> Billing to the client is based on > CPU-seconds. Batch takes about 80% of system resources in terms of cpu. > Given the same DB/2 batch job ( consider DB/2 on MVS being the same ), > same database volume, same z9 Server and so on, so take only z/VM and the > z/OS cpu cycles into account, would there be a difference and if so, how > big would that difference be ?
There will certainly be a difference, but it's pretty much impossible to tell you what the difference will be because both the OS and the database code are substantially different enough that the results are probably not going to be comparable. DB/2 on z/OS has a lot of code that was never added to DB/2 for VM. The optimizers are different. The OS scheduling algorithms are totally different. You're comparing apples to mangoes here -- both are good, but in totally different ways. The only way you're going to be able to tell is to run multiple jobs multiple times in both environments and get some comparison data. You will certainly need to raise your per-second rate to accommodate the higher cost of z/OS resources, software and support personnel, probably at least 20-30%, but you need to compare real numbers for utilization comparisons.