Test for Scalability (Batch Optimize)
When you optimize a SQL statement, it is important to consider the scalability of the alternatives SQL Optimizer generates before you implement them in your production database. SQL performance and scalability are two important factors that determine a database application's ability to handle an increase user load. The performance of a SQL statement changes as the number of users increase. Performance problems can arise when you use certain SQL statements under a heavy user load.Note: This topic focuses on information that may be unfamiliar to you. It does not include all step and field descriptions.
To test for scalability
- Select the Batch Optimize SQL tab in the main window.
- Select a batch from the Batch List node in the Batch Job List pane.
- Select a job from the batch node.
-
Select a SQL statement in the SQL List pane.
Note: You can only test optimized statements with faster alternatives.
-
Click the arrow beside
and select an option.
-
Review the following for additional information:
Select SQL Statements Description Original SQL
Select checkbox to include your original SQL statement when testing for scalability. SQL statement selected by the Best Alternative Criteria setting
Select checkbox to include the best alternative statement when testing for scalability. SQL alternatives whose Total Elapsed Time is faster than the original SQL
Select checkbox to include alternative statements with total elapsed times faster than the original SQL statement when testing for scalability.
Tip: Select Top N SQL and enter a value to only include N number of statements. The default value is 5.
SQL alternatives whose First Row Elapsed Time is faster than the original SQL
Select checkbox to include alternative statements with first row elapsed times faster than the original SQL statement when testing for scalability.
Tip: Select Top N SQL and enter a value to only include N number of statements. The default value is 5.
Advanced Options Description Minimum number of virtual users
Enter the minimum number of virtual users to use for scalability testing.
Range: 1 to 1,000,000,000
Maximum number of virtual users
Enter the maximum number of virtual users to use for scalability testing.
Range: 1 to 1,000,000,000
Step value of virtual users
Enter the number of virtual users to increase by for each step of scalability testing.
Range: 1 to 1,000,000,000
Distribution Model
Select a latency think time distribution model:
- None—Does not insert a latency think time between executions.
- Absolute—Uses the value entered in the Duration field as the latency think time between executions.
- Uniform—Uses a random duration between 1 millisecond and the value entered in the Duration field as the latency think time.
- Negative Exponential—Uses a random duration based on a mathematical model weighted towards lower latency think times. The value entered in the Duration field is used for calculating think times.
- Normal—Uses a random duration based on a mathematical model weighted towards average latency think times. The value entered in the Duration field is used for calculating think times.
- Poisson—Uses a random duration based on a mathematical model weighted towards discrete latency think times. The value entered in the Duration field is used for calculating think times.
Duration (milliseconds)
Enter a duration in milliseconds.
Note: How SQL Optimizer uses the value you enter depends on the distribution model selected.
Execute Each SQL Statement by
Select one of the following:
-
Duration (seconds)—Enter the duration of the scalability test in seconds.
-
Number of times—Enter the number of times to execute the SQL statement.
Range: 1 to 100
Generate Optimized Scripts
This topic focuses on information that may be unfamiliar to you. It does not include all step and field descriptions.To generate an optimized script.
-
Select the Batch Optimize SQL tab in the main window.
- Select a batch from the Batch List node in the Batch Job List pane.
-
Select a job in the Job List pane and click
.
Tip: Select the Replaced checkbox in the SQL Scanned and Optimized pane to include a statement in the optimized script.
-
Click
to save the optimized script.
Send to Optimize SQL (Batch Optimize)
This topic focuses on information that may be unfamiliar to you. It does not include all step and field descriptions.To send a SQL statement to Optimize SQL
- Select the Batch Optimize SQL tab in the main window.
- Select a batch from the Batch List node in the Batch Job List pane.
-
Select a job from the batch node.
-
Select a SQL statement in the SQL List pane and click
.
Tip: Right-click a SQL statement in the Alternatives pane of Optimize SQL and select Mark as Best in Batch Optimize SQL to save it as the best alternative.
Send to Optimize Indexes (Batch Optimize)
You can send jobs to Optimize Indexes to generate an index set that improves the performance of the SQL statements in the job. See About Optimize Indexes for more information.
Note: This topic focuses on information that may be unfamiliar to you. It does not include all step and field descriptions.
To send a job to Optimize Indexes
- Select the Batch Optimize SQL tab in the main window.
- Select a batch from the Batch List node in the Batch Job List pane.
- Click
. A Scan Code workload session opens in Optimize Indexes.
- The Choose SQL to Include in Workload dialog opens and displays the SQL that you sent from Batch Optimize SQL. Use this dialog to delete or modify SQL statements.
- See Optimize Indexes from Scanned Code Workload for more information about searching for new indexes from a scanned code workload.