Case Study Assumptions
This case study examines a portion of the performance test cycle for a fictitious e-Commerce company. We decided to use an e-Commerce web site for several reasons.
The high graphical content of their pages tends to stress the network. Many retail sites receive sustained bursts of traffic at the holiday season, which requires us to plan for these peak load periods. These sites tend toward higher complexity with various databases on the back end and some special servers to assist with graphic content on the front end.
In short, e-Commerce web sites give us the broadest demonstration of the concepts we discuss in the book.
We decided to make our fictional e-Commerce company fairly large. As you'll notice when you read the case study, our fictional retailer receives 100,000 users per day during the peak season. That's a significant amount of traffic for an e-Commerce retailer, and does not apply to most of the e-Commerce sites we see. On the other hand, 100,000 user visits per day is not the largest site we've ever encountered, and certainly does not put our fictional web site among the Internet's heavy hitters.
A large web site allows us to better demonstrate some of the techniques of test planning more thoroughly (and on a grand scale). It also allows us to illustrate a very important point: Don't be afraid of large numbers. You can develop manageable and correct performance tests, even for large web sites. This case study addresses the highlights of a performance test cycle. Obviously, we can't cover every potential issue, or even address in depth some of the issues our fictional team uncovers. However the case study gives you a feel for an actual test cycle.
|
