I'd be remiss if I didn't share some thoughts about last week's DB2 11 announcement.
The level of CPU reduction immediately caught my attention. According to the IBM announcement letter:
"When compared to running on DB2 10, IBM's internal testing and Early Support Program results revealed that, depending on workload, customers may see 'out-of-the-box' DB2 CPU reductions:
· Up to 10% for complex OLTP,
· Up to 10% for update-intensive batch, and
· Up to 40% for complex reporting."
Unquestionably, IBM continues to drive down the cost of owning DB2 for z/OS.
One of the most critical enhancements for availability is the extended log record addressing capacity (1 yottabyte) for sites with heavy logging. I’ve heard that some customers have to reset their RBA back to zero every two years. Extended logging in DB2 11 should eliminate the need to take such drastic steps.
Big data and data analytics are major themes for DB2 11. IBM projects a 5-10 percent reduction in CPU using separately licensed DB2 Analytic Accelerators. The announcement letter refers to a 5-10 percent reduction in CPU for large analytic statements. In addition, DB2 11 provides integration with the separately available IBM/Hadoop-based BigInsights big data platform using SQL. This is designed to provide customers with a simple way to integrate traditional applications on DB2 for z/OS with big data analytics.
One thing to remember about upgrading to DB2 11 is that you can't skip from DB2 9 to DB2 11; you'll have to first upgrade to DB2 10. Fortunately, migrating from DB2 10 to DB2 11 should go very quickly. The catalog conversion to DB2 11 is a whopping 16 times faster than the conversion from DB2 9 to DB210.
Another advantage to upgrading to DB2 11 is that existing DB2 business applications can run in compatibility mode. This protects these applications from incompatible SQL and XML changes.