OPS Area RMONMIB WG Meeting Minutes 49th IETF December 11 & 13, 2000 Minutes by Andy Bierman Review Material --------------- (A) draft-bierman-rmonmib-vds-mib-01.txt (B) draft-bullard-pcap-00.txt (C) draft-ietf-rmonmib-apm-mib-02.txt (D) draft-ietf-rmonmib-appverbs-01.txt (E) draft-ietf-rmonmib-dsmon-mib-03.txt (F) draft-ietf-rmonmib-hcrmon-06.txt (G) draft-ietf-rmonmib-iftopn-mib-03.txt (H) draft-ietf-rmonmib-tokenring-00.txt (I) draft-ietf-rmonmib-tpm-mib-01.txt (J) draft-kalbfleisch-sspmmib-01.txt Agenda ------ 1) WG Status - new RFCs - RMON Protocol Identifier Reference (2895) - RMON Protocol Identifier Macros (2896) - HC-RMON MIB documentation update (F) 2) Transport Performance Metrics MIB (I) 3) Application Performance Measurement MIB (C) 4) RMON Extensions for Identifying Application Protocol Verbs (D) 5) Token Ring Extensions to the RMON MIB (H) 6) RMON Extensions for Differentiated Services (E) 7) RMON Extensions for Interface Parameters Monitoring (G) 8) Remote Packet Capture (B) 9) Definition of Managed Objects for Synthetic Sources for Performance Monitoring Algorithms. (J) 10) RMON Extensions for Virtual Data Sources (A) Minutes ------- 1) Status of Internet Drafts under IESG review The WG Chair presented a status summary of the I-Ds before the WG. All HC-RMON issues have now been resolved, and the HC-RMON MIB (F) will soon be completed. The RMON1-HC MIB and the RMON2-HC MIB will be terminated as Internet Drafts. The MIB modules within these documents will be extracted and archived at an 'official' WEB site, (such as the RFC Editor pages site). The HC-RMON MIB will contain pointers to these updated MIB modules. Refer to the slides for more details on these issues. 2) TPM MIB The latest Transport Performance Measurement MIB (I) was presented and discussed. The highlights of the discussion are listed here. Refer to the TPM slides for details on the presentation. The purpose of TPM is troubleshooting. The expected usage: APM reports that a transaction is out of scope (e.g. delay too long or service unavailable). TPM provides the drilldown into a (potentially) complex application transaction by collecting transport metrics relevant to the particular application being measured. The group discussed streaming protocols and how APM/TPM handles some applications of this type. Metrics such as message jitter and stream jitter are difficult to describe in a generic way (for all applications). There is also concern that the APM and TPM metrics need to be defined in a rigorous manner, consistent with the existing set of IPPM metric definitions (perhaps by the IPPM WG instead of the RMONMIB WG). The specification of APM and TPM metrics remains an open issue. Changes expected in the next revision of the TPM MIB: - The metric definitions are still missing. They must be added soon, or perhaps defined by the IPPM WG - The 'study capabilities' MIB objects from the PM-CAPS MIB (now defunct) need to be adapted, and moved to this MIB. - Any INDEX linkages between APM and TPM or SPPM and TPM need to be finalized 3) APM MIB The latest Application Performance Metrics MIB (C) was presented and discussed. The highlights of the discussion are listed here. Refer to the APM slides for details on the presentation. The WG discussed the notion of user-defined applications, and their representation in the apmUserDefinedAppTable and/or the protocolDirTable from the RMON2-MIB module. It is unlikely that an application developer will be able to support user-defined applications in a vendor-independent manner, since there are no real guidelines for populating values of the apmUserDefinedAppApplication object for a given application. This new feature is in conflict with the PD user-extensibility design, as specified in RFC 2895. Protocols which cannot be identified within the 4-octet-per-layer scheme are given static assignments by the WG, within the 'ianaAssigned' base layer number space. The WG also discussed the behavior of an APM agent for measurements of the availability metric for streaming protocols. It is difficult to describe the criteria for distinguishing between degraded service and unavailable service, in a generic manner. The APM algorithm for measuring availability needs to be specified in more detail. Currently, there is no notion of measuring degraded service at a given moment in time. The availability 'singleton' metric is a '0' or '100' score, with no possible values in between. The WG discussed how this measurement could be achieved and conveyed in the MIB. There was no group consensus that a new 'degraded service' metric, or changes to the existing availability metric, are desirable or achievable. The issue will be discussed further on the mailing list. 4) APP-VERBS I-D The 'Protocol Directory Extensions for Application Verbs' (D) draft was presented and then discussed. Only minor changes and corrections were made to the latest draft. There were no comments on this draft. The WG mailing will be used for comments and suggestions for application verb additions, The next version of this document will contain more application verb definitions. 5) TR-RMON MIB The group briefly discussed the Token Ring RMON MIB I-D (H), which is needed for the TR-RMON MIB to advance from Proposed to Draft Standard (RFC 1513 is in SMIv1 format). The WG Editor will make some minor changes, and the new draft will then be subject to a WG Last Call. 6) DSMON MIB The group discussed the latest version of the Diffserv Monitoring MIB (E), and mailing list comments related to this MIB. The draft will be updated and subject to a 2 week WG Last Call. A summary of the recommendations reached by group consensus is listed here. Refer to the slides for more details. - a DSMON capabilities scalar object of type BITS will be added, similar to the probeCapabilities object in RFC 2021. - the dsmonPdistTimeMark TimeFilter INDEX in the dsmonPdistStatsEntry will not be removed - the INDEX order of the dsmonPdistStatsEntry will not be changed. [Ed. note - actually, Lester is right; the order of the protocolDirLocalIndex and dsmonAggGroupIndex INDEX components should be reversed. The current ordering forces the NMS to retrieve virtually the entire table to determine which protocols are present on an interface, which share the same aggregation group ID. The instance fan-out will be at the protocol INDEX, not the aggGroup INDEX.] - smilint errors - dsmonHostAddress must have SIZE restriction This error exists in RMON2 as well. A range of (0..65535) will be added to remove the error without changing the semantics of the object at all. - some identifier names longer than 32 characters These names will be shortened in the following manner: - dsmonPdistControl* -> dsmonPdistCtl* - dsmonPdistTopNControl* -> dsmonPdistTopNCtl* - dsmonPdistTopNControlGeneratedReports -> dsmonPdistTopNCtlGeneratedReprts - dsmonHosttControl* -> dsmonHosttCtl* - dsmonHostTopNControl* -> dsmonHostTopNCtl* 7) IF-TOPN MIB The Interface TopN MIB (G) was presented and then discussed. There was some discussion clarifying the normalization mechanisms. There are some smilint errors (refer to the slides) that will be addressed by the author. A new version of the draft will be posted soon and subject to a 2 week WG Last Call. 8) Remote Packet Capture A proposal for a new approach to packet capture (B) was presented and then discussed. A probe, router, or switch would save a timestamp for each packet as it was received on the monitored interface. A new packet is created, containing the desired packet slice to be captured, and a header is prepended which contains packet info such as the real packet length, ifIndex of the ingress interface, the ingress timestamp, etc. This new packet is then sent to a pre-configured IP host for collection and analysis of the captured packet slice. Refer to the slides for details. The group did not seem eager to take on this work at this time, although this work could be viewed as an extension of the SMON portCopyTable. This proposal should be discussed further on the WG mailing list to determine interest and feasibility. 9) SSPM MIB The SSPM Framework draft (J) was not discussed at this meeting, although a new version has been published. The WG is waiting until APM and TPM are more stable before officially starting work on configuration of synthetic traffic sources for performance management. Hopefully, this work can begin soon after IETF #50. 10) VDS MIB The Virtual Data Source MIB (A) was not discussed at this meeting. A new version has been published, containing minor enhancements and corrections. The WG will possibly consider this document for future work, after the current work is substantially completed. 11) PM-CAPS MIB The Performance Metrics Capabilities document has been withdrawn, and was not discussed at this meeting. The dynamic and variable nature of the PM Study is not applicable to APM. This functionality will be moved (in some form) to the TPM MIB. 12) SMON MIB Advancement This topic should have been on the agenda, but was missed. The WG will start the required tasks needed to advance the SMON MIB (RFC 2613) from Proposed to Draft Standard. Commercial implementations are already available, and this MIB is eligible for advancement. An announcement will be sent to the WG mailing list soon.