Session ID: 281
Learn how to use simple algorithm analysis for Oracle performance tuning. Finding the mathematical functions that describe the computational complexity of different operations can help us understand and predict performance. This presentation will discuss common functions such as 1/N, LOG(N), N, N*LOG(N), N^2, and N!, and how they can help us understand operations such as sequence caching, bulk collect limit, index access, full table scans, hashing, sorting, joining, parallel processing, and choosing execution plans. This presentation also discusses times when the computational complexity is negated by implementation details. With a new way of looking at simple mathematical functions we can add algorithm analysis to our performance tuning toolkit.
Objective 1: Learn how to use simple algorithm analysis to better measure performance.
Objective 2: Better understand indexes, hashing, sorting, parallel processing, and the optimizer.
Objective 3: Stop wasting time on settings that cannot theoretically improve performance, such as excessive sequence caching and bulk collect limit sizes.