ietf-openpgp
[Top] [All Lists]

Re: Principles and Principals

1997-09-24 21:59:35
-----BEGIN PGP SIGNED MESSAGE-----

On Wed, 24 Sep 1997 16:47:15 +0100 Ian Brown 
<I(_dot_)Brown(_at_)cs(_dot_)ucl(_dot_)ac(_dot_)uk> wrote:

Sorry, I didn't answer this explicitly enough. I did not mean
that e-mail addresses should not be included in certificates. I
was merely agreeing with (I think) Charles Breed or Jon Callas
that a certificate shouldn't HAVE to include an address. This
would make it difficult for a system-wide distributed keyserver
scheme as you then need some other method to look up a key. An
e-mail only keyserver scheme is fine, however; keys which are
intended for use by e-mail can be certified, and looked up
using their address as a hint where to go for the lookup. I'm
not sure if there's enormous need for keyservers in areas not
related to the DNS.

Okay, I guess I was reading too much into what you said.  I
would also concur that an "e-mail address" should not be a
mandatory part of a certificate.  However, I think that the
converse IS a requirement.  To wit, if you DO use the key for an
e-mail address, then it needs to be bound to the key.  This is
probably true for identifying names or addresses in other
potential PGP applications for which I cannot speak
definitively.  Does anybody else have some useful examples to
contribute?

Which is what DNSSEC would allow; the domain of the e-mail
address would certify (not directly within the certificate, but
by providing the key as a trusted domain) that it at least
believes the stated key belongs to the stated address. Which is
a good start, and perhaps all that is necessary for the
majority of correspondence between strangers.

This, however, I think misses the point a little.  Even if we
leave hierarchical certification out of the discussion, when I
download a PGP key, I'd really like to be able to verify the
signatures of various vouching parties.  I single integrity
check from a faceless service provider (who with DNSSEC
would not be MY service provide) carries a lot less weight than
the signature from a key that I already trust.  Who knows what
their local identity or key verification procedures are?

This is not to say that that DNSSEC is valueless.  I think the
DNS could be used quite well for this.  However, there are
probably at least as many political barriers as there are
technical issues involved in using the DNS for this.  (Probably
more, in fact.)  A lot of folks automatically rebel at any
suggestion or a new use for the DNS; although I am not one of
these.  :-)

Also, you need to consider the situation where the DNS
maintainer has relatively little control over the e-mail
addresses in a particular domain.  My domain, for instance,
makes fairly heavy use of wildcard MX records.  This allows
our company a lot of flexibility without having to maintain
our own full time servers.  My service provider maintains these
DNS entries, but has no way to verify what e-mail addresses
we are publishing and using.  From their point of view,
they don't associate the address to the mailbox until a
message hit the queue.  I think this style of usage is on the
rise because of the increased number of virtual domain in use on
the Internet.

I hope this helps!

Chris



 ---------------------------------------------------------------
 |  International Electronic Communication Analysts, Inc.      |
 |  Christopher D. Bonatti                 9010 Edgepark Road  |
 |  Vice-president                     Vienna, Virginia 22182  |
 |  bonattic(_at_)ieca(_dot_)com   Tel: 301-212-9428   Fax: 703-506-8377  |
 |  PGP public key available from "http://www.ieca.com/";       |
 ---------------------------------------------------------------

-----BEGIN PGP SIGNATURE-----
Version: 2.6.3
Charset: noconv

iQCVAgUBNCmB4K8wqXqPmmVZAQGRegP/YfYNwNnOi1NAPFboheCnPbp9O1B9CEQ+
bN7LlbKHVDkEdSXbm8jk6vvIZN5y03S1I8UX/Ic8Dg22mwToXArzQv1aAoDiVho1
/iRws8irrkxVCp2QrxJFZubS6I2cqeSfQJC13w6TtqDiv04v+zG93gHRDcpGdDSg
FRwtQk3ep2I=
=A+Kt
-----END PGP SIGNATURE-----

<Prev in Thread] Current Thread [Next in Thread>