The Role of the Object Table

Preloading controls are available for entity objects, and for entries (handles) in the object table (objectTable) cache relating to entity objects.

The objectTable cache is a large cache that provides mappings for all entities in the cache. The object table contains the object id and information about the object such as its extId, class name, type, backing store table name and cache name. The object table is used to find the actual object either in the cache or in the backing store. The object table is also written to the backing store.

The object table can grow become very large, up to hundreds of millions of rows in extreme cases. You can preload the cache cluster’s object table at system startup to improve performance after the system has started up. (See Domain Objects Configuration in TIBCO BusinessEvents Configuration Guide for details.)

The following figure shows how the object table in cache is used at runtime.

Use of ObjectTable at Runtime
  1. A rule requests an object X to be fetched using its internal ID. The object is not in the Rete network.
  2. The object is not in the local cache.
  3. TIBCO BusinessEvents looks up the ID in the object table and gets its cache name. Its cache table is not found in the cache cluster.
  4. TIBCO BusinessEvents gets the name of the object’s backing store table from the object table, locates the object’s table in the backing store, and returns the object to the calling function.

Note that if the object table is not preloaded with the entry for object X, then an additional processing step is required, to load the object table in the cache for that object’s entry.