Xref: utzoo comp.protocols.nfs:1882 comp.arch:21254
Path: utzoo!news-server.csri.toronto.edu!cs.utexas.edu!uunet!kddlab!titcca!cc.titech.ac.jp!necom830!mohta
From: mohta@necom830.cc.titech.ac.jp (Masataka Ohta)
Newsgroups: comp.protocols.nfs,comp.arch
Subject: Re: The merits of FCFS for file servers
Message-ID: <7348@titcce.cc.titech.ac.jp>
Date: 6 Mar 91 06:35:25 GMT
References: <28975@cs.yale.edu> <PCG.91Feb28195347@odin.cs.aber.ac.uk> <476@appserv.Eng.Sun.COM> <PCG.91Mar4210505@aberdb.cs.aber.ac.uk>
Sender: news@cc.titech.ac.jp
Organization: Tokyo Institute of Technology
Lines: 22

In article <PCG.91Mar4210505@aberdb.cs.aber.ac.uk>
	pcg@cs.aber.ac.uk (Piercarlo Antonio Grandi) writes:

>Some argue that having a
>segment length of 1 (FCFS) is not that bad, actually, depending on
>conditions. I was myself surprised by this, but on second thoughts it is
>not that incredible; if the filesystem is careful in laying out blocks
>in a physically clustered way

I'm afraid you are correct only 10% of the time.

What if there are two uncorrelated processes writing something on
different locations of a disk?

As you are talking about a network file server, such a situation is
very common.

						Masataka Ohta

PS

disk partitioning makes the situation even worse.