It’s been more than 2 years and a half since I last blogged on Oracle performance. I had the feeling I could not carry on this extra time activity when I started managing a 20 people DBA team and 4 technologies (including 2 Oracle competitors- Microsoft and Sybase which has ASE and IQ). It would have been difficult in the first place not to be labeled Oracle-biased by Oracle competitors if I did. I also moved from in-depth troubleshooting, which provided me material to write, to architecture design. This is more on assessing business requirements and translating them into technologies, with an emphasis on scalability and Business Continuity . Although ideas could still have been shared, perhaps towards a different audience, I decided it was time to stop, at least while I was holding this position. I left the blog online though and it still attracts 3000 hits a month as of August 2010, something that slightly worries me as the material will invariably become obsolete. If you read those lines, please note I was working on 10.2.0.4 or on older 10g releases.
August 26, 2010
I submitted an abstract to the Oracle OpenWorld 2010 that was found to be worthwhile enough to be accepted. Here are the session details:
Speaker(s) Christian BILIEN, BNP PARIBAS Corporate Investment Banking, Head of the DBAs
Monday, September 20, 2:00PM | Moscone South, Rm 236 60 min.
Session ID: S314649
Title: Large-Scale Oracle RAC (and Sun Oracle Database Machine) Scalability Design
Abstract: In this technical session, hear how a large corporate investment bank tames a yearly 50 percent to 100 percent data growth rate with Oracle Real Application Clusters (Oracle RAC) and Sun Oracle Database Machines. With databases totaling hundreds of terabytes, an integrated design of databases, storage, and servers is the cornerstone of scalability. This presentation highlights storage design from an Oracle RAC perspective as well as the specific challenges of Oracle RAC data warehouses and Oracle RAC geographical clusters. Real-life measurements on Oracle Solaris, Linux, EMC² storage, and RDS over InfiniBand (outside and inside the Sun Oracle Database Machine) illustrate performance and scalability perspectives and bottlenecks.