Dynamic caching

In general, caching improves response time and reduces system load. Caching techniques are used to improve the performance of World Wide Web Internet applications. Most techniques cache static content (content that rarely changes) such as graphic and text files. However, many websites serve dynamic content, containing personalized information or data that changes more frequently. Caching dynamic content requires more sophisticated caching techniques, such as those provided by the WebSphere Application Server dynamic cache, a built-in service for caching and serving dynamic content.

The caching behavior of the WebSphere Application Server dynamic cache service is specified by cache policies that are defined by <cache-entry> elements in cache specification configuration XML (cachespec.xml) files.

As the dynamic cache service places objects in the cache, it labels them with unique identifying strings (cache IDs) constructed according to <cache-id> rules specified in the <cache-entry> elements. Once an object with a particular cache-id is in the cache, a subsequent request for an object with the same cache-id is served from the cache (a cache "hit"). The <cache-id> rules define how to construct cache-ids from information that is associated with an application server request (to execute a servlet, JSP, or command). This definition includes how information might be obtained programmatically from CacheableCommand objects.

Cached objects are removed from the cache according to information provided in their <cache-entry> elements, such as the<timeout>, <priority>, and <invalidation> elements.

The <timeout> and <priority> elements configure expiry and eviction policies. When the available cache memory is full, a least recently used (LRU) caching algorithm removes cached objects with lower priority. The algorithm can also offload cached objects with lower priority to disk if the disk offload capability is enabled before those with higher priority.

The <dependency-id> and <invalidation> elements define rules that generate dependency IDs and invalidation IDs. Together, these IDs specify that certain objects should be removed from the cache when certain requests (such as requests that update cached information) are processed. When an object is cached, its generated dependency IDs are associated with it in the cache. When a request causes invalidation IDs to be generated, all objects that are associated with those invalidation IDs are removed from the cache.

The <inactivity> element is used to specify a time-to-live (TTL) value for the cache entry that is based on the last time that the cache entry was accessed. The value is the amount of time, in seconds, to keep the cache entry in the cache after the last cache hit.

The dynamic cache service responds to changes in the cachespec.xml file. When the file is updated, the old policies are replaced. Objects that are cached through the old policy file are not automatically invalidated from the cache. The objects are either reused with the new policy or eliminated from the cache through its replacement algorithm.

WebSphere Commerce uses WebSphere Command Caching internally such as with MemberGroupsCacheCmdImpl in the preceding cache filter; however, WebSphere Commerce does not support caching of commands that contain non-serializable objects.