SIGTRAN Thursday, August 03, 2000 Chair: Lyndon Ong (long@nortelnetworks.com) Reported by Matt Holdrege (matt@ipverse.com) Intro and Agenda Bashing The agenda was accepted. There were approximately 170 people attending the meeting. SCTP status SCTP is now in the RFC Editor's queue and will soon get an RFC number assigned. A well deserved round of applause was given to the editors, commentors and implementors of the spec. A number of other groups are now looking at using SCTP, such as beep. The document is now frozen while we do implementation and testing. There are a number of proposals for enhancements and we will track them on a list. The 1st bakeoff was held in June, based on v9. The next bakeoff will be in September or October, tentatively in Chicago. Email Lyndon if you are interested in this event. A 3rd bakeoff is proposed for March 2001. Lyndon will document the progress of the implementations in order to take SCTP to Draft Standard. SCTP MIB (draft-ietf-sigtran-sctp-mib-01.txt, Maria-Carmen Belinchon ) The MIB is being updated regularly and work will continue. Version 02 should be out in a month or so. The MIB is valid for the final version of SCTP. It uses SMIv2 and SNMPv3 will be the protocol. It supports both IPv4 and IPv6 (See ftp://standards.nortelnetworks.com/sigtran/presentations/pitt00/SCTP_MIBv01.ppt) One remaining issue is the assignment of the Object Identifier branch. Scott Bradner said there are people in the research community looking at TCP MIB's and it was recommended that we might get some good information from them to use in the SCTP MIB. Matt Mathis was suggested as a contact for this work. Chip Sharp mentioned that the Checksum error table should be measured per association rather than covering the whole SCTP layer. SCTP Socket API (Randy Stewart, draft-stewart-sctpsocket-sigtran-00.txt) Work is going on for an SCTP Socket API. The intention is to create an Informational RFC for this, possibly as an individual submission since this is not be an official working group item. The API contains both TCP-like and UDP-like APIs. Slides are at ftp://standards.nortelnetworks.com/sigtran/presentations/pitt00/socketsapi.ppt. SCTP Applicability Statements (Loede Coene draft-ietf-sigtran-sctp-applicability-01.txt) Work is proceeding on two applicability statements, one specifically about SCTP, and the other more generally about signaling transport using Sigtran protocols, including the adaptation layers. The editor, Lode, requests more participation to complete the drafts, especially information on the adaptation layers, and it was suggested that a small team be put together. Lode volunteered to host a drafting session in Belgium to progress the draft. Slides are at ftp://standards.nortelnetworks.com/sigtran/presentations/pitt00/applic.ppt. Adaptation Layers IUA/M2UA (Malleswar Kalla kalla@research.telcordia.com draft-ietf-sigtran-iua-04.txt, Ken Morneault, kmorneau@cisco.com draft-ietf-sigtran-m2ua-03.txt) The next draft of IUA should be ready for WG last call. M2UA had a number of technical comments from the participants, especially about whether full changeover with retrieval of packets needed to be supported or was not so useful. M2UA is targeted for WG last call after completion of IUA work. Slides are at ftp://standards.nortelnetworks.com/sigtran/presentations/pitt00/iua-m2ua.ppt. Proposed V5UA (Sanjay Rao < rsanjay@nortelnetworks.com>, draft-rao-sigtran-v5ua-00.txt) This proposes addition of adaptation layer handling for V5.2, based on IUA. It was agreed to make this a Working Group item, as a separate document that points to IUA for common features, and should be very similar to IUA as an adaptation layer, but with separate port number. Slides are at ftp://standards.nortelnetworks.com/sigtran/presentations/pitt00/v52_ietf.ppt. M3UA (Greg Sidebottom gregside@nortelnetworks.com, draft-ietf-sigtran-m3ua-03.txt) There should be a v4 draft in the next few weeks and then perhaps go to WG last call then, once IUA is completed as well. Greg is looking into providing a reference implementation after the document is completed. Maria and others have been working on a MIB. Finally, there seems to be considerable interest in an M3UA bakeoff in the near future, possibly towards the end of this year or early next year. Slides are at ftp://standards.nortelnetworks.com/sigtran/presentations/pitt00/m3ua.ppt. M2PA (Tom George tom.george@usa.alcatel.com, draft-george-sigtran-m2peer-01.txt) The Working Group agreed to make this a WG item, again as a separate document, although there was some concern expressed that it should be very close to M2UA, perhaps a smaller subset. It will be given lower priority compared to IUA, M2UA and M3UA. SUA (John Loughney john.loughney@nokia.com, draft-ietf-sigtran-sua-00.txt) SUA work is proceeding along, and will be treated with lower priority compared to the IUA, M2UA and M3UA. Rspool group The RSPOOL BOF was discussed and pointers to the list were given. Rspool takes material that has been discussed in Sigtran such as registration and application server management, but will look at a broader approach and scope than our previous discussions. SCTP Bake-off and other issues (Michael Tuexen, michael.tuexen@icn.siemens.de) Michael Tuexen described the results of the first SCTP bakeoff which was held in June in Munich, Germany. There were 12 companies participating. Results of the bakeoff were extremely good given the early stages of the protocol. Michael suggests doing more measurement-oriented testing in the next bakeoff, but Randy Stewart suggested that enough details had changed from v9 to v13 that we may need to retest everything that was tested in Munich, especially initiation of sessions. Michael also identified some performance testing results done with University of Essen that compared SCTP and TCP performance. Finally, Michael believes that it may be useful to look at security using TLS rather than IPSEC, and will plan to submit a draft on this. Slides are at ftp://standards.nortelnetworks.com/sigtran/presentations/pitt00/bakeoff.ppt. Mobility Chern Yap (cny@DCS.SHEF.AC.UK) spoke on using Mobility for SCTP. The chair emphasized that this group is not addressing mobileip issues, only any impacts there may be on SCTP itself. Conformance Test Suites Dhruv Sharma (dsharma@hns.com) spoke on conformance test suites for SCTP and the adaptation layers that have been submitted as drafts. Conformance test suites are not on the group's charter, and will not be added unless there is a strong demand for this. However, interested parties can still obtain the drafts, and treat them as individual submissions.