Mike, I think the term "peak" is ambiguous. Does that mean the highest average utilization?
You can get an average and maximum for a given time interval from RMF. Correct RMF interval size can be somewhat subjective, but the problem is not having the distribution of samples to plot the standard deviation and coefficient of variance. If your workload is relatively stable, you should be able to chart the peak-to-average ratio which prevails during times of adequate system responsiveness. You can then use the trended average as a rough gauge of when you expect to hit the existing capacity limit. How rough? Like the guy in the Meineke commercial says, "some rougher than others." I agree with Adam Gerhard that what matters is how long dispatchable work has to wait in the queue. How much queueing your loved ones can afford needs to be assessed. However, sometimes you just need a quick answer for planning purposes and you have neither the time nor the audience for mathematical modeling and queueing theory. Dave Barry Sr. Performance & Capacity Planning Analyst UPS -----Original Message----- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@bama.ua.edu] On Behalf Of Ward, Mike S Sent: Friday, April 15, 2011 4:58 PM To: IBM-MAIN@bama.ua.edu Subject: CPU utilization/forecasting Hello all, can someone tell me if it's better to use CPU peak or CPU average to project growth. My way of thinking is if you use peak then you're sure to show where you need extra horse power. In other words if you can't process during your peaks what good are the averages. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
