Entity managers

Entities and entity types

An entity represents the logical relationship among two or more member records. Each unique relationship has a corresponding entity type such as identity or household. For example, 10 records that describe the same person is an identity entity. Three records that describe the same household is a household entity. An entity type is the basis for any algorithm configuration. In creating an algorithm configuration through InfoSphere® MDM Workbench, you create its associated entity type first.

Entity management

Entity management is the process by which record comparison, data linkage, and task creation occur. The management settings affect how entity type comparisons are conducted after member data is derived. There are two primary methods:

• Synchronous management. The management for a synchronous entity type happens immediately after a member record is updated. The update is done through a put transaction. Before the transaction occurs, the target member record is cross-matched against multiple candidates. As a result, synchronous management occurs based on built-in logic and without any entity management processes. It is unaffected by external processes, input queues, and any polling interval and work unit settings for an entity type.
  Note: If your deployment uses synchronous entity management, then dynamic frequency based bucketing (DFBB) is not possible. For more information about frequency based bucketing, see Enabling frequency based bucketing.
• Asynchronous management. In contrast, the management for an asynchronous entity type happens based on member records that are in an entity manager input queue (that is, the mpi_entique_xx table, where xx is the applicable entType, or entity type, code). The put transaction creates records for the queue; an asynchronous process queries the input queue and completes the cross match for the target member record. The data is stored in the database, and comparison and linkage and task management are processed periodically. If you use asynchronous entity management, you also have the option of further controlling when entity management occurs by priority.
  • Priority management: There are two ways in which you can override your normal asynchronous entity management process. Priority entity management manages the process either by source priority or by set priority. Priority is set in InfoSphere MDM Workbench.
    • With entity management by source priority, you set a default entity priority for definitional sources that controls the order in which the sources are processed by the entity manager. The lower the source priority, the higher the entity management processing priority. For example, a source with a default entity priority of 1 always has entity management run before other sources. Source priority entity management is disabled by setting all of your sources with the same value (the default value is 100).
    • For entity management by set priority, you set the entity management priority at the time of the member write (this option overrides any source entity management priority setting). For example, you can set a lower priority for batch-loaded members. The set priority is used with the member put interactions.

Within the InfoSphere MDM Workbench Configuration editor, you set the "Asynchronous", "Uses an input queue" and "Priority" properties to determine whether an entity type is processed synchronously or asynchronously.

• Asynchronous
  • true for asynchronous processing (default).
  • false for synchronous processing.
• Uses an input queue
  • true for asynchronous processing.
  • false for synchronous processing.
• Uses an output queue
  • true for providing input to event processing.
  • false when event processing is not enabled.
• Priority
  • By Source - an integer 1 - 32767 in the Default Entity Priority field.
  • By Set - defined at the API interaction level (an integer 1 - 32767); not in InfoSphere MDM Workbench.

For entity management, it is possible to implement one of the scenarios described in this table.