Analysis of performance data from a relational database system for applications using stored procedures or SQL

Abstract

Analyzing performance data from a relational database to discover poorly performing SQL statements includes creating event records of performance data for query statements executed by a relational database management system, and condensing the performance data in the event records. A performance summary can be generated for each condensed event record. The performance summaries can then be analyzed based on performance characteristics to identify poorly performing SQL statements.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS[0001]Under 35 USC §120, this application is a continuation application and claims the benefit of priority to U.S. patent application Ser. No. 11 / 080,171, filed Mar. 15, 2005, entitled “Analysis of Performance Data from a Relational Database System for Applications Using Stored Procedures or SQL,” which is incorporated herein by reference.FIELD OF THE INVENTION[0002]The invention relates to the field of database management and performance analysis.BACKGROUND OF THE INVENTION[0003]Data records in a relational database management system (RDBMS) in a computer are maintained in tables, which are a collection of rows having the same columns. Each column maintains information on a particular type of data for the data records of the rows. Data from the database tables may be accessed using instructions in a language that the database management system recognizes. For example, Structured Query Language (SQL) statements are often used to query, access, ...

AI Technical Summary

Benefits of technology

[0015]The compression of raw data from an event monitor into a compact form, such as one line per SQL statement, is performed to show relevant details for the statements, such as the text of static or dynamic SQL statements. This allows a user to more easily see the complete flow of operations and performance of each statement, both within a single application context and across the system.

Problems solved by technology

However, a few poorly performing statements can consume a significant amount of the database management system resources, resulting in a slowdown of the system.

As a result, searching this amount of data to locate a problem statement is very time consuming, and is therefore often not done.

If the problem statement is executed infrequently, such as once at start-up, then the overall impact on system performance may be insignificant, and the problem statement can be left in place.

Without a tool to summarize the performance records of the executed statements, identifying the problem statement, and the statement's overall impact on system performance, is very difficult.

As a result, most users who are not familiar with the internal storage used by the database management system often attempt to locate the text of the static SQL statement within the source code of the application program using the package name and section number from the event monitor, but this can be time consuming and error prone.

This may also be impossible, if the user does not have access to the application source code.

However, if the event monitor is from another system, then the user may have to ask a customer to find the text of the static statement.

Images