ca.portal.admin

Re: HPSPO query

Discussion created by ca.portal.admin on Feb 17, 2006
BT tested this when it was back in beta status.

Basically, HPSPO is a kind of halfway between no protection, and the old full-on storage protection.
You split up your storage pools so that no pool has both user and system storage. User storage is then put into a different
MVS storage key, so that if a user program tries to overwrite system-owned storage it will fail. HPSPO won't stop
a task overwriting another task's storage.

As far as performance goes, the overhead of having HPSPO on is negligible.


I
Iain Robertson British Telecom
Tel : 020 8456 7108
iain.dk.robertson@bt.com
pp2.11, Watford TE, Exchange Road, Watford, WD18 0GB
-----Original Message-----
From: IDMS Public Discussion Forum
[mailTo:IDMS-L@LISTSERV.IUASSN.COM]On
Behalf Of Boyce, Bill (GE Consumer Finance)
Sent: 17 February 2006 15:48
To: IDMS-L@LISTSERV.IUASSN.COM
Subject: [IDMS-L] HPSPO query


Back in the olden days there was two storage
protection options; on or off. Storage protection set to on
was costly from a storage management CPU consumption
standpoint. Then it was determined that running with storage
protection set to on at the system level and off at the
program level was around 30% more CPU effective than running
with storage protection off at the system level.

We have been running this way a long time (mid-eighties???).

I am preparing to upgrade all my production regions
from release 14.1 to 16.0 SP2 and I very confused as to where
the new High Performance Storage Protection Option fits into
the storage management CPU efficiency list:

Standard storage protection High CPU overhead
Storage protection off Moderate CPU consumption
system on/program off Low CPU overhead

In perusing through what documentation I can find,
plus prior post to this list group, I can find no definitive
answer. I only find conflicting statements:

The new option is the same as the 'system on/program
off' method.
A special form of storage protect is available for
the production system which provides negligible processing
overhead....
Performance testing has revealed that using this
option is virtually indistinguishable from running without
storage protect in terms of CPU cost.

Setting up benchmarks and testing out the options in
my current environment will be difficult, but I will do what
I can to get the answers I need. But before jumping into such
a project I thought it would be a good idea to query the
experiences of this IDMS list group.

Does HPSPO use more, less, or about the same as
storage protection off? Would the 'system on/program off'
method use less CPU than HPSPO? Does 'system on/program off'
still work as it did before HPSPO?

OPSYS Z/OS 1.4

TIA,
Bill Boyce

William M. Boyce
GE Consumer Finance - Americas
Sr. Systems Engineer
IDMS, Mainframe, and AS400 Technologies

T 203 944 6183
F 203 944 3509
C 203 650 7755
D *370-6183
E bill.boyce@ge.com
www.ge.com

6 Corporate Drive
7th Fl - 7101
Shelton, CT 06484-6206



>

"
IDMS Public Discussion Forum
IDMS-L@LISTSERV.IUASSN.COM
SMTP
IDMS-L@LISTSERV.IUASSN.COM
IDMS-L@LISTSERV.IUASSN.COM
SMTP








Normal

Normal
Re: HPSPO query
"An addition to what Pete has written. Originally, the overhead of storage
protection was approx. 30% for most shops. The horrendous increase up to
100+ % was due to a microcode change in the IBM z900 series hardware that
changed the instruction path length of the change key instruction. The
overhead we noticed was greater than 100% when we upgraded to the new
mainframe. Instruction path lengths can and do change in hardware microcode
as vendors develop the hardware to maximize performance for certain
software, so performance can be affected in unexpected ways when upgrading
to a new mainframe.

Linda Campbell
Informatix, Inc.

Outcomes