We have 10+ people at a time on our boxes running PPP. Not however with
1.3.72.
> that just keeps sending data (like rcp local_machine:/sdmach /tmp),
> netscape download in another window slows down to 30 bytes per second.
> I can't even use telnet at this speed :)
One problem with rcp is that it as standard uses rsh and proceeds to
blast volumes of data over your link at "interactive" - ie high priority.
Netscape properly uses a lower priority and gets clobbered. It should
use a low priority as ftp does.
> Looks like if a remote host is sending data too fast, Linux will do nothing
> to slow it down. TCP/IP standard includes "Source Quench" messages for this
> purpose, why doesn't UNIX send and recieve them? Also, if Linux doesn't
> acknowledge packets that are coming too fast, won't the other side assume
> that network is conjested and slow down?
Source Quench is almost never used. The problem with source quench is that
having the answer to "network clogged up" as "send more packets" is not
very good. TCP will deduce the round trip time from the acks properly, as
well as the pipe size (ie number of packets in flight). Nothing protects
you from the fundamental problems with modems
o Lack of bandwidth
o High latency
o Large internal queues.
I would be interested if you can repeat your tests with 1.2.13 and maybe
1.3.32 and see if the performance has changed to verify its what I expect
and not a recently arrived bug making a bad situation worse.
> Another solution would be to add priorities to traffic from selected ports,
> like term. When I used it, I was able to open an ftp session and a Mosaic
> window and trsh was still usable?
You can already do that. Al Longyear added the support a while ago. Get
ipfwadm and read the documents.
Alan