Distributed Load Testing

Overview

 

Today, Web applications/Web sites get a great deal of traffic. To effectively handle the traffic and offer improved performance for business-critical Web applications, companies need to predict the maximum load the web server can adequately handle using a distributed architecture. To measure the performance of such a system, QEngine provides the Distributed Load Testing option.

 

Illustration

 

 

Understanding Distributed Load Testing

 

QEngine distributed load testing generates load from one single Test Controller and simultaneously distributes the configured load in the remote play engines. The distributed playback engine has the following components:

 

Test Controller

 

Test Controller is the machine were QEngine is installed. The functions of Test Controller are as follows:

QEngine Distributed Play Engines

 

Distributed play engines are the machines where QEngine is installed with RMI Server for remote test execution. The functions of play engines are as follows:

Common Database

 

The common database is used both by the Test Controller and the distributed play engines. The user requests generated for the configured test parameters, the responses received from the distributed play engines and the test reports generated and sent by the distributed play engines are stored in the common database.

 

Steps to enable Distributed Load Testing

 

The steps to enable distributed playback are as given below:

 

Note: You can skip steps 1 to 3, if you have already recorded the business cases and configured the test parameters for load test execution.

  1. Record the business cases as per the steps given in the topic Recording Business Case.

  2. Parameterize the recorded URLs (if required) as per the steps given in the topic Handlng URL Parameters.

  3. Define test parameters such as user profiles and workloads. Configure server monitoring and virtual Ips (optional). Create test cases to associate the profiles and workloads. To know the details, refer to the steps given in the topic Defining User Profiles, Defining Workloads, Configuring Server Monitoring, Configuring Virtual IP Addresses and Creating Load Test.

  4. Install QEngine in a machine which acts as the Test Controller.

  5. As per your requirements to distribute the load, install QEngine in machines with different hardware configurations. Start remote agent in each of the distributed machines using StartDistributedAgent.bat/.sh file in <QEngine_Home>/bin directory.

  6. Open Web Performance Test Studio from Test Controller, choose Play->Play Settings menu item from the menu bar. In the Other Details panel, select the check box Distributed Playback and click OK.

  7. Choose Play->Add Play Engine menu item from the menu bar. The Play Engine Configuration screen will be displayed.

  8. In this screen, configure the IPs, RMI port and maximum virtual users to be simulated in the distributed play engines. This is to enable the Test Controller to remotely contact the distributed play engines to perform distributed load testing. For details of configuring Play Engine IPs, Port, and Max Users, refer to the context-sensitive help in the following link.

  9. After configuring the IPs, Ports, and Max Users, start the load test execution as per the steps given in the topic Executing Load Test.

  10. The test reports will be stored in the common database. The test logs will be stored in the respective directories in the Test Controller. To know the report details, refer to the topic Load Test Graphs/Reports.



Copyright © 1999, AdventNet Inc. All Rights Reserved.