Date: Mon, 8 Oct 2001 11:06:45 +0200 From: Barry Irwin <bvi@itouchlabs.com> To: =?iso-8859-1?Q?Rickard_Svor=E9n?= <rickard@xpress.se> Cc: freebsd-isp@freebsd.org Subject: Re: Dummynet & traffic-shaping Message-ID: <20011008110645.A26200@itouchlabs.com> In-Reply-To: <1002492561.3bc0d2916b700@webmail.xpress.se>; from rickard@xpress.se on Mon, Oct 08, 2001 at 12:09:21AM %2B0200 References: <1002492561.3bc0d2916b700@webmail.xpress.se>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon 2001-10-08 (00:09), Rickard Svorén wrote: > Hardware = Dual PPro 200Mhz, 128 RAM, 1 Intel EtherExpress NIC & 1 Realtek 8139 > > The catv customers should have 3.5 mbps and the rest should have 500kbps. > Have tested dummynet with my workstation and pico bsd, it worked well. > But what considerations should be taken regarding hardware and queue sizes > when over 300 users are flowing through the bridge? What should be considered > as "normal". Trial and error is an option of course but this would generate > a lot of angry phonecalls to me ;-) I would suggest dumping the Realtek, and getting an additional Ether Express 10/100 (fxp) or 3Com 10/100 (xl) card. Doing this should reduce the likely hood of loosing packets and the like. Realtek cards are nice for cheap and nasty solutions, but are nit really the best for high traffic loads. Barry To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-isp" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20011008110645.A26200>