× The internal search function is temporarily non-functional. The current search engine is no longer viable and we are researching alternatives.
As a stop gap measure, we are using Google's custom search engine service.
If you know of an easy to use, open source, search engine ... please contact support@midrange.com.


  1.  Home
  2. MIDRANGE-L
  3. November 2017

Re: SAN and Performance

On Sat, 2017-11-25 at 16:20 +0000, Steinmetz, Paul wrote:
Larry,

I wouldn't use SETOBJACC, too many objects to deal with.
I was only referring to it as an example.

Perhaps IBM will come out with something similar to SETOBJACC that would keep objects in main storage when ample memory is available.

They already have the engine to identify the objects with STRASPBAL - Media Preference (MP) balancing.
Would IBM consider enhancing the engine to keep "HOT" objects in main storage?

I guess you could say it's partially being done with paging. If enough memory is available, the pages stay in memory.

I remember years back, before SSD, and when we had a serious disk I/O problem. 40% of our I/O was the result of page faults.
We doubled our memory, majority of the faults went away.

From the old days, IBM always said, more memory the better.

Paul

I remember when one of the overriding ideas behind the "400" was the
idea that new storage ideas would be transparent to the
user/programmer... the thinking was for when static/persistent memory
became available nothing would need to be done because that would be all
handled transparently by the slic...

Guess what, that day is nearly upon us:
https://software.intel.com/persistent-memory

Its only taken 29 years, but the reality of persistent memory might
finally be realised.

Cost wise, from the vids, it seems that its closer to ram prices than
flash SSD's, although as someone who is more used to PC hardware prices
I am always astonished at how expensive i/power hardware (disk) is in
comparison to desktop stuff, but given that this is the kind of thing
that will be brought by companies that price differential might not be
such an issue especially where a company might be looking at sticking
half a TiB's of memory in a box, or more (especially as AMD's EPYC, a
"pc chip", can handle 2TiB of ram per chip package - an incomprehensible
amount only a few years back).

Speed wise, its closer to ram than SSD especially latency and its
ability to be accessed bytewise instead of blocks is a huge advantage
for things such as psudo "in memory" DB's. Although because it doesn't
have the inbuilt error checking/handling and stuff that "disk"
subsystems have, that all needs to be done at the software level. But
that is where the i has the advantage, because all the lifting is done
by the "os" so its IBM's problem and all they have to do is have a *SSM
(Static Storage Memory) option where applicable to go with *SSD ;-)

Any ways, I think I may have digressed somewhat.

Jon




-----Original Message-----
From: DrFranken [mailto:midrange@xxxxxxxxxxxx]
Sent: Friday, November 24, 2017 10:53 PM
To: Midrange Systems Technical Discussion; Steinmetz, Paul; 'Gad Miron'
Subject: Re: SAN and Performance

I have had sum success with SETOBJACC in specific environments. Thing is that if a particular object truly gets bashed thoroughly and enough memory exists the object will likely be all or mostly all in memory already. If enough memory does not exist then pulling the object into memory will hurt other processes.

Bottom line is to know what you are doing! :-)

- Larry "DrFranken" Bolhuis

www.Frankeni.com
www.iDevCloud.com - Personal Development IBM i timeshare service.
www.iInTheCloud.com - Commercial IBM i Cloud Hosting.

On 11/24/2017 8:13 PM, Steinmetz, Paul wrote:
I had an interesting discussion with the MPG folks recently.
Question was
Can performance be improved over 100% SSD?
The answer is yes.
Put everything in memory, which is becoming more of a reality as memory prices keep coming down.
It's an interesting concept.
Makes sense.
Remember years ago, before SSD drives were available, if you needed to boost I/O performance, SETOBJACC.

The Set Object Access (SETOBJACC) command temporarily changes the
speed of access to an object by bringing the object into a main
storage pool or purging it from all main storage pools. An object can
be kept main storage resident by selecting a pool for the object that
has available space and does not have jobs associated with it.
Repeated use of the command can cause a set of objects to be resident
in a main storage pool.

The I/O would only be from the CPU to memory, virtually eliminating the I/O.

Any thoughts from the group?

Paul

-----Original Message-----
From: MIDRANGE-L [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of
Steinmetz, Paul
Sent: Friday, November 24, 2017 3:40 PM
To: 'Midrange Systems Technical Discussion'; 'Gad Miron'
Subject: RE: SAN and Performance

From the performance folks at MPG.

Strictly from a performance standpoint, internal SSD using the top PCIe3 RAID Adapter will absolutely always outperform a SAN.

Paul

-----Original Message-----
From: MIDRANGE-L [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of
Steinmetz, Paul
Sent: Friday, November 24, 2017 9:15 AM
To: 'Midrange Systems Technical Discussion'; 'Gad Miron'
Subject: RE: SAN and Performance

1) How do determine what model SAN is needed?
https://www.ibm.com/storage/disk

2) Do all SANs offer the same performance?

3) How do you determine which attachment method should be used.
https://www-03.ibm.com/support/techdocs/atsmastr.nsf/5cb5ed706d254a818
6256c71006d2e0a/fe962d4c9cefdcfc8625785600479c87/$FILE/iExternalStorag
eMatrix_v170301.pdf

3) The SAN can be direct attached without using VIOS, correct?

4) Strictly from a performance standpoint, can a SAN outperform internal SSD using the top PCIe3 RAID Adapter?

Our P7 8205-E6C is performing well with 100% SSD drives.
EOL was announced, so planning our next upgrade.
Power9 maybe a game changer.

Paul


-----Original Message-----
From: MIDRANGE-L [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of
DrFranken
Sent: Thursday, November 23, 2017 8:52 PM
To: Midrange Systems Technical Discussion; Gad Miron
Subject: Re: SAN and Performance

See Jim's response. The answer of course is "It Depends" but of course properly designed it will match or exceed internal disk performance.
Poorly designed, notsomuch.

- Larry "DrFranken" Bolhuis

www.Frankeni.com
www.iDevCloud.com - Personal Development IBM i timeshare service.
www.iInTheCloud.com - Commercial IBM i Cloud Hosting.

On 11/22/2017 11:56 PM, Gad Miron wrote:
Hi Larry

Considering a new machine configuration...

Will SAN upgrade / degrade performance of Write-intensive applications ?
(Compared to 7TB internal disks?)

TIA
Gad







date: Wed, 22 Nov 2017 16:01:15 -0500
from: DrFranken <midrange@xxxxxxxxxxxx>
subject: Re: i hosting i question

Sure VIO can do that but SAN better choice.
You get things like easy tier to move hot data to SSD and cold data away.

You get flash copy for doing backups at a point in time that's not
stuck while you are doing the backups.

You can add drawers on the fly unlike internal disk. There are also
more disk options with SAN as well.

I do NOT hate internal disk, reliable and predictable for decades
but SAN disk brings options that internal doesn't.

18TB, I'd be on SAN for sure.


- Larry "DrFranken" Bolhuis

www.Frankeni.com
www.iDevCloud.com - Personal Development IBM i timeshare service.
www.iInTheCloud.com - Commercial IBM i Cloud Hosting.


--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing
list To post a message email: MIDRANGE-L@xxxxxxxxxxxx To subscribe,
unsubscribe, or change list options,
visit: https://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx Before posting, please take a moment to review the archives at https://archive.midrange.com/midrange-l.

Please contact support@xxxxxxxxxxxx for any subscription related questions.

Help support midrange.com by shopping at amazon.com with our affiliate
link: http://amzn.to/2dEadiD
--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing
list To post a message email: MIDRANGE-L@xxxxxxxxxxxx To subscribe,
unsubscribe, or change list options,
visit: https://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx Before posting, please take a moment to review the archives at https://archive.midrange.com/midrange-l.

Please contact support@xxxxxxxxxxxx for any subscription related questions.

Help support midrange.com by shopping at amazon.com with our affiliate
link: http://amzn.to/2dEadiD
--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing
list To post a message email: MIDRANGE-L@xxxxxxxxxxxx To subscribe,
unsubscribe, or change list options,
visit: https://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx Before posting, please take a moment to review the archives at https://archive.midrange.com/midrange-l.

Please contact support@xxxxxxxxxxxx for any subscription related questions.

Help support midrange.com by shopping at amazon.com with our affiliate
link: http://amzn.to/2dEadiD




As an Amazon Associate we earn from qualifying purchases.

This thread ...
Replies:
Follow On AppleNews
Return to Archive home page | Return to MIDRANGE.COM home page

This mailing list archive is Copyright 1997-2024 by midrange.com and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available on our policy page. If you have questions about this, please contact [javascript protected email address].

Operating expenses for this site are earned using the Amazon Associate program and Google Adsense.