15.10 - Workload Identification - Teradata Database

The objective of workload identification is to isolate the set of requests, called a workload, you want to analyze for the possibility of enhancing their performance by redefining the secondary and single‑table join indexes defined on the tables they access.

There are several methods for identifying workloads, including these.

Ad hoc qualitative assessment of the performance of some queries.

This is a “something is not right here” approach.

Your queries are taking longer to complete than you think they should, but you are unable to identify precisely what the cause of the slow responses is.

Sift through the following data stores in search of candidate queries for performance enhancement:

The database query log using the Teradata Index Wizard (see Teradata Index Wizard User Guide for details).

The query capture database using tools like Visual EXPLAIN or Teradata System Emulation Tool.

After determining the set of queries to be analyzed, you can identify the workload using the Teradata Index Wizard tool or the Teradata System Emulation Tool (see Teradata System Emulation Tool User Guide).

See Teradata Index Wizard User Guide for details about this process.

The process of identifying workloads is indicated by the following graphic:

Key
Input Option	Description
View query plan	You can invoke the Index Wizard from a query plan viewed by means of the Visual EXPLAIN tool. The query used for index analysis is the active plan in Visual EXPLAIN at the time the Index Wizard is invoked.
Feed SQL text	You can directly type the text for the queries that are to make up a workload or you can load them from the Teradata platform.
Query DBQL	You can fetch logged Database Query Log SQL text for index analysis by means of the Index Wizard.
Get target query	You can export a set of queries for analysis using Teradata System Emulation Tool.