Vulnerability Development mailing list archives

Re: spoofing the ethernet address


From: dimitrios.petropoulos () REUTERS COM (Dimitrios Petropoulos x9234 Singer / 4)
Date: Wed, 8 Mar 2000 10:28:41 +0000


One helpful measure owuld be to impliment the authentication in the
network hardware to reduce CPU usage. However, a real solution to this
class of attack is needed.


I think the point here is that verifying a MAC address might not be very
useful to a remote host who only knows its peer by IP address (MACs are
only useful in LAN but not in internet level). On the other hand, if MAC
*and* IP addresses are bound with a certificate then each time a host is
reconfigured with a different address (either IP because it's moved between
networks or MAC because a different NIC is installed) the certificate will
be invalidated.

Furthermore, authentication protocols involve real-time exchange of
messages which do not necessarily fit in with models where communication is
sometimes one way or connectionless. In challenge-response protocols a
verifier has to provide the challenge in order to be convinced of a
claimant's identity (otherwise there's no variant parameter in the message
and therefore messages from other protocol executions can be replayed). A
typical exchange would be

- A identifies itself to B
- B challenges A to prove its identity (e.g. sends random number)
- A responds to the challenge (e.g. sign random number with private key)
- B verify response (e.g. verify signature using A's public key)

(assuming a valid copy of A's public key is available to B)

The above does not always fit very well in situations where a communication
between parties consists of e.g. only one message or is of connectionless
nature. For established communications this is taken care of in higher
levels (network level or higher).

Regards,
Dimitrios

-----------------------------------------------------------------
        Visit our Internet site at http://www.reuters.com

Any views expressed in this message are those of  the  individual
sender,  except  where  the sender specifically states them to be
the views of Reuters Ltd.


Current thread: