Re: SSL and certificates

To: Michael Brennen <mbrennen@fni.com">mbrennen@fni.com>, Jon Tegethoff <jet@cypher-sage.com>
Subject: Re: SSL and certificates
From: "Bob Denny" <rdenny@dc3.com>
Date: Thu, 29 Aug 1996 14:38:19 -0700
Cc: "'Www-Security@ns2.rutgers.edu'" <Www-Security@ns2.rutgers.edu>
In-Reply-To: Michael Brennen <mbrennen@fni.com> "RE: SSL and certificates" (Aug 29, 10:26)
References: <Pine.LNX.3.95.960829100503.27318C-100000@ns1.fni.com>

This list is not replyable (list address is in cc:), so I think I posted my 
last message directly back to the author. I gotta watch that here...

> > I believe that this is the major reason for not using PGP for this type
> > of application.  The trust model is not solid enough.  Trusted CAs are
> > required! 

The hierarchical trust model is a degenerate case of PGP's more general "web 
of trust" model. By agreement, PGP can be run in a hierarchical trust mode 
without any changes to the software. Just agree to trust only "CAs" to sign 
keys.

> Overall I agree, and the CA infrastructure is being worked out. 
> [...] 
> As in most secure contexts, it becomes a matter of risk assessment and a
> careful examination of what I base my trust on. 

Exactly. And it is your assesment of risk that should guide your selection of 
who you trust to sign keys for PGP. For some reason, this is not well 
understood, yet it is so simple.

  -- Bob

References:

RE: SSL and certificates

From: Michael Brennen <mbrennen@fni.com>

From: