Re: [Hipsec] The HIT prefix once again (was Re: Re: Type 1 and 2 HITs)

 In your previous mail you wrote:

   >> Now, this goes back to the discussion whether to divide the
   >> first byte in the HIT into two fields, one being fixed for
   >> now ("This is a HIT" and not an IPv6 address)

=> HIP is not the only protocol which could take avantage from
a dedicated prefix marked as not an address/not routable.
I wrote a short I-D draft-dupont-ipv6-cgpref-01.txt which could
work for at least HIP and a TMSI-like location privacy solution
for MIPv6. I consulted IANA at an IETF meeting just before writing
the first version.
The idea is both to get such a prefix ASAP and "to be enough" in order
to get a very short one. IMHO alone we have no chance to get better
than a 16 bit prefix...

   >> As I have stated before, I still think that instead of
   >> defining a single 8-bit prefix (to be allocated from the IPv6
   >> address space) it is probably a better idea to have an e.g.
   >> 5-bit prefix and then have initially 3 bits (8 choices) for
   >> encoding the hash.

=> I believe it could be better to share the prefix with other (and with
weaker requirements) usages, i.e., to win one bit (2 x) at the cost of 1/8?

   > ii) not knowing what IANA would like to do:  assign HIP a /5 or a / 
   > 8 out of the IPv6 space, or neither.

=> I am afraid someone is dreaming... at least you should get strong
support! (:-)

   > It would help me to decide if I understood the process by which IANA
   > will grant or deny such requests, and whether we can leave the issue
   > "pending IANA resolution" for now.

=> easy, we need an IPv6 WG RFC.

   My (perhaps flawed) understanding of the situation is that IANA
   more or less does whatever the IETF/IESG asks it to do.  See
   Section 4 of RFC3513.  Based on that my understanding of the situation
   is that we need to achieve IETF consensus on the prefix anyway.

=> IMHO we need a bit more.

   In practise, that probably means convincing the following bodies,
   roughly in this order:

      IPv6 WG chairs (Bob Hinden and Steve Deering)
      IPv6 WG
      (IAB)
      INT area ADs
      INT area
      IESG
      IETF in the large

=> there are not (or at least should not be :-) kings at the IETF.
The path should be the IPv6 WG (chairs should follow the WG), IESG
and the IETF in the large.
The first step is to have an I-D (so I propose to join us), the second
is to get support of the IPv6 WG about it (WG item, WG last call, etc).

   So, I would suggest that we start talking to Bob and Steve ASAP,
   and once we've got some opinion from them, ask for cross-WG review
   from the IPv6 WG.  I'll try to remember to take this up with Bob
   on Sunday.  (Please remind me during the week.)

=> so you need an I-D. IMHO it will be easier with a short one than
with an "en passant" paragraph in a HIP architecture one.

Regards

Francis.Dupont@enst-bretagne.fr

PS: if you agree to join us, we can ask the IPv6 WG chairs to poll
the IPv6 WG about a WG item. I can't see technical concerns so it should
be easy and the discussion (in the mailing list about the WG item) should
only be about the length of the prefix.

_______________________________________________
Hipsec mailing list
Hipsec@lists.ietf.org
https://www1.ietf.org/mailman/listinfo/hipsec