Date: Mon, 29 Mar 1999 18:27:03 -0500 From: Craig Metz <cmetz@inner.net> To: alk@pobox.com Cc: freebsd-net@FreeBSD.ORG Subject: Re: FreeBSD as a router Message-ID: <199903292322.XAA11026@inner.net> In-Reply-To: Your message of "Mon, 29 Mar 1999 15:34:58 CST." <14079.61724.162248.667212@avalon.east>
next in thread | previous in thread | raw e-mail | index | archive | help
In message <14079.61724.162248.667212@avalon.east>, you write: >Since the discussion is occurring at this level, it is probably >helpful to note that the *size* of the packets also plays a large role >in determining maximum routing throughput: The box has to do a lot >more routing for 512B packets than it does for 1024B packets at the >same bandwidth. The box doesn't have to do "more routing" for a 512B packet than a 1024B packet, but, if you're filling the same number of bits/second, there are twice as many packets to move for 512B packets than for 1024B packets. (I believe this is what you meant, I just wanted to make sure people didn't get the wrong interpretation) Most router benchmarks like to talk about PPS as opposed to bits/second, and this is why -- most of the routing overhead is per-packet, not per-byte. >I'm guessing that a thorough search would show up some clock >vs. packet-size vs. bandwidth limit graphs for FreeBSD, Linux, >various commerical routers. I'd love to see well-done test data to substantiate or refute this sort of discussion; people know what the good and bad properties of the hardware and the software are and can take reasonably good guesses, but they're still just guesses and not measured performance numbers. I know that there is data out there, but I don't know how good it is. -Craig To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199903292322.XAA11026>