Date: Fri, 3 Jun 2005 13:53:06 -0700 From: Charlie Schluting <schluting@gmail.com> To: freebsd-net@freebsd.org Subject: Re: route metric Message-ID: <83946540050603135324d6b8cd@mail.gmail.com> In-Reply-To: <20050603202109.GA22098@gargantuan.com> References: <20050603181636.GA54906@gicco.homeip.net> <20050603191351.GA54164@ip.net.ua> <20050603202109.GA22098@gargantuan.com>
next in thread | previous in thread | raw e-mail | index | archive | help
> it would be nice to have a feature like this, where you could have > multiple same-prefix, same-metric routes in a FIB, and the packets would > be balanced to the next hop, either on a per-flow or per-packet basis. > i have seen a lot of answers to this request over the years along the > lines of ``FreeBSD isn't a router'', which is sad since it does perform > the task of packet routing exceedingly well, and a heck of a lot cheaper > than vendor C. all of the usual reasons that OSS is better apply here, > too. who wouldn't like SSH on all of their routers without paying $$$ > for a crypto image?!? >=20 It does do many things well enough, but have you tried to use dot1q on 5.x with an Intel chip? Those bugs are reason #1. You can't have a production router that reboot when you run tcpdump or traceroute :) Reason #2 is latency. Vendor C put a lot of time and money into features like CEF that take advantage of hardware packet forwarding. A purely software-based device simply can't keep up with large flows, and definitely introduces latency--especially when filtering. My $0.02 :) -Charlie
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?83946540050603135324d6b8cd>