Hi, everyone. We've released a new draft of TCP-ENO, available in the
usual place:
https://datatracker.ietf.org/doc/draft-bittau-tcpinc-tcpeno/
In this draft, we've tried to address all of the feedback we received on
the list. Some suggestions we incorporated directly. In particular:
* The requirement for 128-bit security was softened to prohibit
"substantially less than 128-bit security."
* The forward secrecy requirement was rephrased to require "forward
secrecy some bounded, short time after the close of a TCP
connection."
* One of the application-aware settings is now reserved, to avoid
confusion over two different values.
* Language was added stating that applications SHOULD treat the session
ID as a monolithic opaque value. Other bytes have to be
computationally indistinguishable from random.
* The description of simultaneous open was clarified/corrected to
retransmit the ENO option in case one of the original SYNs is lost.
On other issues, where we couldn't determine consensus on the list, we
added discussion to an Open Issues section, to ensure that we do not
lose track of these issues. In particular, there are now subsections
on:
* Simultaneous open design points.
* Whether there should be multiple and secret session IDs
* Whether each suboption should have a length byte
I hope this addresses or at least acknowledges most of the major issues
people have raised on the list.
David
_______________________________________________
Tcpinc mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/tcpinc
