Analyze the Statistical Data

Indexes not Accessed Randomly
- All or Mostly Prefetch (SAMOS)
Objects having highest rates of Access
- Unusual/Unexpected Access/Activity
Overall, and Object I/O Rates
- 160 I/O per Second is Approximately 1 MIP of Processor

Previous slide Next slide Back to the first slide View Graphic Version

Notes:

What do we look for in the statistical analysis:

Indexes with a high % of sequential access, or Sequentially Accessed Mostly (SAMO). The indexes and tablespaces that are Randomly Accessed Mostly (RAMO). Additionally, we break these into ‘small’ and ‘large’ working sets. This then gives classifications of RAMOS, RAMOL, SAMOS, SAMOL.

The end result of our analysis is to group similarly accessed objects together. The RAMOL/RAMOS and SAMOL/SAMOS objects.

Additionally, we must know specifically how they are accessed, and the average and maximum number of pages they had resident in a pool at any point during the collection period.