Re: A quick question about velocity goals set high
I think you have answered your own question. The velocity is set so high that it's unreachable. My velocity goal for CICS and similar tasks is 40. And things are fine. __ Dave Thorn * Senior Technology Analyst * SunGard Computer Services * 600 Laurel Oak Road, Voorhees, NJ, 08043 Tel 856 566-5412 * Mobile 609 781-0353 * Fax 856 566-3656 CONFIDENTIALITY: This e-mail (including any attachments) may contain confidential, proprietary and privileged information, and unauthorized disclosure or use is prohibited. If you received this e-mail in error, please notify the sender and delete this e-mail from your system. Estimeed lister I'm in a shop were I see for some subsystems (CICS,CONTROM etc) a very high velocity goal, that is 90%. PIs range from 1.4 to 2 and it NEVER reaches 1 nor it's below 1. I read some pepres that a very high velocity goal (80%) is not good if not useless and infact in my previous shop in some CICS we lowered our goals without any delay in response time. Is still bad to put velocity goals higher than 80% ? In IMHO it is. Thank you in advance Max Scarpa -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: A quick question about velocity goals set high
On Wed, 28 Sep 2005 15:24:54 +0200, Max Scarpa [EMAIL PROTECTED] wrote: I'm in a shop were I see for some subsystems (CICS,CONTROM etc) a very high velocity goal, that is 90%. PIs range from 1.4 to 2 and it NEVER reaches 1 nor it's below 1. I read some pepres that a very high velocity goal (80%) is not good if not useless and infact in my previous shop in some CICS we lowered our goals without any delay in response time. Not necessarily true. I've been at shops with CICS regions that have a very high velocity goal (90) for some loved regions and were able to make that goal. It all depends on the environment. Is still bad to put velocity goals higher than 80% ? In IMHO it is. No, it's bad to define unatainable goals - period. This is what you have in your shop according to what you wrote. Regards, Mark -- Mark Zelden Sr. Software and Systems Architect - z/OS Team Lead Zurich North America and Farmers Insurance Group mailto: [EMAIL PROTECTED] Systems Programming expert at http://Search390.com/ateExperts/ Mark's MVS Utilities: http://home.flash.net/~mzelden/mvsutil.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: A quick question about velocity goals set high
Max, I am going to use the IBM reason, It depends WLM is relative to what your other goals are set to. It also depend on what importance you set the velocity go to. A high velocity with and importance of 1 will react a lot diffenently than a high velocity goal and an importance of 5. In the situation you stated the goal was designed to never be met so for the most part will get the resources it needs. Is this a good thing? Again, it depends. IMHO a high velocity goal is not a good thing, but again it depends on your other goals. Thanks, Fletch -Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Max Scarpa Sent: Wednesday, September 28, 2005 8:25 AM To: IBM-MAIN@BAMA.UA.EDU Subject: A quick question about velocity goals set high Estimeed lister I'm in a shop were I see for some subsystems (CICS,CONTROM etc) a very high velocity goal, that is 90%. PIs range from 1.4 to 2 and it NEVER reaches 1 nor it's below 1. I read some pepres that a very high velocity goal (80%) is not good if not useless and infact in my previous shop in some CICS we lowered our goals without any delay in response time. Is still bad to put velocity goals higher than 80% ? In IMHO it is. Thank you in advance Max Scarpa -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: A quick question about velocity goals set high
Thank you all for reply. No, I see that our (or better 'their') goals are NEVER reached, PIs are constantly above 1 (sometimes very above). I'm monitoring it since a many weeks. In my opinion the system is stressed trying to reach these goals but these goals are NOT set for CICS only but for many others AS (21) and for this reason it's not possible to say that goals are high because they are applied to 'loved ones' that, by definition, cannot be ALL CICS regions. For this and other reason I think the goal is wrong in THIS environment. Anyway in a previous shop I tested very high velocity but the difference between a goal of 90% and a goal of 70% was minimum for batch jobs and not so high for 1 or 2 CICS. Infact we had CICS velocity goals ranging from 30% to 50%. Thank you again and best regards Max Scarpa -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: A quick question about velocity goals set high
Max, I have to agree, an unattainable goal is an invalid goal. If you're getting a PI of 1.4 to 2 (or higher) that likely means that the work that is running in that particular service class is actually getting a velocity of 64 (at best ~ 1.4 PI) down to 45 or worse (at = 2 PI). We don't have any velocities over 50 for anything, no matter how important. Most of them are around 30 for started tasks. If it's so important that it needs the highest priority, it goes to SYSSTC. It sounds like a good review of your goals, importance levels, and actual service levels (i.e. how workloads and service classes are performing now) is in order. It gets even more fun if you're attempting to balance across a CECPLEX for IRD and ensure adequate performance for high importance workloads at 100% physical busy times. Best regards, Gary Diehl -Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] Behalf Of Max Scarpa Sent: Wednesday, September 28, 2005 8:25 AM To: IBM-MAIN@BAMA.UA.EDU Subject: A quick question about velocity goals set high Estimeed lister I'm in a shop were I see for some subsystems (CICS,CONTROM etc) a very high velocity goal, that is 90%. PIs range from 1.4 to 2 and it NEVER reaches 1 nor it's below 1. I read some pepres that a very high velocity goal (80%) is not good if not useless and infact in my previous shop in some CICS we lowered our goals without any delay in response time. Is still bad to put velocity goals higher than 80% ? In IMHO it is. Thank you in advance Max Scarpa -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: A quick question about velocity goals set high
Is still bad to put velocity goals higher than 80% ? In IMHO it is. ... Empirically, if you don't have I/O as part of the velocity equation, you can never achieve a velocity of more that 45-50. With I/O, the issue is not what you set it to. Rather, can you achieve it. If you set it to high, any goal will be unachievable. And, sometimes the SRM/WLM tandem will throw up its (metiphorical) hands, in disgust, and do nothing for the workload. If you need to achieve a goal, and aren't, find our why. And, fix it. Setting arbitrary rules (of thumb) and sticking to them is not performance analysis. -teD In God we Trust! All others bring data! -- W. Edwards Deming -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: A quick question about velocity goals set high
Empirically, if you don't have I/O as part of the velocity equation, you can never achieve a velocity of more that 45-50. What do you base this statement on? Adam -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: A quick question about velocity goals set high
On Wed, 28 Sep 2005 00:00:00 GMT, Ted MacNEIL [EMAIL PROTECTED] wrote: Is still bad to put velocity goals higher than 80% ? In IMHO it is. ... Empirically, if you don't have I/O as part of the velocity equation, you can never achieve a velocity of more that 45-50. With I/O, the issue is not what you set it to. Rather, can you achieve it. Huh?! With I/O priority management off, it doesn't matter. With it on it can only lower the velocity (if there are delays), not make it higher. Mark -- Mark Zelden Sr. Software and Systems Architect - z/OS Team Lead Zurich North America and Farmers Insurance Group mailto: [EMAIL PROTECTED] Systems Programming expert at http://Search390.com/ateExperts/ Mark's MVS Utilities: http://home.flash.net/~mzelden/mvsutil.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: A quick question about velocity goals set high
Empirically, if you don't have I/O as part of the velocity equation, you can never achieve a velocity of more that 45-50. What do you base this statement on? ... Empirical evidence. Statements from one of IBM's Canadian Performance Guru. -teD In God we Trust! All others bring data! -- W. Edwards Deming -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: A quick question about velocity goals set high
I dunno. I could easily design a cpu intensive workload (branch on count, anyone?) that could achieve and maintain a very high velocity for very long periods of time. I could pretty much guarantee that it could get over 80% using samples during just about any period, provided it was given a high enough importance. I'd expect calculation-heavy engineering jobs to perform in a similar manner. I would agree, however, that any I/O heavy workload would struggle to ever cap 50% velocity and stay there. Gary p.s. I hesitate to say never anymore, though I do still use the word occasionally. -Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] Behalf Of Ted MacNEIL Sent: Tuesday, September 27, 2005 7:00 PM To: IBM-MAIN@BAMA.UA.EDU Subject: Re: A quick question about velocity goals set high Empirically, if you don't have I/O as part of the velocity equation, you can never achieve a velocity of more that 45-50. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: A quick question about velocity goals set high
Empirically, if you don't have I/O as part of the velocity equation, you can never achieve a velocity of more that 45-50. What do you base this statement on? ... Empirical evidence. Statements from one of IBM's Canadian Performance Guru. Well .. I hate to disagree but I've got plenty of reports showing service classes with no I/O velocities getting velocities of over 90%. I'm afraid that there is absolutely nothing that restricts such a velocity ... Empirical or otherwise. Adam -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: A quick question about velocity goals set high
Huh?! With I/O priority management off, it doesn't matter. With it on it can only lower the velocity (if there are delays), not make it higher. ... I/O 'using' also counts. And, if it's higher than delay, you will get higher velocity. Especially, with a responsive DASD farm. We implemented Goal Mode when DISCONNECT counted as part of using. So, when we turned IO on, our velocities went up. And, when we turned on Dynamic PAV's, it went up again. My point was/is not so much the value of the velocity. Can you achieve it? If not, why not? If you have to, fix it! -teD In God we Trust! All others bring data! -- W. Edwards Deming -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: A quick question about velocity goals set high
Is it safe to translate the sometimes to once PI 4? ... I've been involved in situations where the PI was less than 4, but nothing could be done to fix it. -teD In God we Trust! All others bring data! -- W. Edwards Deming -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: A quick question about velocity goals set high
I dunno. I could easily design a cpu intensive workload (branch on
count, anyone?) that could achieve and maintain a very high velocity for
very long periods of time
...
Artificial workloads don't count! (8-{]}
But, if I ever worked in a shop with a lot of engineering type work I might
have said otherwise.
What part of 'empirical' do you not understand.
Besides, I eventually have to do IO.
But, this is an aside.
My point was/is: not the value; the acheivement is what counts.
-teD
In God we Trust!
All others bring data!
-- W. Edwards Deming
--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: A quick question about velocity goals set high
On Wed, 28 Sep 2005 00:00:00 GMT, Ted MacNEIL [EMAIL PROTECTED] wrote: We implemented Goal Mode when DISCONNECT counted as part of using. So, when we turned IO on, our velocities went up. That was changed in OS/390 R8 - R10 (via PTF). A long time ago now. There is another thread already referencing R8 of z/OS. Mark -- Mark Zelden Sr. Software and Systems Architect - z/OS Team Lead Zurich North America and Farmers Insurance Group mailto: [EMAIL PROTECTED] Systems Programming expert at http://Search390.com/ateExperts/ Mark's MVS Utilities: http://home.flash.net/~mzelden/mvsutil.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: A quick question about velocity goals set high
That was changed in OS/390 R8 - R10 (via PTF). A long time ago now. There is another thread already referencing R8 of z/OS. ... Sigh! YES! I know. My point was not about values, rather the achievement of goals. As usual, IBM-Main goes on about the side comments. -teD In God we Trust! All others bring data! -- W. Edwards Deming -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: A quick question about velocity goals set high
Trying to gently veer this thread back on course, unlike everyone else my reaction is who cares ???. If CICS regions are (as I would expect) managed via transaction goals, high (generally unmet) velocity goals are meaningless. We set ours high to help startup/shutdown, then forget about them. Shane ... -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: A quick question about velocity goals set high
I was into the “who cares?” mentality early. But, others weren't. The questions on the table, gentle-beings, are: 1. Are your goals being met? 2. Are your SLA's being met? If the answers are “YES” (#2 being more important), go do something else. If the answers are “NO”, the question is: 3. Are your business processes being impacted? If the answer is “NO”, go do something else. If the answer is “YES”, the question is: 4. What can I do to fix it. The question is NEVER what is a 'BAD', or too 'HIGH', goal. Goals are relative to each other, NOT absolutes. The real answer is: We do what we have to, to get the job done. Rules Of Thumb do NOT rule! -teD In God we Trust! All others bring data! -- W. Edwards Deming -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
