From: David Newall Subject: Re: defrag deployment status (was Re: [PATCH] ext4: allow defrag (EXT4_IOC_MOVE_EXT) in 32bit compat mode) Date: Wed, 10 Mar 2010 02:49:41 +1030 Message-ID: <4B96751D.10804@davidnewall.com> References: <201003072132.10579.borntraeger@de.ibm.com> <4B94367E.9080506@garzik.org> <201003080853.42978.borntraeger@de.ibm.com> <4B9518DA.8010201@davidnewall.com> <4B952437.8020607@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: Christian Borntraeger , Jeff Garzik , linux-ext4@vger.kernel.org, linux-kernel@vger.kernel.org, Akira Fujita To: jim owens Return-path: In-Reply-To: <4B952437.8020607@gmail.com> Sender: linux-kernel-owner@vger.kernel.org List-Id: linux-ext4.vger.kernel.org jim owens wrote: > No. Your logic would be correct if rotating disks had > similar speed at all locations. Current disks are much > faster at the 0 end than at the middle or highest address. > I think my logic is still correct, although I wished I had said "closer to the middle." In fact, simplistic ideas for placement of files are unlikely to produce fabulous results (and that includes placing commonly used files towards the middle of the disk, say at the inside edge of the outermost zone.) The effort that BSD went to in FFS, placing directories with files and meta-data in cylinder groups, illustrates that disk performance is a sophisticated problem. Why don't we use BSD FFS/FFS2?