HIP BOF Review

["James Kempf" <kempf@docomolabs-usa.com>]

I basically agree with EKR's review, but have a few additional comments. I
think it might be helpful to get them to try separating out the components
of HIP that EKR identified and think about other possible options to replace
them. In particular, either though experiment or careful analysis, I think
they need to explain what benefits the cryptographic identity binding has
beyond simply separating the identity and location. If they can't come up
with a very good reason for that, then generating the identifier
anynonomously or by hashing the DNS name might be a better option, since it
would require fewer changes to things like the key exchange protocol, and
the need for DNSSEC.

Also, a major task I see for the WG is coming up with a good solution to the
rendezvous server problem. Right now, they are thinking of using dynamic DNS
or their own, custom solution. Both of these have properties that make them
less desirable alternatives than Mobile IP from a deployment or utilization
standpoint. On the other hand, if they could come up with something that
didn't require new infrastructure, such as MIP does, but has good
properties, HIP could be more attractive in some ways.

Much of the discussion at the BOF revolved around "why should IETF charter a
WG to do experimental drafts?" Clearly, the charter for a WG targetting
experimental publication rather than standards track must be different, and
there were some indications in the proposed charter that, deep down, the HIP
guys really think it should be standards track (like the inclusion of a
MIB).  I think the ADs may have to keep the WG's feet to the fire in this
regard, so they don't drift off into thinking they are doing a standards
track protocol. A good place to start is the charter. Besides being cleansed
of standards track-like goals, I think the charter should empahsize some
engineering research questions the WG is trying to answer.

The discussion at the end of the meeting about whether a WG should be
chartered was confused by a question about whether people had enough
information to decide, then the room needed to indicate that, etc., but in
the end, it seemed most folks felt that a WG is a good idea.

            jak

            jak