Re: [pim] AD Review of draft-ietf-pim-bfd-p2mp-use-case-05

[Alvaro Retana <[e-mail]>]

On August 18, 2021 at 1:53:29 AM, Michael McBride wrote:


Mike:

Hi!


I just saw Greg's response -- and it prompted me to go back to your
message.  In short, it is not clear to me what the difference is
between the options you proposed.


> Please consider either 1) dropping the "use-case" from the draft name and
> keep it a specific pim-bfd-p2mp solution and less generalized or 2) keeping
> the existing name and make it more generalized with more use cases as Alvaro
> suggests.

The name of the draft in unfortunate because it is a specification,
not an use case.  But it is just a name.  In the grand scheme of
things, the title of the document, which Greg has already changed to
"The Use of BFD for Multipoint Networks in PIM-SM", is much more
important.  IOW, changing the name won't solve anything at this point.
[OTOH, it will create confusion because a new name implies a new
document which hasn't been adopted, WGLC'd, etc..]


The part that I don't understand is what does "a specific pim-bfd-p2mp
solution" mean to you?   The way I read the document, there is only
one BFD p2mp solution for PIM: use the new Hello Option to bootstrap
the tail (rfc8562).  IOW, this is not an option -- and I don't see how
it could be "less generalized".   But I wonder if we're talking about
the same thing...?


The generalization part that I suggested to Greg, and where I would
appreciate more discussion, is as follows:

The document says that "any PIM-SM router, regardless of its role, MAY
become a head of a p2mp BFD session", so I believe there is no need to
constrain the application of BFD to monitoring the DR (as currently
written).  The action taken after a failure is detected is simply to
follow the base spec (rfc7761) -- there are no other changes!

As you can tell from the thread, we are in agreement that other use
cases are possible: monitor a downstream router that sent a Join, for
example.  Monitoring that router is within the scope of this document
because "any PIM-SM router, regardless of its role, MAY become a head
of a p2mp BFD session", and the resulting action is already specified
in rfc7761.

The net effect of generalizing the application of BFD is probably a
shorter document (no need to be specific about the DR, or any other
role) and maybe a couple of additional sentences about any
considerations related to a false positive (which will likely be the
result of misconfiguration or a rogue router).


I hope this is clearer -- sorry if it wasn't before.

Alvaro.