On Tue, Feb 19, 2008 at 06:19:39PM +0800, Kacheong Poon wrote:
> So I guess the issue is on how to grow the window size of each
> channel.  And only those channels which have used up their

Exactly.

> windows will need to have the windows adjusted.  I guess your
> scheme may work.  Just let the sender know the receiver's buffer
> size to avoid over subscription.  Using an appropriate fairness
> control and grow the bulk trasnfer channel's window.  This makes
> sure that there is always data in TCP to be sent and TCP will
> send as fast as it thinks it can.

Thanks.  I'll try to play with this sometime.

I'm thinking: start by setting SO_SND/RCVBUF to very large values and
then implement a dynamic window size scheme like I described.  I think
estimating available bandwidth and congestion at the SSHv2 layer may
well work, though there will be an aliasing problem -- I suspect that
a slow-start-like window growth scheme may keep the aliasing problem
under control, but if not then it's possible that we'll see a see-saw
pattern, but with very wild swings.  So this will take some
experimenting to get right.

Reply via email to