Re: [lisp] Comments to LISP+ALT draft

[Sam Hartman <hartmans-ietf at mit.edu>]

>>>>> "Vince" == Vince Fuller <vaf at cisco.com> writes:

    Vince> I am a little hesitant to simply make all of the changes
    Vince> that you have offered because the LISP+ALT spec pre-dates
    Vince> LISP-MS and is still intended to be independent of it. I do
    Vince> agree that this historical fact has created some
    Vince> inconsistencies between the two documents but need to
    Vince> consider how to modify the wording in LISP+ALT to clarify
    Vince> things without at the same time sacrificing its
    Vince> independence from LISP-MS.

Speaking as an individual.  I think that addressing these comments
while preserving independence is quite important.  It seems that
lisp+alt can distribute mapping information whenever it somehow knows
what BGP prefixes to announce.  That can happen either because the
prefixes are local, because "static routes" are configured, or because
the router learns the prefix through something else.  One something
else is that the router may be a map server (or receive information
through some mechanism from a map server).  However I don't see any
reason why the map server relationship should be special in this
specification.

I think that during the specification process we want to be careful to
define clean architectural interfaces around the components.  However
I do see the points that are made here and believe that addressing
these concerns could be an excellent opportunity to move towards these
cleaner interfaces.

I look forward to reviewing the editors' proposals for resolving these issues.