Re: 'Fast' readout of sequential data stored on hard disks.

From: Drew Kramer <drew_at_ny.ox.com>
Date: Fri, 27 Jan 95 10:54:53 -0500

I would use one of two options to process such a large data set.

1) a TLZ87 - a native 20GB tape drive (compression will get you more) with a
   data rate of 1.25MB/sec

2) Use LSM and/or AdvFS to set up a "stripe" set. This will spread the data
   evenly and equally between two or more drives. You should be able to
   sustain the maxiumum SCSI throughput of about 10MB/sec

Of the two choices above, the first is simple, and would run you about $5,000.
The second, could run a little more than that, depending on how many drives
you use, and you would want the AdvFS Utilities license. You do get much more
bang for your buck on choice two. There has been a bunch of discusion
regarding LSM and AdvFS on this list, so see the archives for more info. Or
you can just ask me, I'm out of the office right now, but if you want/need
more info on them, I can point you in the right places. Good luck.

-drew

-------
USMail: OTA Limited Partnership E-mail: drew_at_ox.com
        1 Manhattanville Road Phone: +1 914 694 5800
        Purchase, New York 10577 FAX: +1 914 694 5831
                       "Welcome, to the machine..."
Received on Fri Jan 27 1995 - 10:55:48 NZDT

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:45 NZDT