Re: [tcpm] question about TCP-AO and rekeying

[Joe Touch <[e-mail]>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1



Eric Rescorla wrote:
...
> OK, I fear we're talking past each other.
> 
> My claim is that the system must enforce the following invariant:
> 
> * There must be only one MKT defined that matches any given set
>   of parameters: (s-ip, d-ip, s-port, s-port, key-id).
> 
> This includes wildcards and ranges.

That one is a lot easier, since you're including the key-id. That's
available only for either established connections (e.g., bound to the
socket after connection establishment for outgoing, or indicated in the
segment for incoming).

For the initial outgoing SYNs from legacy applications, we need the MKT
match to be unique per:

	(s-ip, d-ip, s-port, d-port)

This is the case that's causing the problem for two cases:

a) overlapping MKT patterns, as would be used for catch-alls vs. more
specific patterns (which we could prohibit, but I would like to try not to)

b) identical MKT patterns, as would be used when a a primary and backups
are added

The TSAD/TAPD had a notion of grouping MKTs together when patterns
overlapped, so that (b) would allow selection of a 'currently designated
primary' to be used for outgoing SYNs. Getting rid of that structure -
which you claimed was an implementation issue - is omitting structure in
the key groupings that allows (b), and also enables (a).

If we omit the TSAD/TAPD, how do we express the notion that a given key
is preferred for outgoing SYNs?

Joe
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAko7xZsACgkQE5f5cImnZrscrQCfYfhimRCWFDvpoFXfBrjTwaK6
HiwAn1EQpeYdIPdd6f+SCXG9cUkfDfnl
=jdMC
-----END PGP SIGNATURE-----