DB2 10 Hash Access: Access Path or Collision Course? - B06

Session Number: 84
Track: DB2 for z/OS - II
Session Type: Podium Presentation
Primary Presenter: Donna Di Carlo [BMC Software]
Time: May 16, 2012 (09:45 AM - 10:45 AM)
Room: Room 111 & 113

Speaker Bio: Donna Di Carlo has been in the software business for over 30 years. Her early days were devoted to Performance Tuning and Capacity Planning for the automotive industry. The last 20 years have been with BMC Software as a Lead Software Developer specializing in DB2 Backup and Recovery.
Audience experience level: Intermediate
Presentation Category: Data Management
Presentation Platform: DB2 for z/OS
Audiences this presentation will apply to: Application Developers, Data Architects, Database Administrators, Systems Programmers
Technical areas this presentation will apply to: Database Performance (DB2 for z/OS), New Release (DB2 for z/OS)
Objective 1: An introduction to the new DB2 10 Hash Access
Objective 2: See how a hashed table and index are physically organized
Objective 3: Determine if Hash Access is suitable for your application
Objective 4: Some rules of thumb on how to define hashed tables and indexes
Objective 5: What happens to performance when hashed objects are defined correctly; and worse, what happens when they are defined incorrectly

Abstract:  Hash Access is a new DB2 10 access path which could reduce the number of GETPAGE requests for a random access. But not all table spaces are candidates for this new feature. This presentation will cover the internals of hash organization and compare it to the index and table structures we’re all used to. Then, we will discuss tips on how to determine which table spaces are good candidates and provide some rules of thumb on how to best define your hashing structures. We will also look at some performance statistics.