CURRENT_MEETING_REPORT_ Reported by Lee Oattes/University of Toronto AGENDA o Administrivia - Attendance list, someone to write minutes o Review of Charter - Objectives, goals of each document o Document 1: FINAL REVIEW (again!) - Behind Timeframe objectives - WILL submit to RFC process on 22 Feb. (Thursday) - DRAFT with minor revisions will be posted to mailing list for final comments THIS MONDAY. Comments will be accepted through 20 Feb. - Revisions to date: primarily syntax, semantics from the draft.01. Major change was moving examples from the statement of problem to an appendix. - My questions: Agreement on logged alert is ASN.1 OCTET STRING? No implication as to contents, and the goal is to avoid the dreaded OPAQUE. String might or might not be a construct (mine aren't right now :-). Since SNMP says primitives, we would be treating the OCTET STRING like an OPAQUE; the guy asking for it has to know it's intern format. Failure to reach closure TODAY means I will mak it a CHOICE of the two. o OPEN FOR QUESTIONS AND COMMENTS ON DRAFT 1, but ... - Protocol implementers are encouraged to voice concerns on specific deviations from SNMP/CMOT specs - Document syntax, layout is still fair game - Document goals have been agreed on for over 6 months, and are no longer a topic for debate (closure was reached). - Criticisms of mechanisms must be backed up with implementations (eg. I tried this, and it cost too much, didn't work, etc). MINUTES Meeting opened with Lou discussing the two documents that the WG is chartered with writing. The first deals with mechanisms that systems should use to control the flow of alert information. It is being authored by Lou. The second, written by John Cook, details specific techniques that an implementer of alerts may wish to consider. Its focus is on how alerts are generated. Final Review of document 1 Logged alerts should be of the SMI type OPAQUE. Lou was hesitant to use this, as OPAQUES use has been avoided. However, 1 the SNMP constraint of not using construct ASN.1 types ruled out the option of encapsulating an SNMP trap in an OCTET STRING. The use of a simple OCTET STRING would be a violation of ASN.1, and would not result in further parsing of the trap. A review of the DRAFT event and logging tutorial by the OIM Working Group followed. Lou agreed to contact Lee LaBarre about resolving any potential conflicts with the alert logging. The only problem centered on the Alert-Man requirement that a full log wrap. OIM shows this wrapping behavior as optional; halting behavior must be supported. The Alert-Man requirements might be presented as a subset of the OSI logging function. No problems were found with the OIM view of feedback/pin. The latest set of changes to the DRAFT will be posted on the 11th of this month. Comments will be accepted through the 22nd. At that time, the DRAFT will be submitted to the RFC process, as a protocol standard. Its initial, requested status will be elective. 2nd Document Lou will again post his sample technique/format to the mailing list. He also agreed to write a short overview of CMOT events that Brian will review for the second document. John discussed the format and status of the second document. Implementation reviews are being solicited for the following techniques: o threshold hysteresis (time and value based) o snapshots o thresholds on exceptions o thresholds built on counters, gauges, tidemarks o sliding window "pins" on each threshold o adaptive thresholds We are also looking for any other techniques in use. Submissions should represent actual implementations (can be in-house code), and can be posted to the working group list in the format Lou is using. John will post the DRAFT in its current state shortly. ATTENDEES Aronson, Cathy cja@merit.edu Cook, John cook@chipcom.com Feridun, Metin mferidun@bbn.com Handspicker, Brian bd@vines.dec.com Minshall, Greg minshall@kinetics.kinetics.com Newkerk, Oscar newkerk@decwet.dec.com Oattes, Lee oattes@utcs.utoronto.ca Perkins, David dave_perkins@3com.com Pokorney, Dave poke@nervm.nerdc.ufl.edu 2 Sheridan, Jim jsherida@ibm.com Waldbusser, Steve sw01@andrew.cmu.edu Wittbrodt, Dave dmw@cisco.com 3