ADSM-L

ADSM and STK Iceberg dasd

1995-12-07 09:42:40
Subject: ADSM and STK Iceberg dasd
From: Raymond Raw <catmgrwr AT IBMMAIL DOT COM>
Date: Thu, 7 Dec 1995 09:42:40 EST
From - Raymond Raw at The Mutual Group
Technical Services, 519/888-3900 x3505
InterNet Address - CATMGRWR AT IBMMAIL DOT COM

CC: IBMMAIL /INTERNET
 SUBJECT:      ADSM and STK Iceberg dasd
 We currently have 4 ICEBERG systems totaling 840 gigs.  We're very
 happy with them.  We were part of the early beta program and have
 been running ICEBERG's for 2 years.  They perform pretty much as STK
 says although we've seen some inconsistencies with PAGE datasets on
 them.

 We've had ADSM databases on ICEBERG right from the start without any
 problems.  However, there is one major point of consideration when
 putting ADSM databases on ICEBERG.  (I suspect that RAMAC might have
 the same problem.)

 The ICEBERG size (200gig, 300gig etc.) is based on an assumption of 3
 to 1 data compression within the box.  The real round and brown
 backing a 200gig ICEBERG is only about 75gig in size.  (The 3 to 1
 ratio has actually changed with the newer boxes.)  At any rate, the 3
 to 1 compression ration is generally very conservative and for most
 data we achieve better compression ratio.  However, if you put
 precompressed ADSM data onto an ICEBERG, you don't achieve the 3 to 1
 ICEBERG compression and the box fills up faster than expected.  The
 result being that a 200gig ICEBERG might only store 150gigs of
 compressed ADSM data.

 In the last few months STK has moved to newer faster compression
 chips that do a better job of ADSM precompressed data so this problem
 might be insignificant now.  My suggestion would be to talk to STK
 specifically about ADSM data.  They are certainly aware of what we
 experienced.  Your STK rep can also run one of their field utilities
 against your 3390 databases and tell you exactly what compression
 rations you should achieve on the ICEBERG.
 *** END OF MESSAGE ***
<Prev in Thread] Current Thread [Next in Thread>