Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!munnari.OZ.AU!news.mel.connect.com.au!news.mel.aone.net.au!grumpy.fl.net.au!news.webspan.net!news1.exit109.com!en.com!lawrenceville.mccc.edu!news-xfer.netaxs.com!tezcat!gail.ripco.com!nntp.neu.edu!not-for-mail
From: mkagalen@lynx.dac.neu.edu (Michael Kagalenko)
Newsgroups: comp.os.linux.misc,comp.os.linux.networking,comp.os.linux.setup,comp.unix.bsd.bsdi.misc,comp.unix.bsd.misc
Subject: Re: Linux vs BSD
Date: 21 Jan 1997 18:41:12 -0500
Organization: Northeastern University, Boston, MA. 02115, USA
Lines: 55
Message-ID: <5c3k6o$qro@lynx.dac.neu.edu>
References: <32DFFEAB.7704@usa.net> <5c155c$p6u@raven.eva.net> <5c19pg$rf6@lynx.dac.neu.edu> <5c341j$3dp@cynic.portal.ca>
NNTP-Posting-Host: lynx.dac.neu.edu
X-No-Archive: yes
Xref: euryale.cc.adfa.oz.au comp.os.linux.misc:153334 comp.os.linux.networking:65572 comp.os.linux.setup:93098 comp.unix.bsd.bsdi.misc:5654 comp.unix.bsd.misc:1952
Curt Sampson (cjs@cynic.portal.ca) wrote:
]In article <5c19pg$rf6@lynx.dac.neu.edu>,
]Michael Kagalenko <mkagalen@lynx.dac.neu.edu> wrote:
]>J.C. Archambeau (jca@bighorn.accessnv.com) wrote:
]>]preferrential to BSD. It also seems the FreeBSD is more solid that
]>]Linux in the networking department.
]>
]> I see this claim now and then. Can you post some specific data, comparing
]> Linux and BSD networking ?
]
]I see this claim regularly to, and I recently tried to investigate
]it. Unfortunately, I have found only one Linux user who is willing
]to give me any statistics on NFS performance, and his network and
]server were heavily loaded.
That's right, there is a conspiracy amongst Linux users to deny you
the data.
]Under NetBSD on 486-class machines (Alpha Multia, Sparc IPX), I
]get 300-400K/sec writing NFS v3 and 920 K/sec reading. I expect
]that the writing performance would improve with faster servers and
]clients; I do not think the read performance will improve at all,
]since even on significantly slower machines (Sparcstation 1) I get
]similar read performance (though worse write performance). I hope
]to do some testing with more powerful machines soon. My current
]results of disk and NFS testing with bonnie are available in
]<http://www2.portal.ca/~cjs/computer/benchmark/>
]
]Under Linux, last time I had direct experience with it I saw
]50-100K/sec read and write performance under NFS. The person I
]talked to more recently, with the heavily loaded network and servers,
]saw 150-200K/sec performance for read and write. If anyone out
]there has better figures (preferabily bonnie output, along with
]descriptions of the machines and whether this is NFS v2 or NFS v3),
]I would love to see that information.
Clearly, you have no idea what you are talking about. Disk performance
benchmarks are limited by the narrowest bottleneck in the system.
Linux system in question is most likely to use IDE drives, whereas
workstations (to which you misleadingly refer as "486-class machines";
no workstation is "486-class machine", even if CPU speeds are comparable)
use SCSI-drives (and high-performance buses). I encourage you to read
the documentation that comes with benchmark that you are using.
To make meaningful comparison, you have to run your benchmarks on the
same hardware using different OSes. Which you won't have any trouble to do,
if you want the data, since Linux runs on Sun hardware. Otherwise, you
are blathering.
--
ABILITY,n. The natural equipment to accomplish some small part of the meaner
ambitions distinguishing able men from dead ones.
-- Ambrose Bierce, "The Devil's Dictionary"