I worked in a shop that went from a 3 CP machine to 2. Although the each of 2 individually ran at a higher speed than the engine in the 3-way, overall system throughput suffered. Even if CICS is single threading transactions, it is not the only thiing running on your box. We had a task that hogged one engine, cpu bound work, and everything else had just the 1 CP to work on. Don't forget, I/O runs on a CP, too. It isn't going to a multiple engine machine that is bad for you, it is believing adding the speed of the two equals the speed of one that is wrong. Your throughput benefits from more than one, cpu bound work benefits from faster engines. Your quandry is now figuring out just how fast an engine you really need if you have more than one.
On Thu, 6 Dec 2007 04:27:48 -0800, Patrick Falcone <[EMAIL PROTECTED]> wrote: >I would be reluctant to move to a 2 way from a uni and potentially cut my rating by CP by half. I wouldn't want to be taking the calls when peak hits. I'm not comfortable with taking a .2 second CPU transaction and making it possibly a .4 second CPU transaction. Most likely there is also some latent demand but how much? I have a hard time believing that this scenario is workable. Those VPS CPU spikes can be painful, I resource capped. > ---------------------------------------------------------------------- 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
