As you already know, when a product is tied to a database, it is the database that provides features to enhance the performance rather than the product itself. In other words, database of a product plays a vital role when it comes to performance.
Besides, already known capabilities like materialized views, massive parallelism, etc., there are a few new popular techniques like in-memory, columnar processing, compression, and data skipping, which helps the user to find the needed data more quickly.
Let’s see what its peers are saying about Oracle Exadata:
- Uses columnar processing; it uses columns to compresses data. However, while handling a query, it decompresses it.
- Uses flash-cache, which is located on the storage server [not on the database server], for faster processing
- Uses Smart Scans feature, which is capable of querying historic data rather than real-time data.
- Uses workload management, which is based on first come, first serve basis; which sometimes results in keeping resources idle.
- Uses query queuing, which consumes lot of space that in turn hampers performance.
- Possess no ability to control the Storage Server CPU, which is consumed by a workload; in turn results in difficulty in maintaining SLAs.
- Recommends extra hardware to improve performance
What you suggest?