home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!uknet!root44!root.co.uk!numb
- From: numb@root.co.uk (Matthew Newman)
- Newsgroups: comp.protocols.nfs
- Subject: Re: NFS I/O Ops/seconds
- Message-ID: <numb.711969731@root.co.uk>
- Date: 24 Jul 92 09:22:11 GMT
- References: <1992Jul22.061146.15641@u.washington.edu> <l6r4uvINNf0p@exodus.Eng.Sun.COM> <nkok0p0@rhyolite.wpd.sgi.com>
- Sender: news@root.co.uk (News System)
- Organization: UniSoft Ltd., London, England
- Lines: 24
- X-Address: Spa House, Chapel Pl, Rivington St, London EC2A 3DQ
- X-Organisation: UniSoft Ltd
- X-Quote: To imagine is everything. - Albert Einstein
- Nntp-Posting-Host: dylan
- X-Job: Systems Engineer
- X-Phone: +44 71 729 3773 x215, Fax: +44 71 729 3273
-
- In <nkok0p0@rhyolite.wpd.sgi.com> vjs@rhyolite.wpd.sgi.com (Vernon Schryver) writes:
- >In article <l6r4uvINNf0p@exodus.Eng.Sun.COM>, beepy@tabitha.Eng.Sun.COM (Brian Pawlowski) writes:
- >> The number "300 NFS ops/s per ether" is a rule-of-thumb
- >Yes, everyone keeps saying, for ethernet.
- >One can hope to do significantly better on FDDI.
- >Imagine a multi-homed FDDI NFS server (>1 FDDI link), with FDDI-to-ether
- >routers or bridges that can run at ethernet speeds.
- This is misleading, there are two different issues here:-
-
- 1) Number of NFS ops to saturate the Ethernet/FDDI
-
- 2) Max number of NFS ops a given client/server arrangement
- can achieve on a given network.
-
- In some fairly detailed tests on our FDDI network and our departmental
- Ether, we found little increase in NFS ops over FDDI as the
- clients/server/disks were the bottleneck, not the network.
-
- FDDI has a higher bandwidth and therefore can support more simultaneous
- NFS servers at full speed, but in practice each server/client is
- not any faster.
- --
-
- Matt Newman
-