********************************************************************** IETF 84 PWE3 - TUESDAY, July 31, 2012 - 17:00-18:30 (90/90 min) Room Regency C ********************************************************************** Chairs: Matthew Bocci and Andy Malis Secretary: David Sinicrope Minutes: Scott Mansfield (scott.mansfield@ericsson.com) 1. Agenda bash, WG Agenda and Status - Andy MALIS and Matthew BOCCI - Openflow presentation has been moved to the end of the session. - Four new RFCs since last IETF and two in the editors queue. - There are three documents with the IESG. PW redundancy, VCCV Implementation Survery Results, and Ethernet OAM Interworking. - P2MP PW Requriements sent back to the working group. - ICCP is on the agenda, status-reduction has expired. Luca needs to refresh this draft, and a last call is needed, also a LC on ICCP is needed. - A new version of mpls-tp-oam-config is to be submitted after IETF 84 - vccv-for-gal scope needs to be reduced - mpls-tp-ms-pw will be refreshed Chairs request that IESG comments/AUTH 48 comments be done in a timely fashion. The best time to comment is before the IESG chat. 2. Virtual Circuit Connectivity Verification version 2 (VCCV2) - Tom Nadeau http://tools.ietf.org/html/draft-nadeau-pwe3-vccv2-00 Tom Nadeau presenting VCCV is described in 4 different places. Goal is to put everything about VCCV in one place. The goal is to reduce the number of CC types down to 3 types. YJS -- at the Paris meeting, everyone wanted to drop the RA, but not TTL so down to three types. YJS -- To understand VCCV there are 4 documents and more if using BFD. The options are either to write one document that includes everything, or to write a short document that only obsoletes RA and adds Type 4. The WG needs to determine the way forward. Himanshu from Ciena requested clarification (the draft is correct, the slides are not clear.) TTL can be used for non-CW PWs? YJS -- If using CW -- Type 1, if not using CW, can use TTL. If using TP -- Type 4. With Ethernet over TP you need to use Type 1. Shahram, in favor of writing a .bis and also wants to get rid of TTL? YJS and George, the TTL Type is needed and there is no consensus to remove TTL processing. Type 3 needs to be deprecated. Stewart: A short document that clarifies the capability will be much faster to get out of the working group. Then do a rewrite if needed. Sense of the room: Brief document for just CC mode: 3 people Comprehensive document: 6 people Underwhelming response in the room. This discussion will go to the list 3. PW Congestion Considerations - Yaakov STEIN https://tools.ietf.org/pdf/draft-stein-pwe3-congcons-01.pdf This is cross area work, David Black would like this work to remain in PWE3. PW traffic is not inherently congestion-responsive and PWE3 has not defined any congestion mechanisms. There is an RFC that provides UDP usage guidelines (David Black). There are two cases discussed in this draft. 1) elastic PWs (can be responsive to congestion), 2) inelastic PWs that can not respond to congestion. elastic PWs are automatically TCP-friendly, inelastic PWs are often TCP-friendly (usually do not rrequire any additional mechanisms). Look at the PDF version of this draft to see the graphs. Request to accept this item as a PWE3 work item with the goal of producing an informational RFC. Call for feedback from the congestion-control community David Black, strongly agrees with accepting this work in PWE3. Giles: Elastic case, TCP backoff is not the main method for congestion control, this issue is not that big of a problem. What is fairness? YJS, in the original version of the presentation, a universal fairness theorem was introduced, the individual TCP flows are more important than the PWE aggregates. Thomas: What problem is trying to be solved here? YJS, not trying to solve a problem, trying to show that there is no problem. Discussion moves to the list. 4. PW endpoint fast failure protection - Yimin SHEN http://tools.ietf.org/html/draft-shen-pwe3-endpoint-fast-protection-02 Local repair is fast but temporary, while global repair is the opposite. All of the comments received have been addressed, the draft is ready for working group adoption. There is a fair proportion of the people that have read the document, think that it is ready for working group adoption. 5. Transporting PTP messages (1588) over MPLS Networks - Shahram DAVARI http://tools.ietf.org/html/draft-ietf-tictoc-1588overmpls-02.txt Draft that is active in the TICTOC working group. There are a number of people at the mic, but there is no time for questions. Stewart points out that this document should be discussed on the PWE3 list, because that is where the experts in PW are. Luca: There is no need for this, should not transport timing across the network. Shahram disagrees. YJS, as TICTOC chair, take the discussion to the TICTOC list. Post Meeting Note (8/20/2012): While the http://tools.ietf.org/html/draft-ietf-tictoc-1588overmpls-02.txt document is referred to above as "active in the TICTOC working group", it expired last April (Expires: April 9, 2012). Post Meeting Note (YJS): It should also be noted that in discussion of the 1588overMPLS at TICTOC meeting in Paris, IETF-83, it was recommended to split the document into four documents (TICTOC, MPLS, IGP, and PWE3) that will address changes in related areas and conduct reviews in most appropriate WGs, not only on TICTOC list. 6. Transparent SDH/SONET over Packet - Gert MANHOUDT http://tools.ietf.org/html/draft-manhoudt-pwe3-tsop-00 Today's solution is RFC 4842. Looking for a structure agnostic way to transport STM over MPLS. Done in a similar manner as SAToP (RFC 4553). Basic solution is extending SAToP to support STM. YJS: 4553 like techniques were considered (at the time). This draft is "hacks" SDH in a number of bad ways. When physical layer continuity is broken, the SDH architecture has requirements. Gert: This should be similar to Transport over OTN. Shahram: What is the need for this? YJS: I have seen this used in many places. Further discussion to list and create a liaison for the SG 15 meeting. 7. Pseudowire redundancy on S-PE - Jie DONG http://tools.ietf.org/html/draft-dong-pwe3-redundancy-spe-02 Author thinks the draft is complementary to draft-ietf-pwe3-redundancy-bit and would like working group adoption. Andy: Have more people read the draft and comment on the list, then discuss working group adoption. 8. ICCP extension for MSP - Yuxia MA http://tools.ietf.org/html/draft-hao-pwe3-iccp-extension-for-msp-03 PW redundancy. People need to read the draft, and send comments to the list. 9. Definition of P2MP PW TLV for LSP-Ping Mechanisms - Sami BOUTROS http://tools.ietf.org/html/draft-jain-pwe3-p2mp-pw-lsp-ping-00 Define new LSP Ping procedures for P2MP PW. Defines new FEC 130 and 131 Pseudowire sub-TLV. Authors are looking for more feedback and also for working group adoption. Andy: Please read this draft, and PWE3 is looking for a new P2MP editor (Sami's draft depends on the P2MP draft. 10. MPLS-TP Pseudowire Configuration using OpenFlow 1.3 - Andrew MCLACHLAN/David MEYER http://tools.ietf.org/html/draft-medved-pwe3-of-config-01 This draft is about using OpenFlow for provisioning and OAM for PW. There is no OAM in OpenFlow today, so material was created in this draft to show how OpenFlow could be used for OAM. James Kempf: There is a paper that does exactly this. YJS: OAM? OpenFlow doesn't support this, so is this just an academic exercise? Andrew: Looking for some way to do provisioning and OAM in a programmitic way. There are other ways to this, ForCES for example. OpenFlow seemed to be a good fit to make the network more programmatic. Luca: OpenFlow is the wave of the future, so OpenFlow would need to support PW as well. Looking at the future of data center provisioning. Andrew: Where should this work live? Andy: Discuss further on the list. Is this appropriate for the PWE3 working group? Also, should have comments to the PWE3 list. 11. Static pseudowire configuration checking using GAP - Sami BOUTROS http://tools.ietf.org/html/draft-jc-pwe3-static-config-check-00 Andy: Ran out of time, look at the slides and read the drafts. 12. Requirements for IP/MPLS network transmission interruption duration - Fan Peng http://tools.ietf.org/html/draft-fan-opsawg-transmission-interruption Andy: Ran out of time, look at the slides and read the drafts. ********************************************************************** REMOTE INFORMATION FOR THE PWE3 SESSION(S) ********************************************************************** Remote Participation Info: http://www.ietf.org/meeting/84/remote-participation.html Audio will be on channel 5 (Regency C) Jabber room is "pwe3"