IM Relationship Resolution Information Center, Version 4.2

Product architecture

IBM Relationship Resolution is a multi-tier system, in which data from data sources is loaded into the system from acquisition programs and processed by the pipelines that are hosted by pipeline nodes. The results of the processing are written to the entity database and can be routed to other systems or other databases.

In a typical deployment, enterprise data from multiple data sources is sent to acquisition programs, where the data is transformed into Universal Messaging Format (UMF). Each acquisition program uses a transport to send the data into one or more pipelines. Many of these transports are bi-directional, and the system can be configured to provide the acquisition program with responses.

IBM Relationship Resolution Product Architecture diagram

One or more pipeline processes run on pipeline nodes. Each pipeline maintains its own connection to the entity database. As the pipeline receives the UMF data from one or more acquisition programs, it processes the data record-by-record through its three core processes: recognize, resolve, and relate. As each record is processed, the pipeline stores the results of the processing in the entity database.

Users interact with the system using these interfaces:

IBM Relationship Resolution uses an embedded version of IBM WebSphere Application Server. This application server hosts the Configuration Console, the Visualizer, and the Web services.

The system also features an application monitor database. The application monitor database is used to store configurations for the nodes that are registered for system monitoring and routing.

This robust architecture provides scalability for any deployment. Pipelines can be deployed on any number of small or large machines. Pipeline performance can scale to any desired level, given enough database capacity.



Feedback

Last updated: 2009