Re: I-D ACTION:draft-ietf-multi6-hba-00.txt

Personal comments:

I believe this is also ready to hand over to the future WG.

Just a couple of remarks.

1. You don't discuss the DNS at all - it clearly isn't a requirement
for the HBA mechanism itself to have any DNS entries, but surely
in reality at least one of the addresses will have to go into DNS?

2. A related point - in the discussion in 7.1 of MITM attacks, the
attack you describe only makes sense if the other end has no independent
check of *any* of the addresses in the address set. If even one of
them is (for example) in a trusted AAAA record, a MITM is excluded,
I think.

     Brian