, are open. All parties, even IAB members, are encouraged to attend and make their positions known and have them argued in the open. It is important to appreciate that development of technology does not occur in a vacuum for a community as large as the internet. It is also important to appreciate that the needs of the community must be met in a timely fashion. IETF has worked over two years on the Ethernet MIB. The MIB was approved by the Working Group, the IETF plenary, and the IESG. Only at that late date, the IAB unilaterally raised issues that had been previously resolved. Because of the pressing needs of the community we feel coerced into revising the ethernet MIB as we have done today. Unfortunately the technical merit of this document has suffered owing to the lack of 802.3 expertise available to us on such short notice. As a result of this decline in technical quality, this document is less useful for management, and the manageability of the internet will suffer. It must be emphasized that during the two years of preparation, numerous 802.3 experts contributed to the process and all the issues later raised by the IAB were thoroughly discussed, repeatedly argued, and resolved to the satisfaction of the experts. We feel that the editing session, hastily convened owing to the misplaced paternalism of the IAB, has undone some of the work of those experts. 4 Attendees Steve Alexander stevea@i88.isc.com Karl Auerbach karl@eng.sun.com James Barnes barnes@xylogics.com Steve Bostock steveb@novell.com David Bridgham dab@asylum.sf.ca.us John Burruss jburruss@wellfleet.com Jeffrey Case case@cs.utk.edu Chi Chu chi@sparta.com Gigi Chu gigic@hpspd.spd.hp.com Henry Clark henryc@oar.net John Cook cook@chipcom.com Tracy Cox tacox@sabre.bellcore.com Dave Cullerot cullerot@ctron.com James Davin jrd@ptt.lcs.mit.edu Gary Ellis garye@hpspd.spd.hp.com Mike Erlinger mike@mti.com Jeff Fried jmf@relay.proteon.com Shari Galitzer shari@gateway.mitre.org Shawn Gallagher gallagher@quiver.enet.dec.com Kenneth Goodwin goodwin@psc.edu Phillip Hasse phasse@honchuca-emh8.army.mil Ron Jacoby rj@sgi.com Mike Janson mjanson@mot.com Frank Kastenholz kasten@europa.clearpoint.com Manu Kaycee kaycee@trlian.enet.dec.com Mark Kepke mak@.hpcndk.cnd.hp.com Kenneth Key key@cs.utk.edu Christopher Kolb kolb@psi.com Deidre Kostick dck2@sabre.bellcore.com Bobby Krupczak rak@cc.gatech.edu Cheryl Krupczak cheryl@cc.gatech.edu Chao-Yu Liang Shane MacPhillamy slm@netrix.com Ron Mackey rem@dsiinc.com Keith McCloghrie kzm@hls.com Evan McGinnis bem@3com.com Robert Meierhofer Lynn Monsanto monsanto@eng.sun.com Karen O'Donoghue kodonog@relay.nswc.navy.mil Michael Patton map@lcs.mit.edu David Perkins dperkins@synoptics.com John Pickens jrp@3com.com Brian Price brian@bss.com Marshall Rose mrose@dbc.mtview.ca.us Jonathan Saperia saperia@decwrl.enet.dec.com Mark Schaefer schaefer@davidsys.com John Seligson johns@ultra.com Timon Sloane peernet!timon@uunet.uu.net Bob Stewart rlstewart@eng.xyplex.com Emil Sturniolo emil@doe.dss.com Bruce Taber taber@interlan.com Maurice Turcotte dnmrt@rm1.uucp 5 David Waitzman djw@bbn.com Drew Wansley dwansley@secola.columbia.ncr.com David Ward dward@chipcom.com Gerard White ger@concord.com Linda Winkler lwinkler@anl.gov Mark Wood markl@dsiinc.com Joseph Zur fibrontics!zur@uunet.uu.net 6