We have a TSM 4.2.0.0 server running under OS/390. We are getting ready to install a Hitachi Data Systems 9960 disk subsystem. We are planning to put some disk storage pools, among other things, on the new subsystem. About 40 of the emulated 3390 model 3 volumes on the subsystem will be used this way. All of them will be carved out of a RAID group consisting of four physical disk drives with a capacity of 73 gigabytes each. There will be sixteen gigabytes of cache memory between the physical disks and the mainframe. How should I subdivide the available space into storage pools? For example, should I make each pool out of datasets that are as large as possible and fit on as few virtual disks as possible, or should I use smaller datasets and stripe each pool over as many virtual disks as possible?
