[Info-vax] Accuweather new contract

[Bill Gunshannon]

In article <55185cb3$0$1575$c3e8da3$e408f015 at news.astraweb.com>,
	JF Mezei <jfmezei.spamnot at vaxination.ca> writes:
> re: improving the IP stack.
> 
> Some newer OS have IP stacks that are able to handle TCP much more
> efficiently, not only offloading work to the ethernet card, but also
> sending ACKs for multiple packets instead of ACKs for each packet.

That was tried 30 years ago.  Believe it or not there was usually no
improvement in performance and in some cases it weas actually worse.
(As i have stated here before, I still have a couple of those cards
for PC's.  They were almost always used in "dumb" mode with the CPU
removed.  I also remember doing cards, EXCELAN I think, for the VAX.)

> 
> aka: if you have received sequence number 562 and 563, you can ACK 563
> and this implicitely says you have received all packets until 563, and
> this includes 562. For fast data transfers, this ends up greatly
> reducing ACK processing and traffic on the uplink.
 
That has always been the case with the TCP/IP spec.  I can't imagine
why any implementation would not have been taking advantage of this 
unless it was just plain broken. (Unless it did not allow any kind
of windowing, but then, considering that things like Kermit were
doing this decades ago I guess that would just be another sign of how
broken an implementation was.)



bill

-- 
Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolves
billg999 at cs.scranton.edu |  and a sheep voting on what's for dinner.
University of Scranton   |
Scranton, Pennsylvania   |         #include <std.disclaimer.h>