Re: [midcom] SNMPv3 as MIDCOM protocol: Opinions?

Randy Presuhn writes:
 > Hi -
 > 
 > > From: Michael Thomas <[email protected]>
 > > Message-ID: <[email protected]>
 > > Date: Mon, 9 Dec 2002 17:08:25 -0800 (PST)
 > > To: Wes Hardaker <[email protected]>
 > > Cc: Michael Thomas <[email protected]>, [email protected],
 > >         "Harrington, David" <[email protected]>,
 > >         "Chris Elliott" <[email protected]>,
 > >         "Wijnen, Bert (Bert)" <[email protected]>,
 > >         "[email protected] (E-mail)" <[email protected]>
 > > Subject: Re: [midcom] SNMPv3 as MIDCOM protocol: Opinions?
 > > In-Reply-To: <[email protected]>
 > > References: <6D745637A7E0F94DA070743C55CDA9BA256F06@NHROCMBX1.ets.enterasys.com>
 > > 	<[email protected]>
 > > 	<[email protected]>
 > > 	<[email protected]>
 > ...
 > >    I'm not saying that IPsec would be a good fit;
 > >    quite the opposite, actually. I'm arguing that app
 > >    layer security without key management is, well,
 > >    useless was the first thing that popped into my
 > >    mind, but I should probably be more charitable.
 > >    It really makes deciding what to do at any sort
 > >    of scale *very* hard given the inferior choices.
 > > 
 > >    As I said, for MIDCOM this is a very distinct
 > >    negative for SNMPv3 as a choice, all other
 > >    things being equal. Any other wg considering 
 > >    SNMPv3 as the basis for their protocol would be
 > >    well advised to take that into consideration.
 > ...
 > 
 > Could you give an example of a protocol in use that does
 > initial key distribution, key update, authentication and access
 > control management well?  (All four, not just one or two.)
 > With SNMPv3 USM+VACM, we get three out of the four.  If there
 > were a good way to get all four, I think the WG would like
 > to know about it.

Irrelevant, but IKE,KINK/IPsec and TLS fit your
definition. This working group decided for its
own reasons that app layer auth/priv/authz was
required. As such, none of the usual suspects can
be used for key management for those SA. We are
left with a void. This manifestly causes trouble
with generally poorly conceived and reviewed
stopgaps.

Trying to say that the world is an imperfect place
therefore lets bury our head in the sand is a huge
cop out. In the context of MIDCOM, SNMPv3 should
be dismissed out of hand because it doesn't have
key management. It's that serious.

		Mike