Re: is manual keying mandatory

EKR <ekr@terisa.com> Mon, 23 March 1998 22:39 UTC

Received: (from majordom@localhost) by portal.ex.tis.com (8.8.2/8.8.2) id RAA26635 for ipsec-outgoing; Mon, 23 Mar 1998 17:39:20 -0500 (EST)
To: bkavsan@ire-ma.com
Cc: Dave Carrel <carrel@ipsec.org>, Steve Sneddon <sned@cisco.com>, "Theodore Y. Ts'o" <tytso@MIT.EDU>, ipsec@tis.com
Subject: Re: is manual keying mandatory
References: <199803232019.MAA28635@weenie.redbacknetworks.com> <3516DD14.D23C6BDE@ire-ma.com>
From: EKR <ekr@terisa.com>
Date: Mon, 23 Mar 1998 14:54:08 -0800
In-Reply-To: Bronislav Kavsan's message of "Mon, 23 Mar 1998 17:07:16 -0500"
Message-Id: <3hg4pqar3.fsf@kmac.terisa.com>
Lines: 23
X-Mailer: Gnus v5.5/XEmacs 20.4 - "Emerald"
Sender: owner-ipsec@ex.tis.com
Precedence: bulk

Bronislav Kavsan <bkavsan@ire-ma.com> writes:
> Here are the reasons:
> - there is no "standard" key distribution mechanism for symmetric keys (I guess I
> can get on the phone with another guy and negotiate key values)
> - there is no "standard" mechanism for negotiation key lifetimes (should I also
> use the phone?)
> - how to re-key? - (get on the phone again?)
> - what is the encapsulation context - tunnel/transport? (my phone bill is getting
> higher?)
> etc, etc, etc.
These are operational reasons why it's inconvenient for USERS to do
manual keying. They don't have anything to do with why implementers
can't do manual keying, which is the question at hand.

What, precisely, is so incredibly difficult about adding this to
one's implementation that people are willing to make a big deal over
this, instead of just letting the settled issue stay settled?

-Ekr

-- 
[Eric Rescorla                             Terisa Systems, Inc.]
		"Put it in the top slot."