We are using a lot of data striping to get better performance on various files. However, we are using a ton of DASD in our stiping pools and are actually using less that 55% of it. Is there something else out there that we can do to get the performance without using so much DASD?
You haven't given enough information to really answer this question. The answer is "it depends". Is the data being accessed in TSO, batch, or online (DB2/CICS/IMS)? Perhaps you can trade memory for I/O. Can you increase buffer sizes? Can you use BSLR (batch LSR)? Can you use other SMS performance options? Is VLF or DLF an option?
If you (or a someone else in your organization) are a CMG member, you can find lots of information from their web site and proceedings. See www.cmg.org.
If batch performance is your biggest concern, you may want to review this IBM Redbook: System/390 MVS Parallel Sysplex Batch Performance.
This was first published in April 2006