home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!ukma!cs.widener.edu!dsinc!ub!niktow!pavlov
- From: pavlov@niktow.canisius.edu (Greg Pavlov)
- Newsgroups: comp.unix.ultrix
- Subject: Re: Recommendations for high disk throughput
- Message-ID: <2001@niktow.canisius.edu>
- Date: 28 Jan 93 10:32:54 GMT
- References: <C1BFFz.EEr@news.iastate.edu> <1993Jan27.205507.13234@grc.genroco.com> <1k7la9INNk6c@sol.tis.com>
- Organization: Canisius College, Buffalo NY. 14208
- Lines: 24
-
- In article <1k7la9INNk6c@sol.tis.com>, mjr@tis.com (Marcus J Ranum) writes:
- >
- > IPI's going to be a lot better
- > for database or other applications like imaging than it will
- > be for NFS. .....
-
- Imaging, maybe, but I see this statement made a lot re databases and I
- don't understand it. "Database" does not equate with large sequential
- reads; in some cases it does, in some cases it doesn't, there just ain't
- no generalization since databases span much of the applications spectrum.
-
- Actually, I would say that a lot of people spend a lot of time to cleverly
- construct databases and software to reduce sequential data traversal to a
- minimum, or eliminate it altogether. In the situations where this can be
- done (and I would guess that in the majority of applications, it is the
- correct and most useful strategy), and where there is a lot of user activity,
- I would want to spread the data out among as many controllers and under as
- many read/write heads as I could - to the extent that the cpu and systems
- software can handle it. Cheaper, smaller disks, but more of them, would
- then seem to be the more appropriate choice.
-
-
- greg pavlov
- pavlov@fstrf.org
-