Capacity planning

mlp · New User Joined: 23 Sep 2005 Posts: 91

The SRM constat is the number which used to normalize the CPU seconds and convert them to SUs.

The SU is used to measure the amount service done by the infrastructure. But there is MIPS terminology also, used to guage the usage of infratrsucture. Which of them used to plan the capacity? and which is used for billing?

Note- Not sure "JCL" is the right forum to open this topic. If not, please route to appropriate forum.

vasanthz · Posted: Tue Aug 06, 2013 12:46 pm

Hello,

Robert Sample · Posted: Tue Aug 06, 2013 3:26 pm

You need to separate the concept of "billing" from capacity planning. Billing is an internal IT function and as such may exist in any number of forms -- flat rate billing, usage based on cpu / I/O / disk space / tapes / some combination of these, and as such billing is an accounting function and must be discussed with your site's accounting department.

MIPS has not been a valid term to use with capacity planning for many years -- IBM would prefer you to use the term MSU instead. And even when it was used, MIPS was applied to the entire machine (LPAR these days) -- people would, very incorrectly, say that application X used so many MIPS. No, it did not -- it may have used so many CPU seconds, but unless the application used the entire machine 24 hours a day 7 days a week the application used 0 MIPS. If you know someone still using MIPS for anything, you need to know that there is a person who is (a) many years behind the times, or (b) completely deluded about mainframes.

Capacity planning ALWAYS uses MSU, but there is more to capacity planning than just MSU. It makes no sense for your site to have enough CPU power to process the workload if there aren't enough tape drives, or enough disk space, or if the channels run 90% of capacity for most of the day -- all of these will slow down throughput and impact performance as much as the CPU.