Should I use blocksize larger than 4096 for my IXX?

Products

• Datacom
• DATACOM - AD
• CIS
• COMMON SERVICES FOR Z/OS
• 90S SERVICES
• DATABASE MANAGEMENT SOLUTIONS FOR DB2 FOR Z/OS
• COMMON PRODUCT SERVICES COMPONENT
• Common Services
• CA ECOMETER SERVER COMPONENT FOC
• Easytrieve Report Generator for Common Services
• INFOCAI MAINTENANCE
• IPC
• UNICENTER JCLCHECK COMMON COMPONENT
• Mainframe VM Product Manager
• CHORUS SOFTWARE MANAGER
• CA ON DEMAND PORTAL
• CA Service Desk Manager - Unified Self Service
• PAM CLIENT FOR LINUX ON MAINFRAME
• MAINFRAME CONNECTOR FOR LINUX ON MAINFRAME
• GRAPHICAL MANAGEMENT INTERFACE
• WEB ADMINISTRATOR FOR TOP SECRET
• Xpertware

Issue/Introduction

When defining the index for a CA Datacom database, the default blocksize is 4096.  For many types of files, changing a blocksize to a larger value will help performance of that file. However, this is not true for all types of files.

Should I use blocksize larger than 4096 for my IXX?

Environment

Release: DBCISF00200-15.1-Datacom-Database-Option for CICS Services
Component:

Resolution

The default value of 4096 has been identified as the optimum blocksize for the index. In fact, this was the only option for the longest time, and an option to use 8192 was introduced for a specific, unique situation. The tradeoff here is that while you might save very minimally on your I/O to the index, you will also increase your CPU usage and search times scanning the index records.

Consequently, there is no real benefit in increasing the blocksize for most databases, and there is a possible detriment to overall performance. In addition, unless the blocksize for every database is changed to this higher value, there is a potential increase in the amount of storage used for index buffers, and a possibility that half of this buffer storage is wasted by having to allocate buffers at the largest blocksize.

Additional Information

As always, please contact CA Technologies support for CA Datacom if you have further questions.