Future and Internals for DBA - Part I - D05

Session Number: 288
Track: DB2 for LUW - II
Session Type: Podium Presentation
Primary Presenter: Matt Huras [IBM]
Time: May 16, 2012 (08:30 AM - 09:30 AM)
Room: Room 106

Speaker Bio: Matt Huras is an IBM Distinguished Engineer, and a senior architect of DB2 on the Linux Unix and Windows platform. He focuses on the database kernel, which includes the data management, index management, locking, concurrency and other protocols. Matt has been with IBM for 27 years, and has a Bachelor of Applied Science from the University of Waterloo, and a Master of Engineering from the University of Toronto. Matt's most recent project was DB2 pureScale - a new feature focusing on delivering new levels of scalability and availability. Matt is DB2 pureScale's lead architect.
Audience experience level: Beginner, Intermediate, Advanced
Presentation Category: Data Management
Presentation Platform: DB2 for Linux, UNIX, Windows
Audiences this presentation will apply to: Database Administrators, Systems Programmers, IT Managers
Technical areas this presentation will apply to: Database Performance (DB2 for LUW), Disaster Recovery (DB2 for LUW), High Availability (DB2 for LUW), New Release (DB2 for LUW), pureScale
Objective 1: Learn key DB2 LUW internal algorithms
Objective 2: Learn best practices for influencing DB2 LUW internal behavior
Objective 3: Learn about the internals associated with the latest releases

Abstract:  DBAs often need to understand what's going on 'under the covers' in the relational database system, in order to fully appreciate how best to tune the database, and to have full confidence in their database installations. This presentation gives DBAs definitive knowledge about DB2 on Linux Unix and Windows' internal algorithms and how to control them. It covers key concepts like how tablespaces are designed, SMS vs DMS vs Automatic Storage internals, space management within tables, index management, memory management and STMM, logging and recovery, locking, parallelism, shared nothing and shared data clustering, and how threads and processes are used. In each of these areas, the presentation covers internal design fundamentals at a conceptual level, and then follows up with guidance on how to influence the described DB2 behavior via tuning knobs and controls. The material in this presentation will be up-to-date as of the latest versions of DB2 available on the conference date in May 2012.