ATG Campaign Optimizer is designed to produce the most efficient scenarios possible to support heavy traffic. To minimize the performance impact of running tests, certain features exist in ATG Campaign Optimizer to limit the number of events generated in a system, thereby improving the overall performance of the system. Testing shows that the drop of performance for a system running ATG Campaign Optimizer is between10-20% for up to 10 concurrent running tests with 2 groups each and four elements each. The expectation is that most sites will likely have no more than 4-6 concurrent tests.
One way to improve application performance is to reduce the number of events that are fired. ATG Campaign Optimizer provides three mechanisms for doing so:
MIME type filtering prevents events from being fired for non-page requests, such as requests for images. See Substitution Mime Type Filtering.
Each request for a template, page, page fragment, or slot fires an event in an attempt to determine whether there’s a substitute for the requested item and, if so, to locate that substitute item. This operation produces an item (either the original or the substitute), which is cached for use by subsequent requests. See Substitution Caching.
A test subject registry keeps track of the items affected by active tests and prevents events from being fired when users request items that aren’t used in tests. See Test Subject Registries.
By default, ATG Campaign Optimizer is configured to use these three performance-enhancing mechanisms. It’s unlikely that you should need to modify or disable any of them. ATG Campaign Optimizer also provides a cache for log entries that improves the performance of the reporting module. You may need to adjust the size of the cache; see Changing Logging Cache Sizes for details.
