Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!news.mel.connect.com.au!munnari.OZ.AU!spool.mu.edu!news.sol.net!uniserve!van-bc!news.cyberstore.ca!math.ohio-state.edu!howland.reston.ans.net!gatech!news-feed-1.peachnet.edu!paperboy.wellfleet.com!news3.near.net!sol.caps.maine.edu!web.ddp.state.me.us!gatekeeper.ddp.state.me.us!isdmill From: isdmill@gatekeeper.ddp.state.me.us (David Miller) Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: FreeBSD router, as good as a harware router ? Date: 8 Jan 1996 17:47:00 GMT Organization: Maine State Government Lines: 32 Message-ID: <4crlak$1h4@web.ddp.state.me.us> References: <4cof7j$59@news.mistral.co.uk> <4cpjil$k1l@agate.berkeley.edu> NNTP-Posting-Host: gatekeeper.ddp.state.me.us X-Newsreader: TIN [version 1.2 PL2] Jordan K. Hubbard (jkh@violet.berkeley.edu) wrote: : In article <4cof7j$59@news.mistral.co.uk>, Pete <plaker@cybar.co.uk> wrote: : >I've been convinced that FreeBSD would be a good OS for my pentium : >mail/web/news/ftp LAN server, but can FreeBSD on a seperate : >386 really be as reliable and more monitorable and configurable than : >the 'black box' option ? If so, this is much much cheeper, why : >doesn't EVERYBODY do this instead of spending a fortune on a hardware : >router ? : There's also throughput to consider. If I was routing 3-4 (10Mb) LANs : with a lot of inter-lan traffic to deal with, I'd probably get the : router also. A box that does nothing else but route packets is going to : blow the doors off an all-in-software solution that's capable of doing : everything from routing packets to running vi. 200-300K/sec is what you'll : typically get in lan-to-lan routing with a FreeBSD box, whereas 700-800K/sec : is more likely with a nice Cisco router. Just one datapoint here. I setup a bsdi PC (90 MHz pentium, lots of ram, 3com and WD cards) and got well over 800K/sec with some FTP's. Caveat - very simple setup, nearly empty routing tables, and no gated. Just thought I'd throw it in as a datapoint. I've been told that the indicated CPU utilization (3%) was simply not possible on an ISA bus, but PCI should do away with the problems soon enough:) -- David Miller Usual disclaimers apply