RE: draft-purser-pppext-pppcn-00.txt

Kevin Purser (QA/EMC) writes:
> > No additional security risks? Really? How about revealing an 
> > IP address
> > before you have even authenticated the peer? Or accepting an unlimited
> > number of IP packets waiting for some negotiation to complete?
> > Neither of these significant security issues are present in RFC1661,
> > and both represent leaks that can easily be abused. If anything, this
> > draft _adds_ serious security risks.
> 
> Perhaps the text wasn't clear enough.  While the IP address may be "revealed" to the peer, it's of no consequence if packets sent by that peer will *not* be forwarded by the authenticating peer until *after* completing the auth phase, right?  And the sending of IP packets was really only a further optimization to the proposal, not a requirement.

I agree with that.  The proposal doesn't omit the required security
steps, and IP addresses seem to me to be hardly something that could
be considered "secrets."

-- 
James Carlson, IP Systems Group                <[email protected]>
Sun Microsystems / 1 Network Drive         71.234W   Vox +1 781 442 2084
MS UBUR02-212 / Burlington MA 01803-2757   42.497N   Fax +1 781 442 1677