[793] in cryptography@c2.net mail archive
Re: Full Strength Stronghold 2.0 Released Worldwide
daemon@ATHENA.MIT.EDU (Adam Shostack)
Sat May 10 14:04:01 1997
From: Adam Shostack <adam@homeport.org>
In-Reply-To: <33737230.41C6@netscape.com> from Tom Weinstein at "May 9, 97 11:51:28 am"
To: tomw@netscape.com (Tom Weinstein)
Date: Sat, 10 May 1997 11:17:23 -0400 (EDT)
Cc: adam@homeport.org, cryptography@c2.net
Tom Weinstein wrote:
| Adam Shostack wrote:
| > | Yeah. I get a lot of mail, much of it from mailing lists. Of
| > | course, almost none of it is encrypted today. It would be nice if,
| > | at some time in the future, all of it were encrypted.
| >
| > If its on a public list, you want authentication, not
| > encryption. Seems pretty silly to make glacier.mcom.com re-encrypt
| > each message to every member of the list.
|
| Why not? It makes traffic analysis much harder.
|
| In any case, you don't have to reencrypt the whole message, just the the
| bulk key. S/MIME handles multiple recipients by encrypting the message
| with a bulk encryption key, and then encrypting the bulk key with each
| recipient's RSA key.
Yes, you don't have to re-encrypt the whole message, just the
expensive parts of it. If you have a list with say, 1000 recipients,
then that is 1000 RSA ops per message. at 1/10 of a second per, thats
100 cpu seconds per message. That limits your list to about 860
messages per day, ignoring the cost of running a mail program. If you
take a mailing list like firewalls, with 10000 or so members, you're
talking about 86 messages per day maximum. Firewalls has gone over
that on a regular basis.
I don't see how it makes traffic analysis any harder on an
open list which anyone can subscribe to.
Adam
--
"It is seldom that liberty of any kind is lost all at once."
-Hume
