In many high end production environments, it is simply impossible to perform realistic load testing manually. Such testing, however, is a crucial task that must be performed to ensure that a product is ready for live production situations. To solve staffing and hardware limitations, Hewlett Packard offers LoadRunner as a testing and performance analysis tool.
LoadRunner provides quantitative analysis of application performance while automating the process of actually creating a “load.” This circumvents the need for dozens, if not hundreds, of live users sitting at just as many computers. LoadRunner replaces human testers with “Vusers” that can use only a handful of hosts to generate thousands of requests on an application. When testing an application, these Vusers act out specific script driven scenarios. The results are recorded in detail and used to create a post-test report.
LoadRunner consists of the Controller, VuGen, Analysis, and Tuning modules. The Controller allows the operator to configure individual load tests, assigning such factors as the number of Vusers involved, the scripts to be run in the scenario, and the frequency of requests. VuGen is responsible for the creation of usage scripts that reflect the behavior of a real user as closely as possible. The Analysis module interprets test results to find performance bottlenecks and other problems. The Tuning module assists in isolation of bottlenecks, finding the exact cause behind the slowdown.
Through Controller configuration, the user can run all kinds of performance testing on an application. Component testing allows the engineer to place load on a specific module of the application. General load testing places the application under the load expected in real-world situations. Stress testing finds the breaking point of an application, subjecting it to a simulated spike in activity. Volume testing allows the engineer to simulate a prolonged period of high activity.
LoadRunner is frequently used in conjunction with Web based servers and database systems.