> From: George Michaelson <[email protected]>
>
> > end-to-end performance for small-object web transfers, and time to achieve
> > full window state on the network in general.
>
> "General release" users run at modem speeds;
...
> I think you'll find that even on modem lines, now that 28.8 and above is
> ubiquitous, the difference between time to get to a 2-3 window (not 1-2)
> from a start of 1, and time to achieve viable transfer if the initial window
> is 2-3, does make a difference.
Most current TCPs start with a window of 2 (the initial of 1 doubles
with the SYN/SYN-ACK exchange).
Going from 2 to 4 takes 1 additional RTT, unless there are other
problems (packet loss), in which case the larger window would have to
be used for restart, which is a whole other problem (see the related
I-D on restart).
> If you can shave off 1/3 of the round-trip cost per median object,
> isn't that worth it?
It's more like 15% (1/7) - you burn 2 RTTs in the open and
initial get. It takes another 3-4 'rtts' of transmission. That's
6, or 7 if the smaller initial window is used.
And both persistent connections and multiple connections defeat this.
> In any case, unless effect of an increase is provably *bad*, Wouldn't this
> be a possibly ok thing? Like for those pursaps on Hughes DirecPC where there
> is a direct-to-pc satpipe coming in..
How 'general' are such satelliet connections?
And do we now accept proposals "unless they're provably bad"?
(rough consensus and the absence of running, bad code?)
Joe
This archive was generated by hypermail 2b29 : Mon Feb 14 2000 - 16:14:42 EST