Re: [Tsvwg] Re: Last Call: Robust ECN Signaling with Nonces to Experimental

[Neil Spring <nspring@cs.washington.edu>]

On Wed, Jan 08, 2003 at 12:53:29PM -0500, Mark Allman wrote:
> We agree that eavesdroppers can exploit ECN and the ECN
> nonce as a mechanism for DoS.

Can either of you explain this statement a little bit more?

If an evesdropper is able to spoof an acknowledgement with a
faulty nonce, as described in Yogesh's prior mail, it would
seem just as capable of spoofing a RST to tear down the
connection in one packet (not probabilistically four).
While the nonce doesn't solve the problem of spoofing
evesdroppers, I don't yet see how the nonce increases the
potential for misbehavior in the network.

-neil