Re: draft-ietf-msdp-spec-16.txt vs. security

["Steven M. Bellovin" <smb@research.att.com>]

In message <200305121501.h4CF19903792@windsor.research.att.com>, Bill Fenner wr
ites:
>
>MSDP is an evolutionary dead end, except it's all we've got.  There are
>no promising ideas waiting in the wings for supporting interdomain any-
>source multicast.  Given that, I'd expect it to keep living for a long
>time, and for people to want to build on it even though it's not what we
>want...
>
>I think that MUST in an Experimental is meaningful.  Although 2119's
>introduction mentions standards track documents, it seems to me that
>the definitions still apply within the context of the document.
>
>I think MUST have keyed-MD5 and MUST interoperate with implementations
>that don't have it would work.  The latter allows existing implementations
>to play, and the former supplies some modicum of security when moving
>forward.
>

That works for me.  Alex?



		--Steve Bellovin, http://www.research.att.com/~smb (me)
		http://www.wilyhacker.com (2nd edition of "Firewalls" book)