RE: I-D ACTION:draft-ietf-ipsec-isakmp-gss-auth-03.txt

Pyda Srisuresh <srisuresh@yahoo.com> Mon, 11 October 1999 21:39 UTC

Received: from lists.tislabs.com (portal.gw.tislabs.com [192.94.214.101]) by mail.imc.org (8.9.3/8.9.3) with ESMTP id OAA10569; Mon, 11 Oct 1999 14:39:33 -0700 (PDT)
Received: by lists.tislabs.com (8.9.1/8.9.1) id QAA14297 Mon, 11 Oct 1999 16:00:05 -0400 (EDT)
Message-ID: <19991011201148.18587.rocketmail@web1405.mail.yahoo.com>
Date: Mon, 11 Oct 1999 13:11:48 -0700
From: Pyda Srisuresh <srisuresh@yahoo.com>
Subject: RE: I-D ACTION:draft-ietf-ipsec-isakmp-gss-auth-03.txt
To: Markku Savela <msa@hemuli.tte.vtt.fi>, ipsec@lists.tislabs.com
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-ipsec@lists.tislabs.com
Precedence: bulk

Sounds like your approach should work. I believe, many vendors use this
type of an approach to get their initial IKE implementations (with 
certificate based authentication support) out the door.

cheers,
suresh

--- Markku Savela <msa@anise.tte.vtt.fi> wrote:
> > From:	Waters, Stephen [SMTP:Stephen.Waters@cabletron.com]
> >
> > One of the suggestions that came out (that I was fond of :) ), was you
> > provide clients with just their own private key, and the public key of the
> > server.  The server has its own public/private key, and access to a secure
> > database containing the clients private keys.
> 
> As I'm trying to specify a suitable test environment for demonstrating
> IPSEC (more than just PING over fixed end points), the above appears
> to be fairly close what I have in mind.
> 
> But, as I'm still a bit out of my area when public keys and ISAKMP is
> involved, there may be some catches that I am not aware of. And to
> verify, if I understood all correctly, I present my ideas and
> understanding the above here:
> 
>  - we have a server host with known fixed IP address (for example,
>    SMTP + POP + IMAP, perhaps IRC, WWW or whatever),
> 
>  - we want to allow client IPSEC device to connect this server from
>    random IP address (e.g. client calls any suitable ISP and uses
>    whatever dynamic address it gets)
> 
>  - only known users must be able to connect (eg. only company
>    employees, or people who have paid for the service)
> 
> Would this be achieved with the following setup
> 
>  - the organizaion runs own server specific CA setup
> 
>  - the server ISAKMP only accepts clients that have their keys signed
>    by the server CA directly (e.g. server ISAKMP is configured only to
>    accept single CA = self)
> 
>  - the access to server is given to client by having their public key
>    certified by the servers key (including policy changes for IPSEC on
>    the client). That is, client is given 3 pieces of information:
>    IPSEC policy changes, servers public key and clients
>    server-certified public key.
> 
> Is there ISAKMP negotiation mode that can establish a connection from
> client to server based on above information? As far as I can see, the
> clients IP address should not be needed (the server policy requires
> IPSEC for ALL connections by default, except of course having port 500
> clear!)
> 
> It would seem that certificate revocation would be fairly easy to
> manage as the control is all within organization. Revoked certificate
> list could be on the server and directly available to the
> ISAKMP. Thus, if a client is careless with the private key,
> certificates based on it can be quickly revoked.
> 
> -- 
> Markku Savela (msa@hemuli.tte.vtt.fi) Technical Research Centre of Finland
> Multimedia Systems, P.O.Box 1203,FIN-02044
> VTT,http://www.vtt.fi/tte/staff/msa/
> 


=====

__________________________________________________
Do You Yahoo!?
Bid and sell for free at http://auctions.yahoo.com