In today’s complex web application world, developers need to test applications that go beyond simple HTTP-based pages. They need to test Rich Internet Applications (RIAs) that incorporate complex technologies like Adobe's Flex.

Adobe Flex applications may be different from applications you worked with previously.  With applications that deliver HTML as the user interface, the server does all of the processing. Users simply page back and forth to the server. Flex applications are different. These applications download the Flash client application and run in the browser or on the desktop for Adobe AIR applications. They only go back to the server periodically to request or send data. For this reason, server loads reveal a very different profile.

With web applications, performance is measured mainly by the round-trip times between the client’s web browser and the server that provides the data to the browser. Flex technology can cause significant increases in the number of browser-to-server HTTP calls made in the background. This increase in traffic can have a profound effect on performance. While users might not be aware of the round-trips between the browser and a distant server, they will definitely notice performance problems if the application is slow or doesn’t work because of the increased load. This situation makes load testing very important when working with Flex.

With innovative RIA technologies like Flex, there are many new challenges to address. If you can address these challenges, you can expect success with Flex. If you’re developing or deploying applications in this fresh environment, you need to think about load testing in new ways. This article will to discuss the new challenges in load testing Flex applications and how to best address them.
 
What You Want to Test?
Before you start testing an application, it is important to determine your application testing objectives. Do you want to test an application's robustness, performance, infrastructure and/or stability for a predefined user load?  If yes, then you want to simulate user experience at a constant rate over a period of time. The test results should answer the following questions:

• Is the server capable of handling a certain number of simultaneous users?
• Is the average response time for pages acceptable under this set load?
• Does the server revert to normal behavior after a load peak?