Re: Large scale transparent proxying

From: Simon Rainey <>
Date: Fri, 15 Oct 1999 07:29:14 +0100


We're suffering from poor throughput on loaded boxes when using the Intel
EtherExpress 10/100 under Linux. There seem to be other problems with that
NIC too, such as lockups in the presence of multicast traffic. I'm about to
replace all the Intel EtherExpress 10/100 NICs with a NIC based on the DEC
Tulip, which appears to perform well under Linux.


>> On Thu, 14 Oct 1999, Pete Yandell wrote:
>>> I'm running squid-2.2.STABLE4 under NetBSD 1.4.1 doing transparent
>>> proxying without caching, and with about 200 concurrent users (I don't
>>> have a figure for number of hits per second) the load gets really high
>>> (85% CPU usage or thereabouts) on a Pentium II 450 with 256Mb of RAM.
>>> A lot of that load seems to be system CPU time dealing with the
>>> required NAT translation.
>>> It seems to me that load shouldn't really be high...what it's doing is
>>> really pretty simple, especially given that it's not caching.
>> Just a thought, but what sort of network card are you using?
>Intel EtherExpress 10/100 (running at 100).

Simon Rainey Direct Line: 01235 823238
Principal Internet Consultant, Engineering Fax: 01235 823424
RM Internet for Learning E-mail:
New Mill House, 183 Milton Park, Abingdon, Oxfordshire, OX14 4SE, England
Received on Fri Oct 15 1999 - 00:37:20 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:48:55 MST