8th November 2010 MPLS Session 1 (1 hour) Monday, Afternoon Session II 1510-1610 Room Name: Valley Ballroom C ---------------------------------------------- MPLS Session 2 (2.5 hours) Thursday, Morning Session I 0900-1130 Room Name: Valley Ballroom C ---------------------------------------------- 0/ WG Status Chairs, 20min 5 new RFCs since last time 3 Chinese operators and 2 European asked for consideration of draft-fang, which includes oam consideration. They are surprised that it was not put on the agenda. Huub: requested draft-bhh, surprised not to see it on the agenda Adrian: The discussion should first take place on the mailing list. If this happen, then discussion/presentation on the meeting can happen. But on the mailing list there were no discussion, not even from author. The meeting purpose is to progress discussion already started and not waste time discussing on the agenda. Italo: They wanted to understand if they were going to the right direction Alessandro: not agree with BFD CC. Concern from the industry, everybody from industry shall understand. China: disappointed not to be discussing really important points. First requirement draft Adrian: a formal complain on the WG shall do a letter to the AD, but not the WG meeting. The agenda was out for debate on the WG. People agreed that we should move on. Italo requires that this should be recorded (that the WG didn’t want to discuss draft-bhh and draft-ZANG 5 draft in IESG processing several wg drafts not on the agenda First last call on the identifier draft. This time, despite the last time, they believe it is ready. (see IANA slide) Draft on demand CV they made an early IANA allocation. Everybody agreed. But now 2 draft have conflicting IANA suggested values. Luca: not sure understand the rule. Propose to put a rule from the WG that the author insert a specific IANA section for allocation. Loa: until we don’t have a IANA allocation not allowed to put any value on the draft (put just TBD) Adrian suggest to continue this discussion on the routing area meeting Steward recommend to read the allocation draft 1/ Interop testing report Bijan, 15min Luyuan on behalf of Bijian. Isocore MPLS-TP interop testing. Participating company: Cisco, Ericsson, NEC, Hitachi and Ixia. Bidirectional co-routed LSP, MPLS-TP LSP, transport optimized OAM including LSP ping and BFD. 2/ Open discussion - Signal Degrade Ehud Doron, Nurit, 20min [missed] Nurit. Presenting the SD draft. Good distinction between SD and survivability and protection mechanisms. Recovery mechanisms can be triggered by SD can be triggered by… Not the job of the Survivability FW to define what SD is or how to measure it Not the job of the Survivability FW to define interactions between network conditions (that is policy) SD consideration of protection mechanisms Yacoov: it is a physical layer function. But for example it is not coming from traffic congestion. Dave Allan: Distinguished events from protection, because this may do things worse Kam Lam distinction between SD and SF. Nurit: yes, there is a distinction. Loa: how do we use this information on the MPLS layer? How does it have an effect on MPLS-TP? The intention of the author is to signal it to the far end of the stack. (to the LER of the LSP). Rahul: what is the goal of this mechanism? Propagate the info to the operator Initiate protection switching This is not clear how should it be treated differently from the normal SF case Nurit: in fact we are discussing it (on the slides as well) Julien. (referring to the triggering slide). In his knowledge is something more used to convey information. Nurit. Wanted just to show the different available mechanisms. Julien doesn’t agree that the trigger could be control plane. Russ Housley repeat that MPLS-TP need one single OAM solution (this was already been decided on the 2nd of August in Geneva) Ross Callon stress this is a good idea to ensure quality work 3/ Open discussion - MIP MEP Rolf, Yoshinori, Greg, 20min Yoshinori explains the slides (quick fault notification) Questions: David Allan: Proposal is the closest I've seen to a practical solution to the problem, but what it does do is PHP within the box, and so inherits all the OAM limitations of PHP. We'll have to evaluate the effort-reward accordingly before really proceeding. Manuel Paul: Ach-tlv, not to sure if helps, since also another draft was out but not much use of it has been done. Loa: when talking about MIP/MEP they were into architecture/modeling, but now they are moving into hardware solution. Ando also, when looking at idea at placing MIP/MEP in a point of the box, is it a function in a specific point in the box? Yoshinori: we need to continue it. He understands this is a big challenge, but for adapting MPLS-TP to transport we need that. Since transport already has this kind of characteristic. Dave Allan: recast the terminology but not change the framework document Julien: The transport network mentioned by Yoshinori is mainly layer 1 OAM, so we could relax all this requirements/features (since here we are doing layer 2). 4/ Working Group drafts - 25min draft-ietf-mpls-return-path-specified-lsp-ping-01 Mach Chen, 5min Loa: Need to resolve IANA draft-ietf-mpls-tp-cc-cv-rdi-02 Dave, 5min Question: Would it be possible to simply configure at the sink MEP the period of CV. Dave Allan: The problem is the sink MEP would need to know the period used for every source MEP. The PDU design does not allow both CC and CV period to be communicated. So to avoid a huge table of CV rates, we need a mechanism to reduce the amount of state, such as a universal minimum rate. Alessandro: I do not see the backwards compatibility issue here? Dave: It is not bits on the wire compatibility, it is the broad the number of implementations that can be upgraded to support CV functionality. George: You can send CV in every packet. draft-ietf-mpls-tp-fault-03 George, 5min draft-ietf-ccamp-mpls-tp-cp-framework-03 Lou, 10min No questions 5/ Focus points - 30 min draft-leymann-mpls-seamless-mpls-02 Thomas Beckhaus, 5min Rahul: multicast should be in a separate document and there is already another document on the table, we can discuss if it meets your requirements. draft-chen-mpls-6pe-mib-00 Gang, 5min No questions draft-fang-mpls-tp-use-cases-and-design-02 Luyuan, 10min No questions draft-fang-mpls-tp-security-framework-03 Luyuan, 5min No questions draft-he-mpls-tp-csf-02 Jia He, 5min Matthew: people were concerned about the overlap. This draft should point to the PW static signaling for the PW Jia: Agreed (which is proposal number 1 in the slides) Matthew: yes, and also mention LDP control Plane case Loa: Matthew to close the loops about that with the authors Matthew: just did it 6/ 75 min 6a/ mpls - 30 min draft-jin-jounay-mpls-mldp-hsmp-01 Lizhong, 10min draft-zheng-mpls-ldp-hello-crypto-auth-00 Vero Zheng, 10min draft-li-mpls-mt-applicability-requirement-00 Li Lianyuan, 5min Adrian: Sorry I have not read your draft Adrian: When you are talking about adding a new AS, is this one per VPN? Answer: No, just one additional AS covers all VPNs Adrian: This work impacts on L3VPN and IDR working groups. Please socialize it in both places. draft-raza-mpls-ldp-ip-pw-capability-00 Sami, 5min [missed]: Important work, bits to enable Ipv4 and ipV6 Yes, there is already. 6b/ mpls-tp - 45 min draft-boutros-pwe3-mpls-tp-mac-wd-00 Sami, 5min draft-koike-mpls-tp-temporal-hitless-psm-01 Koike, 5min Alessandro: Telecom Italia has the same concern draft-zhang-mpls-tp-path-segment-monitoring-01 Zhang Fei, 10min No questions draft-flh-mpls-tp-oam-diagnostic-test-01 Feng, 10min No questions draft-xiao-mpls-tp-throughput-estimation-01 Xiao Min, 5min (question) Differences between your draft and the draft in the previous presentation (answer) The previous draft is based in 1731 (question) I don’t understand, in the previous draft it was not discussed 1731 Loa suggest to merge the two drafts George suggest to continue the discussion on the mailing list to coordinate the work draft-huang-mpls-ring-signaling-extension-01 Yong Huang or Yuanlong Jiang, 5min draft-raza-l2vpn-pw-typed-wc-fec-01 Sami, 5min