From owner-freebsd-isp  Thu Oct 23 20:13:11 1997
Return-Path: <owner-freebsd-isp>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.7/8.8.7) id UAA16535
          for isp-outgoing; Thu, 23 Oct 1997 20:13:11 -0700 (PDT)
          (envelope-from owner-freebsd-isp)
Received: from red.juniper.net (red.juniper.net [208.197.169.254])
          by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id UAA16529
          for <freebsd-isp@freebsd.org>; Thu, 23 Oct 1997 20:13:08 -0700 (PDT)
          (envelope-from tli@juniper.net)
Received: from chimp.juniper.net (chimp.juniper.net [208.197.169.196])
	by red.juniper.net (8.8.5/8.8.5) with ESMTP id UAA23657;
	Thu, 23 Oct 1997 20:12:36 -0700 (PDT)
Received: (from tli@localhost) by chimp.juniper.net (8.7.6/8.7.3) id UAA01661; Thu, 23 Oct 1997 20:12:35 -0700 (PDT)
To: dennis@etinc.com (dennis)
cc: freebsd-isp@freebsd.org
Subject: Re: Routing thru a FreeBSD?
References: <3.0.32.19971023201441.00ade850@etinc.com>
From: Tony Li <tli@juniper.net>
Date: 23 Oct 1997 20:12:35 -0700
In-Reply-To: dennis@etinc.com's message of 24 Oct 97 00:14:41 GMT
Message-ID: <82hga7j1lo.fsf@chimp.juniper.net>
Lines: 20
X-Mailer: Gnus v5.3/Emacs 19.34
Sender: owner-freebsd-isp@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

dennis@etinc.com (dennis) writes:

> >> Ob FreeBSD: The point here is that at some point, when you need many
> >> hundreds of thousands of PPS of forwarding, normal processors just
> >> fail to
> >> provide the necessary speed.  Note that for most situations, this is
> >> not
> >> necessary.  A heavily hacked FreeBSD system can get around 100Kpps.
> >> Of
> >> course at this point, you also run out of PCI bandwidth, so you've
> >> maxed
> >> out the rest of the hardware too.
> 
> Of course these numbers change linearly as Intel cranks out faster
> processors. 

That's not at all clear.  Again, exhausting the PCI bandwidth is very much
an issue that's not going to go away with a faster CPU.

Tony