Re: [pim] Benjamin Kaduk's Discuss on draft-ietf-pim-msdp-yang-15: (with DISCUSS and COMMENT)

[Alvaro Retana <[e-mail]>]

On February 28, 2020 at 7:05:28 PM, Benjamin Kaduk wrote:

Ben:

Hi!

Thanks for your review!

Just replying to a couple of points.  I'll let the authors take care
of the rest of the replies and the updates.



> ----------------------------------------------------------------------
> DISCUSS:
> ----------------------------------------------------------------------
>
> I don't think there's enough clarity on what authentication schemes are
> supported and how the YANG configuration interacts with MSDP operation.
> If I'm reading^Wsearching through RFC 3618 correctly, the only supported
> authentication mechanism is TCP-MD5 (RFC 2385), and there have been no
> updates to RFC 3618 that are indicated in the RFC database. However,
> RFC 2385 is obsoleted by RFC 5925 (TCP-AO). Can TCP-AO be used with
> MSDP? What protocol elements or operation are controlled by the
> "authentication" container? What algorithms are valid for use with the
> "password" case? RFC 8177 is the sole reference for both the
> "key-chain" leaf and the "password" case, but that does not seem a
> sufficient reference from which to implement.

Yes, more clarity is needed.

MSDP is only specified to work with TCP-MD5.  While TCP-AO could be a
possibility, there is no current specification or implementation (that
I know of) that goes beyond MD5.



...
> ----------------------------------------------------------------------
> COMMENT:
> ----------------------------------------------------------------------
>
> This document (and the YANG module) lists six authors, which is more
> than the typical five and would normally spark some discussion. I do
> see that the shepherd writeup indicates that all six members of the
> design team were quite active, and presume that this indicates that it
> is appropriate to have the larger author count than is typically seen.

It is.


Alvaro.