Hello Santosh, authors, Alvaro and list members, for the quoted section 3 of draft-ietf-bfd-seamless-base it gets even "worse", as it says for the example in that section:
The IS-IS with SystemID xxx (node A) allocates an S-BFD discriminator 123, and advertises the S-BFD discriminator 123 in an IS-IS TLV. The IS-IS with SystemID yyy (node D) allocates an S-BFD discriminator 456, and advertises the S-BFD discriminator 456 in an IS-IS TLV. So it puts the IGP (IS-IS in this case) into an authoritative role. The IS-IS teams respond with "whoa, wait, I'm just the messenger" :-) Section 3 also says An S-BFD module on each network node allocates one or more S-BFD discriminators for local entities, and creates a reflector BFD session. A bit if a contradiction - who is allocating now, S-BFD or the IGP? I would think it is the S-BFD module that allocates and orchestrates. It uses an IGP or other means to transport discriminators to S-BFD modules on other nodes. The reason is that S-BFD creates the BFD reflector, so it should know what this reflector is for (maybe pure IPv6, or used for specific QoS etc.). There is another aspect: the protocols (OSPF, IS-IS, L2TP) plan to transport a list of discriminators. Okay ... but how is the receiver S-BFD module making sense out of this list? Would have expected something like (type, discriminator). The protocols don't need to understand the details, only that the API transports one or more of these tuples in/out of the protocol module. S-BFD would know/define what a particular type means. Just asking before we send OSPF, IS-IS, L2TP into the wrong direction :-) Regards, Marc On Tue, 8 Dec 2015 16:47:53 +0000, Alvaro Retana (aretana) wrote: > Santosh: > > Hi! > > There are 2 more things (Major) that I need you to address. Sorry, I > almost forgot. > > (A) Please take a look at the SecDir comments here [1]. > > [1] > https://mailarchive.ietf.org/arch/msg/secdir/o7CSlWeRPh4-BZBTI_eDyoBM7ak > > > > (B) > > During the IESG review of draft-ietf-isis-sbfd-discriminator the question > came up about how the mapping of S-BFD discriminators (advertised by IS-IS > in that draft, but there are also similar OSPF and L2TP drafts) to > specific applications/entities is to be done. This conversation ended up > in me placing a DISCUSS on that document [2]. > > The problem is that draft-ietf-isis-sbfd-discriminator (and the OSPF and > L2TP drafts) declare the mapping out of scope. *AND* This document > (draft-ietf-bfd-seamless-base) says this in Section 3. (Seamless BFD > Overview): > > An S-BFD module on each network node allocates one or more S-BFD > discriminators for local entities, and creates a reflector BFD > session. Allocated S-BFD discriminators may be advertised by > applications (e.g., OSPF/IS-IS). Required result is that > applications, on other network nodes, possess the knowledge of the > mapping from remote entities to S-BFD discriminators. The reflector > BFD session is to, upon receiving an S-BFD control packet targeted to > one of local S-BFD discriminator values, transmit a response S-BFD > control packet back to the initiator. > > This text reads to me that S-BFD is expecting ("Required result") the > mapping to be somehow provided by the "applications (e.g., OSPF/IS-IS)". > Note that one possible interpretation is not for OSPF/IS-IS to "know" > anything about S-BFD discriminators/entities, but to transport that > information (similar to transporting discriminators). I can see at least > one relevant use case (in draft-ietf-bfd-seamless-use-case) that seems to > require the ability to distinguish and map: Section 3.8. (Multiple BFD > Sessions to Same Target). > > The point to address is this: what is the expectation (from the S-BFD > point of view) with respect to the mapping? > > Answering may require a WG-wide discussion. Depending on the answer, > there may be obvious effects (and work needed) in the isis, ospf and > l2tpext WGs, so please (with the help of the chairs) work with them. > > > Thanks! > > Alvaro. > > [2] > https://datatracker.ietf.org/doc/draft-ietf-isis-sbfd-discriminator/ballot/ >
