Re: [Isis-wg] Benoit Claise's No Objection on draft-ietf-isis-sbfd-discriminator-02: (with COMMENT)

["Alvaro Retana (aretana)" <[e-mail]>]

On 11/18/15, 2:27 PM, "Les Ginsberg (ginsberg)" <[e-mail]> wrote:

Les:

>I think you are referring to this text from
>https://www.ietf.org/id/draft-ietf-bfd-seamless-base-05.txt
>
>Section 3

Yes, that's what I quoted below (and in the DISCUSS itself).

>
>"...Allocated S-BFD discriminators may be advertised by
>   applications (e.g., OSPF/IS-IS)."

The next sentence says:  "Required result is that applications, on other
network nodes, possess the knowledge of the mapping from remote entities
to S-BFD discriminators."


>This is (in my view) poorly worded bordering on inaccurate.

At the very least the text is confusing.  I think you can interpret it
either way.


>In the context of the IS-IS/OSPF drafts we are defining a way to
>advertise the S-BFD Discriminator(s) assigned to a given node. This does
>not imply any usage of these values by the IGPs.
>
>It is conceivable that the IGPs could find a use for an S-BFD session and
>in that context the IGPs would be an S-BFD client and considered an
>"S-BFD application". But that is NOT what is covered in the current IGP
>drafts - nor is it within scope.
>So I reiterate, the current discussion regarding how applications decide
>which S-BFD discriminator should be used is out of scope for the existing
>IGP drafts.
>
>The entire discussion belongs in the BFD WG.

Agreed, the discussion should be there.  I have asked the Chairs to lead
that discussion (while keeping everyone else in the loop: isis, ospf,
l2tpext).

There is nothing specified in an S-BFD document that talks about the
mapping topic (except for the text above).  As I said below, it worries me
that S-BFD is expecting the "applications" (yes, the draft could've
probably used a better word and not lumped the IGPs in there...but it did)
to do something that is not being done.

Alvaro.


 
> 
>
>   Les
>
>
>> -----Original Message-----
>> From: Alvaro Retana (aretana)
>> Sent: Wednesday, November 18, 2015 10:12 AM
>> To: Les Ginsberg (ginsberg); Benoit Claise (bclaise); The IESG
>> Cc: [e-mail]; [e-mail];
>> [e-mail]; [e-mail]; [e-mail]
>> Subject: Re: [Isis-wg] Benoit Claise's No Objection on
>>draft-ietf-isis-sbfd-
>> discriminator-02: (with COMMENT)
>> 
>> On 11/18/15, 12:47 PM, "Les Ginsberg (ginsberg)" <[e-mail]>
>> wrote:
>> 
>> Les:
>> 
>> >Neither IS-IS nor OSPF are the "application" here. We are simply the
>> >transport for an opaque piece of information. It makes no sense to me
>> >to ask the IGPs to clarify what is an issue between S-BFD and an
>>application.
>> 
>> What you say makes sense.
>> 
>> It also makes sense to me that one could argue that since we're
>>transporting
>> opaque information we could also include an extra piece of opaque
>> information to let someone else know what the first piece was for.
>> Note that I'm offering that scenario just as an example of the fact
>>that there
>> are at least a couple of potential interpretations.
>> 
>> Because of the language used in the Base S-BFD draft (where it does call
>> OSPF/IS-IS an "application", [1]), I just want to play it safe and make
>>sure
>> we're not pointing at each other and that result won't be that no one
>>does
>> the mapping.
>> 
>> Thanks!
>> 
>> Alvaro.
>> 
>> 
>> [1]
>> 
>>    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
>> 
>