marcelo bagnulo braun <[email protected]> writes: > We would like to ask the WG to express their support to adopt one (or > none) of the following documents as WG document that will serve as a > basis for the protocol specification. Of course, the draft, if adopted, > will need to updated according to the WG input. In particular, they need > to be updated to not protect the TCP header. > > > The candidate drafts are: > > https://datatracker.ietf.org/doc/draft-bittau-tcpinc-tcpcrypt/ > https://datatracker.ietf.org/doc/draft-rescorla-tcpinc-tls-option/ > > We plan to discuss this on the meeting but it would be useful to start > the discussion before the meeting, so if you can express your opinions > before the meeting, it would be helpful.
I obviously support tcpcrypt. To meet the new requirements, we would make several modifications to the tcpcrypt draft. The most obvious is that authenticating the TCP header would become optional, and the draft would state that, barring a setsockopt or sysctl to the contrary, implementations SHOULD make unauthenticated TCP headers the default. To prevent FIN injection attacks from altering the TCP stream, we would add a payload-level EOF marker that MUST accompany any FIN segment. Likely this indicator would be a MAC of the FIN's 64-bit extended sequence number position, and would not be covered by the payload MAC so as to prevent it from being interpreted as data should the FIN bit be cleared. Other functions such as flow control would not be protected. In addition, we would modify some of the structures used in tcpcrypt negotiation to make it impossible for there to be downgrade attacks, where an attacker removes header authentication from a connection that actually wants it. We will also generalize this to make tcpcrypt's forward compatibility story more obvious. (Though we believe tcpcrypt in its current form is amenable to newer versions of the protocol, the existing draft does not give a good sense as to how the protocol should be evolved to avoid downgrade attacks.) David _______________________________________________ Tcpinc mailing list [email protected] https://www.ietf.org/mailman/listinfo/tcpinc
