The second keynote speaker at Dallas’ Oracle Day was Syamal Bandyopadhyay. He is the Database Deployment Architect at Verizon Wireless and has worked in IT for more than 25 years. Syamal is also active in the IOUG Exadata SIG and talked about how Verizon Wireless is using Oracle.
Because we need simplification. Business needs a reduced time to market, doing more with less, and with the system always available. IT has to be seen as a business enabler. That means we have to keep up the pace to process an increased volume of data at a reduced cost.
This is a cultural change. In order to change, the mindset must be ready to accept change. At Verizon, we do incremental changes. Look at supporting the current business in addition to adding functionality. There is significant cost involved but it’s worth it. This is a partnership and it doesn’t happen in one day.
- High Availability -> Continuous Availability
- Scalability to Support Growth
- Reduced Time to Market
- Reduced Total Cost of Ownership
- Need for Increased ROI
- Need for Simplification
RAC – Gives us horizontal scalability, use enterprise-class servers
ASM – Add hardware resources on demand
Tools for Availability
- Active Data Guard – Need Synchronous Mode for Best Protection, but There is a Performance Penalty on PROD
- Data Replication – Multi-directional
Engineered Systems / Exa Technology – More than just Exadata, it’s a different mindset. Multiple pieces are there, so we don’t have to test all the integration pieces. Oracle handles some of that for us.
Hybrid Columnar Compression (HCC)
Verizon is putting multiple databases on the Exadata servers. That consolidation reduces the maintenance costs. Have one engineered system replace multiple database servers.
- 2 Data Centers – Data Replication between the 2 production databases.
- 2 Production Databases – 4 instances for each Database
- 4 RAC Instances – Provides local high availability
- Oracle 188.8.131.52
- SUN M5000 Servers
- Hitachi Storage
- Real time Data Replication
35,000 users access the systems through switches and Fusion Middleware. Transactions are submitted through the next available database. Any data that a customer might need is replicated through any of the database. Because critical data is replicated in real-time, the availability can be continuous. Used Data Guard before, but the 20 minute lag time was too long. Business needed continuous access to real-time data.