At 02:46 07.04.00 , Jamshid Mahdavi wrote:
>Section 2.3:
>
> > In low-speed, high error-rate environments (for example, the
> > wireless WAN environment), TCP windows are much smaller, and burst
> > errors must be much longer in duration in order to damage multiple
> > segments. Accordingly, the complexity of SACK may not be
> > justifiable, unless there is a high probability of both burst
> > errors and congestion.
>
>This logic doesn't make a lot of sense to me. I suppose that if your
>error rate is high, and you are constantly timing out, SACK is not
>going to make any difference. But if there is congestion too
>(i.e. the total error rate is even higher), SACK is going to help even
>less.
>
>I'm not an expert, but I would think that in wireless environments,
>changes in surroundings (i.e. someone walks in front of the wireless
>card) would cause a high probability of correlated errors. This would
>imply to me multiple losses per window, exactly the thing that SACK
>is good for. So if things mostly *don't* suck, SACK will help for the
>times when they do.
In http://iceberg.cs.berkeley.edu/papers/Ludwig-WINET99/ we present
measurements from a low-speed wireless WAN environment (9.6 Kb/s GSM
circuit-switched data) and we do find the mentioned correlated losses.
This would speak in favor of SACK which we believe is a win anyway (at
least it never hurts). However, we believe that persistent link layer ARQ
for the class of reliable flows (TCP, Reliable-Multicast, etc.) is an even
better solution. Still, SACK helps for correlated congestion losses and on
wireless links that do not (or cannot, e.g., unidirectional satellite)
implement link layer ARQ.
Also note, that (unlike stated in the above cited section 2.3) the fact
that an environment is low-speed has little influence on the TCP sender's
congestion window if router buffers are large enough. The advertized window
remains to be the limit and the congestion window will usually get there.
///Reiner
This archive was generated by hypermail 2b29 : Mon Jan 28 2002 - 09:12:24 EST