- Should tcpinc support TCP simultaneous open? Yes or No
Yes.
I agree with other posters in noting that this should not
be done if it requires too much work - i.e., a longer
option or additional exchanges.
However, any mechanism intended to work in conjunction
with TCP SYN exchange ought to be able to support
simulatenous open as a true degenerate case without
such a burden - at best, it should require only
trivial tie-breaking.
If it requires more than that, IMO that means you
have a mechanism that isn't well matched to TCP, not
that the WG should "give up" on SO.
-- If "No," should TCP simultaneous open connection attempts
be abandoned or proceed unencrypted?
If SO fails for any reason, it should act exactly
like any other failure of TCO ENO.
Again, if it's treated as a special case when
failing, then that is a strong indication that
ENO isn't designed properly, IMO.
Joe
_______________________________________________
Tcpinc mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/tcpinc
