Minutes of the IP Flow Information eXport (IPFIX) WG 65th IETF, Dallas, Wednesday March 22, 2006 46 people in attendance submitted by Nevil Brownlee and Dave Plonka (co-chairs) based on notes from Ralf Wolter and Tanja Zseby. The text messaging log is available here: http://www.ietf.org/meetings/ietf-logs/ipfix@rooms.jabber.ietf.org/2006-03-22.html The meeting agenda and slides are available here: http://ipfix.doit.wisc.edu/IETF65/ [please see the agenda slides there for the sequence of topics: http://ipfix.doit.wisc.edu/IETF65/0-ietf65-ipfix-agenda.pdf ] ---- IPFIX Drafts Status (Nevil Brownlee) The chair reported that our Area Directors have reviewed all four drafts and had no serious concerns. He proposed one more editing iteration to fix the primarily editorial changes. Also, security considerations sections should be improved before last call. Regarding draft-ietf-ipfix-info-11, it was suggested that the draft include a template to be used to request new information elements. Also, the XML needs fixing. Regarding draft-ietf-ipfix-as-06, there were some concerns whether it properly documented situations in which IPFIX is most applicable and that some examples were "exotic" situations. The authors agreed to gather more feedback. The authors/editors present agreed to submit a new revision of their respective draft(s) by April 7, 2006, with the intent of these being submitted for IETF last call. ---- Interop Report (Elisa Boschi) Elisa reported on the second IPFIX Interop event March 1-2 in Salzburg. New issues were put into the Implementations Guidelines draft. [see slides for details: http://ipfix.doit.wisc.edu/IETF65/02-ietf65-ipfix-interop2.pdf ] ---- IPFIX Charter Discussion (Juergen Quittek) Juergen proposed that IPFIX recharter with a set of short term goals. Nevil suggested that they should be completed in less than one year and that the selections should encourage use of IPFIX as currently defined. To this end (below), we could select these goals from the existing set of IPFIX-related drafts. The chairs will propose a selection of these for discussion on the mailing list to determine consensus. Accordingly, a new charter incorporating those items will be proposed on the mailing list. [see slides for details: http://ipfix.doit.wisc.edu/IETF65/03-ietf65-ipfix-charter.ppt ] ---- The following candidate drafts were presented for consideration: draft-dressler-ipfix-aggregation-02 (Juergen Quittek) draft-boschi-ipfix-implementation-guidelines-01 (Elisa Boschi) draft-boschi-ipfix-reducing-redundancy-01 (Elisa Boschi) draft-bclaise-ipfix-reliability-01 (Benoit Claise) draft-trammel-ipfix-biflow-00 (Brian Trammel) draft-trammel-ipfix-file-00 (Brian Trammel) draft-kobayashi-ipfix-concentrator-model-01 (Atsushi Kobayashi) draft-kobayashi-ipfix-concentrator-mib-01 (Atsushi Kobayashi) draft-schmoll-ipfix-testing-00 (Paul Aitken) draft-stephan-isp-templates-01 (Emile Stephan) [see slides for details: http://ipfix.doit.wisc.edu/IETF65/ ] ---- PSAMP MIB / IPFIX MIB (Juergen Quittek) Juergen pointed out that there are two existing MIB drafts, PSAMP and the IPFIX concentrator MIB, which are nearly complete and could be combined and extended to form a single IPFIX MIB. [see slides for details: http://ipfix.doit.wisc.edu/IETF65/13-ietf65-ipfix-mib.ppt ] ---- Select Working Group Items, Establish new Milestones The chairs introduced discussion of the individual drafts (presented above), inviting suggestions as to which should be chosen as working group items. Dan Romascanu said that working groups can consider API-like work such as file formats, in IPFIX' case. Also he suggested that we mention only the high priority items in a new charter, rather than listing also potential future work. It was suggested that, in rechartering, we consider whether the group is defining probes' functionality or just a transport protocol. Under the existing charter it may have been a combination of both. In response, Dan mentioned that security and reliability concerns sometimes require that both be considered. There was general support in the room for the implementation guidelines and testing drafts as new working group items. On behalf of the security community, support was expressed for the biflow and file drafts. Also efficiency improvements, i.e. the reducing redundancy draft. A number of participants showed support for an IPFIX MIB. It was mentioned that some 20% of flow users use flow export for accounting and billing, so may benefit significantly from the reliability draft being selected. It was discussed whether or not some of the drafts' efforts would benefit from dramatic change in the protocol, such as a structured, multi-level information model, and therefore be deferred. In response it was pointed out that that could be many years away, and therefore a near-term solution must use only existing IPFIX features. The question was raised about how to select amongst the drafts. It was suggested that the drafts' number of authors/editors and dedicated reviewers could help to decide amongst too many "acceptable" drafts. Discussion regarding the choice amongst these candidate drafts will proceed to the mailing list. Nevil will propose a list to begin discussion there. -- $Id: minutes.txt,v 1.5 2006/04/10 16:20:09 dplonka Exp $