Date: Tue, 4 Apr 1995 11:49:22 -0700 (PDT) From: Poul-Henning Kamp <phk@ref.tfs.com> To: rgrimes@gndrsh.aac.dev.com (Rodney W. Grimes) Cc: wilko@yedi.iaf.nl, peter@bonkers.taronga.com, terry@cs.weber.edu, PVinci@ix.netcom.com, hackers@FreeBSD.org Subject: Re: large filesystems/multiple disks [RAID] Message-ID: <199504041849.LAA03075@ref.tfs.com> In-Reply-To: <199504041720.KAA07847@gndrsh.aac.dev.com> from "Rodney W. Grimes" at Apr 4, 95 10:20:52 am
next in thread | previous in thread | raw e-mail | index | archive | help
> > > RAID does have the negative effect of of having to write 20% more data, > > > thus cutting effective bandwidth by 20%. It is actually worse than > > > this in that all writes must write to at least 2 drives no matter how > > > small they are. The removes some of the benifits of stripping. > > > > And that is why some RAID systems use (battery backed up please ;-) RAM > > caches. This works quite nicely. > > And you find these caches will fill up and some point in a sustained > write test and you end up right back at the 20% performance loss I > was talking about. In practice it is much more that 20%. I saw numbers in the 50% range for Veritas' RAID5 implementation, which is a pretty good one. > Pure stripping of drives always outperforms RAID, you always pay some > price for reliability, and it is usually performance or $$$. Actually, you can consider mirroring a good combination. I have seen an almost 80% speedup on a "best effort mirroring" implementation. > Humm.. wonder why no one is doing parrallel transfer SCSI drives, seems > it would be easy to use 2 of the heads at once in a drive to effectivly > double the drive electronics to media transfer rate. This could all > be hidden in the drive electronics giving us disk drives that would > be just as reliable and have 2x the transfer rate they currently do. IBM does. They have a 5.25" device which is really two 3.5" responding as one LU. -- Poul-Henning Kamp <phk@login.dknet.dk> -- TRW Financial Systems, Inc. 'All relevant people are pertinent' && 'All rude people are impertinent' => 'no rude people are relevant'
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199504041849.LAA03075>