 | Level: Introductory Alan Lee (ykalee@ca.ibm.com), DB2 UDB Software Developer, IBM Toronto Lab
Ken DeLathouwer (kenjd@ca.ibm.com.), DB2 UDB Software Developer, IBM Toronto Lab
Punit Shah (punit@us.ibm.com), Advisory Software Engineer, IBM US
03 Jun 2004 Database server performance depends heavily on the I/O subsystem performance. Starting with the IBM® DB2® Universal Database™ V8.2 release, DB2 UDB supports Direct I/O and Concurrent I/O on AIX®, and Direct I/O on HP, Solaris, Linux, and Windows®. The focus of this article is AIX. It discusses various application I/O models available on AIX, and then describes how DB2 takes advantage of the CIO feature.
Introduction
Database server performance, typically measured in terms of the transaction throughput and application response time, depends heavily on the I/O subsystem performance. To attain the best possible I/O throughput, database table data layout demands special attention from database and system administrators. Today this is more pressing than it has ever been, considering the ever-widening gap between processor and disk speed. Chosen data layout has great impact on the manageability and extensibility of the database storage requirements.
Starting with the Version 8.2 release, DB2 UDB supports Direct I/O and Concurrent I/O (CIO). CIO is a new file system feature on AIX. The authors of this article explain DIO and CIO, and help you understand how to use this feature on your DB2 UDB system on AIX. Using this feature can help you optimize your database performance.
Topics covered include:
- The DB2 storage model and I/O modes available using either the file system or raw devices
- Enabling Direct I/O and Concurrent I/O on Stinger
- Recommended OS maintenance levels and fixes
- Test system configuration
- Performance test results
- Recommendations
Read the full article (PDF) here.
About the authors | 
| | Alan Y Lee works as a staff software developer in the DB2 UDB Operating System Services department where he develops and maintains an abstract layer of operating system APIs. These APIs are built into a library called OSSe (Operating System Service Everywhere) to be leverage by other IBM groups. During his first two years in IBM, he was the UNIX team lead in DB2 UDB Development Infrastructure Team. |
| 
| | Ken DeLathouwer works as a software developer at the Toronto Lab, IBM Canada, where his current focus is on DB2 UDB performance for data warehousing. In the past year since Ken joined IBM, he has been working on DB2 UDB performance mprovements and developing tools for benchmarking. He assists customers in fine tuning DB2 UDB to achieve optimal performance. |
| 
| | Punit Shah is an Advisory Software Engineer in the eServer Solution Enablement/Database Engineering group. His primary responsibilities are enabling database servers to use the latest AIX technologies. He has been working in the database server development, performance area for last 8 years. Punit holds a Bachelors and Masters degree in Computer Science and has authored or co-authored several articles for developerWorks. |
Rate this page
| |
