Understanding Entity Caches

Understanding Entity Caches
This appendix is provided for those who want to understand the internal structures of the Oracle Coherence caches used in Cache object management. This information is not required for configuration tasks.
For each entity in working memory, a corresponding cache exists in the cache cluster. Internal entities also have caches for various purposes, explained in this section.
Entity Cache Names Format
Each entity cache has a name, which uses the following format:
cache-type.cluster-name.AgentClassName.entity-name
The elements of the above name are explained below
Cache Type (Caching Scheme)
Cache type is the type of caching scheme (as defined by its cache name in the coherence-cache-config.xml descriptor), for example, dist-unlimited-bs.
Cluster Name
Cluster name is the value of the following property:
java.property.tangosol.coherence.cluster
Agent Name
This field of the cache name is blank because TIBCO BusinessEvents does not support agent-specific entity caches.
All entities are globally scoped and available to all agents.
Entity Name
Two types of entities have caches:
Internal entity names and caches are listed and described in Table 51, Internal Entity Caches.
The ontology entity field of the entity cache name uses the entity’s generated class name, which is similar to its design-time folder path and name, prefixed by be.gen. For example:
be.gen.Concepts.LargeConcepts.ThisLargeConcept
Caches for Ontology Objects
These caches are used to store the objects of types defined in the ontology of the project.
The types of caches created for ontology objects depend on the caching scheme used. If the dist-unlimited-bs caching scheme is used, then the cache names look like this:
dist-unlimited-bs$foo$$be.gen.Order
Where foo is the cluster name.
Caches for Internal Entities
The following internal caches use a pre-defined scheme in the cache configuration file. Do not change this scheme. This information is provided for reference only.
Stores the mapping between type IDs and class names. All ontology objects are tagged with a unique integer ID. Use of IDs avoids the need to serialize and send class name strings between nodes.
AgentTxn-agentId
Each agent in the cluster has an AgentTxn-agentId cache. The agentId is internally generated. It stores the change list for the agent. The change list is used to replicate changes between active-active and active-passive sets of agents in the cluster so that they stay synchronized.
Maintains all entries that are time bound, for example, state machines that can have timeouts at a state machine level or at a state level. This cache maintains an index to all objects that must be re-evaluated after a certain period of time.