I assume that most people saw the announcement that the new standard
mechanisms draft is available from the I-D archive
(draft-ietf-tcpsat-stand-mech-04.txt if your playing along at
home...). In this version, we have made an attempt to address all
the issues raised on the list and in the meeting in LA. In
particular, we have reworked the FEC section a bit. The section on
SACKs was also heavily editted. And, we added something resembling
a real security section. Finally, we made a number of somewhat
minor tweaks to the document...
-added pointer to the PSC TCP implementations web page
-added a little about link level protocols (how they are
important, but basically beyond the scope of the draft) (we
already have such statements about other mechanisms (e.g.,
application protocols) so this seemed like a good idea)
-made some minor changes to the Path MTU section, per comments
given at the LA meeting (notably that the people in charge of
the link should pick the appropriate MTU size and TCP should
use that MTU whenever possible).
-added a note that socket buffers must currently be hand tuned
(and also mentioned autotuning as a potential solution to this)
-a bit of general cleanup
-possibly some other small stuff that has slipped my mind...
As always, comments are very much appreciated.
allman
--- http://gigahertz.lerc.nasa.gov/~mallman/
This archive was generated by hypermail 2b29 : Mon Feb 14 2000 - 16:14:42 EST