Re: status of hip

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

So I've been trying to figure out what I can say about this in 25 works or
less. I sent a long brain dump to Thomas but I didn't want to burden the
joint I* with so much text.

Coming at HIP from a wireless standpoint, it does seem to solve a number of
problems in ways that are solved less elegantly by Mobile IP. For example,
Mobile IP has no solution for a dual stack host moving between an IPv4 and
IPv6 network other than to have the host support both, while HIP solves this
problem with a single mechansim (or so I'm told). And, while the original
motiviation for Mobile IPv4 was to not change all the hosts on the Internet,
with Mobile IPv6, route optimization does, in fact, require changes to all
hosts on the (IPv6) Internet, so this is no different than HIP's
requirement. There are a couple of other mobility problems along these lines
HIP does a better job of solving.

That said, HIP solves the basic rendezvous problem of a correspondent
needing to find a moving host in a way that is similar to MIP, and so it
doesn't reduce the infrastructural requirement much, except to the extent
that it requires a single set of infrastructure for both IPv4 and IPv6.
Also, talking with Dave Crocker yesterday after the IAB Open Meeting, Dave
pointed out that HIP introduces another global namespace, and this is likely
to suffer from the same kind of issues that the current global address space
does. He is working on a solution which he claims is similar to HIP but
won't suffer from this constraint.

I think it would be hard to fit HIP into an IETF stovepipe, since it does
claim to solve so many problems, which may be why it has been classified as
a solution looking for a problem. There's an effort currently underway  (led
by Pekka Nikander) to simplify the basic spec and ultimately to submit it to
IESG as an individual contribution for publication as experimental (or, at
least, that's the plan now). I spoke with Pekka about it last nite, and he
thinks that it might be worthwhile having a higher level of IETF activity
after that has been complete, based on possible applications of HIP.

'Nuff said.

            jak