>Perhaps it's the design of WLM that doesn't account for profoundly diverse behaviors within a single LPAR. >Any explanation is welcome.
Now we're starting to get into opinion. But, I'll start the discourse with the bumps and bruises I earned in the early 1980's. Back then, the attitude was that TSO should have four periods, but thankfully they dropped to three. The attitude was also that n% had to complete in the first period. If TSOP1 was not one of the highest priorities, response was erratic. If the period was sized poorly, other workload (Production Batch, Onlines, etc.) suffered. I used to do three periods, with the third as a MTTW. MTTW does not exist, except for DISCRETIONARY, under GOAL MODE. Also, with only 5 IMP's, there is a lot less granularity. Somebody said that the WLM does better with diverse workloads. Due to my experience with it, I disagree. I have found that TSO is better served in a separate LPAR (within the same SYSPLEX), rather than on with an IMS/CICS/DB2/etc workload. You either have to bump TSO up (hurting Production [especially Batch]), or leave it so low that nobody is productive. That's why I have taken the SYSNAME/DA route for the people who are monitoring. When in doubt. PANIC!! ---------------------------------------------------------------------- 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
