Document:  draft-ietf-ipv6-privacy-addrs-v2-04  
Reviewer: Lakshminath Dondeti [ldondeti@qualcomm.com]
Review Date:  Wednesday 12/14/2005 3:16 PM CST
Telechat Date:  Thursday 12/15/2005

Summary: I feel that the draft is not ready to move forward as DS for two reasons.

1) As has been pointed out by others, we have a report of only one 
partial (the case of no stable storage of a history value)
  implementation.  Don't we need two full implementations?

2) The spec is not algorithm agile:
"This document proposes the generation
    of a pseudo-random sequence of interface identifiers via an MD5 hash."
I'd like to see new Draft Standards to be algorithm agile.  Try "via 
a cryptographic hash algorithm," and specify MD5 as a MUST.  I'd 
prefer some other hash algorithm, but it looks like it is ok to use 
MD5 here since the algorithm seems to rely on the preimage property.

Finally, the security considerations section ought to talk about the 
reliance on the preimage property.  At the moment the section says 
very little; talks only about how privacy capabilities disallow 
filtering capabilities.