1996-08-20 - emscrypt and replay attacks

Header Data

From: Matthew Carpenter <mcarpent@mailhost.tcs.tulane.edu>
To: campbell@c2.org
Message Hash: 6258aa8457053cdf9fc555d6c35663e2eb5fb71579bf51696ccf514446ef3c50
Message ID: <199608192023.PAA20241@rs6.tcs.tulane.edu>
Reply To: <9608191513.AA11169@cfdevx1.lehman.com>
UTC Datetime: 1996-08-20 00:21:58 UTC
Raw Date: Tue, 20 Aug 1996 08:21:58 +0800

Raw message

From: Matthew Carpenter <mcarpent@mailhost.tcs.tulane.edu>
Date: Tue, 20 Aug 1996 08:21:58 +0800
To: campbell@c2.org
Subject: emscrypt and replay attacks
In-Reply-To: <9608191513.AA11169@cfdevx1.lehman.com>
Message-ID: <199608192023.PAA20241@rs6.tcs.tulane.edu>
MIME-Version: 1.0
Content-Type: text/plain


Rick Campbell writes:

>>     P.S. I have an alpha version of a program which may be of interest to
>>     technomads: it automatically executes scripts received by email from a
>>     remote machine and then mails back the results.  The scripts (shell
      ... 
> 
> Does your mechanism do anything to prevent replay attacks?
> 
> 			Rick

Alan apparrently forwarded my message from technomads to cypherpunks,
but since I'm on cypherpunks too, I got this message.  Anyway, yes it
does have a simple replay attack prevention mechanism.  It keeps track
of the most recent time and date stamp from the PGP signature info and
refuses to executed any message that doesn't have a stamp more recent
than previously executed script.  This simple mechanism can cause
unwanted rejection if scripts are received out of order, but multiple
scripts can be batched into a single message to help overcome this.

See the following URL for a discussion of known limitations and security
concerns with emscrypt:

  http://www.bmen.tulane.edu/~carpente/emscrypt/emscrypt_doc.html#limits


--Matt

--
mcarpent@mailhost.tcs.tulane.edu    PGP mail preferred, finger for public key.





Thread