From owner-freebsd-net Tue Oct 26 16:57:45 1999 Delivered-To: freebsd-net@freebsd.org Received: from pail.ircache.net (pail.scd.ucar.edu [128.117.28.5]) by hub.freebsd.org (Postfix) with ESMTP id E61DC14FCB for ; Tue, 26 Oct 1999 16:57:43 -0700 (PDT) (envelope-from rousskov@ircache.net) Received: from localhost (rousskov@localhost) by pail.ircache.net (8.9.2/8.8.7) with ESMTP id RAA48753 for ; Tue, 26 Oct 1999 17:57:43 -0600 (MDT) (envelope-from rousskov@ircache.net) Date: Tue, 26 Oct 1999 17:57:43 -0600 (MDT) From: Alex Rousskov To: freebsd-net@freebsd.org Subject: TCP throughput vs number of aliases Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Hi there, We use FreeBSD 3.1 to benchmark Web caching proxies. We want to use a [very] large number of IP addresses for simulated clients and servers in the benchmark setup. However, we are limited by the number of machines we can dedicate to the tests. Naturally, we are using interface aliases to create IP addresses. With the number of aliases above 600+ we are monitoring a sudden and significant drop in TCP throughput (measured using simple netperf tests). The graph and some details of the setup are at http://www.ircache.net/~wessels/Junk/aliases/ We would like to use more than 600 aliases; probably 2,000-5,000 aliases per box without the loss of performance. Can anybody recommend a solution? Thank you, Alex. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message