Re: [quicwg/base-drafts] Don't amplify 0-RTT (#1616)

[Martin Thomson <[e-mail]>]

martinthomson commented on this pull request.



> @@ -1896,6 +1896,13 @@ dynamically calculate the expiration time.  It is also unlikely that the client
 port number is the same on two different connections; validating the port is
 therefore unlikely to be successful.
 
+Even when a token is valid, a server SHOULD limit the amount of data that it
+sends toward a client in response to 0-RTT data.  Spoofing of source addresses
+and replays of 0-RTT handshake messages can cause tokens to appear to be valid
+when they are not.  No specific recommendation is made regarding the ratio of
+data sent to received, though it is unlikely to be necessary to send more than
+the initial congestion window.

Yeah, this is why it is worded so loosely.  Without this advice, the risk seems to be limited to the server receiving a spoofed ACK for the ServerHello and sending that much more in response, but that's only one more packet than the IW, so it's hardly earth-shattering.

Given that this is largely dictated by congestion controller design (IW specifically), is this really a change we need to implement?  Should we instead close #1309 with no change?

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/quicwg/base-drafts/pull/1616#discussion_r207073216