David S. Miller wrote:
From: Bill Davidsen <davidsen@xxxxxxx>
Subject: Throughput problems on high latency link
Date: Wed, 22 Jun 2005 13:37:40 -0400
I increased memory sizes in net/core/*mem* and rmem-* wmem_* in ipv4, as
well as changing the thp_adv_win_scale. Looking at the traces, it
appears that the number of packets in flight are increasing, but the
throughput is still low.
Those *mem* sysctl values under net/core/ aren't going to
modify TCP behavior at all. You need to change the ones
named net/ipv4/tcp_*mem. They are documented in the file
named "Documentation/net/ip-sysctl.txt" in the kernel sources
and probably elsewhere.
Yes, thanks, had done that as well, and essentially every other thing
google and books on Linux networking could find, including playing with
the window size in the routing table. Still getting very poor
throughput. I can provide more bandwidth, but not less latency, that
speed of light thing gets in the way crossing many timezones and routers.
Thanks for the input anyway.
--
bill davidsen <davidsen@xxxxxxx>
CTO TMR Associates, Inc
Doing interesting things with small computers since 1979
-
: send the line "unsubscribe linux-net" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html