Re: [Tsvwg] F-RTO - open issues?

[Neil Spring <[e-mail]>]

On Wed, Oct 09, 2002 at 06:22:03PM +0300, Pasi Sarolahti wrote:
>> 3. Have you thought about protection against malicious TCP receivers?

> A receiver that sends malicious acks? I think any TCP sender is in
> problems in that case. 

Not all TCP senders have problems with malicious receivers,
only those that trust all acks.

 http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-tcp-nonce-03.txt

is an allocation of a bit in the TCP header to allow the
receiver to prove it has received only un-ECN-marked
original (not re-) transmissions.

That internet draft, whose status I've forgotten though I
thought it was on the path toward submission to the IESG as
a Proposed Standard, was a result of some simple analysis
in:

 http://www.cs.ucsd.edu/~savage/papers/ICNP01.pdf

Although the ICNP paper describes the ECN case specifically,
its predecessor:

 http://www.cs.ucsd.edu/~savage/papers/CCR99.pdf

describes a set of attacks a malicious receiver could employ
to subvert the sender's congestion control policy.  Not all
of these attacks work against robust senders.

enjoy,
-neil
_______________________________________________
tsvwg mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/tsvwg