Process Federation Server components overview
Terminology
The terms federated system or federated instance can refer to either an IBM Business Process Manager or a Business Automation Workflow server that is part of the federated environment.
Architecture
Process Federation Server provides end users with a unified view of tasks, process instances and case instances running on different Business Automation Workflow and Case Manager federated systems.

REST APIs
- retrieving information from federated systems through the retrievers (Launch Entities and Systems Metadata REST APIs).
- performing searches into the Federated Data Repository indexes to return lists of tasks, process instances and case instances (Tasks List and Process Instances list REST APIs)
- storing saved searches in a Federated Data Repository index or a Process Federation Server database (deprecated) (Saved search REST API).
- storing reusable queries in a Federated Data Repository index (Reusable queries V2).
- returning data for the Team Performance Dashboard.
Indexers and retrievers
For each federated system to configure, two components must be configured: an indexer and a retriever.
- Indexers
-
The indexer inserts JSON documents that describe all the instances and/or tasks that run on the federated system into a Federated Data Repository index. Each federated system has its own single index, and depending on the version and on the type (BPD, BPEL or Case) of the system that you want to federate, the indexer runs either on Process Federation Server or on Business Automation Workflow nodes.
Process and BPEL indexers:- constantly poll the federated system’s database to identify which tasks and process instances must be indexed into the Federated Data Repository.
- perform SQL queries to the database to build the document that will be indexed into Elasticsearch for each identified task and process instance.
- perform maintenance operations on the database tables involved in the indexing.
- in Business Automation Workflow 24.0.0.0, the process indexer was redesigned and is now configured and runs on the Business Automation Workflow federated system. For older process versions and for BPEL, the indexer is still configured and runs as part of Process Federation Server.
The Case indexer indexes Case instances into the Federated Data Repository.
- To federate a system hosting process instances and tasks, you
must define the indexer either in Process Federation Server or in Business Automation Workflow, depending on your
version of Business Automation Workflow:
- For Business Automation Workflow 24.0.0.0, there is no indexer to configure in Process Federation Server. Instead, Business Automation Workflow must be configured to index directly in the Federated Data Repository as documented in Configuring Business Automation Workflow to index to the Federated Data Repository.
- For Business Automation Workflow 23.0.2
and earlier, you must configure the indexer in the Process Federation Server configuration
through an
<ibmPfs_bpdIndexer>element. For more details, see Enabling indexing of process-related data in a federated environment.Note: Federating a Business Automation Workflow 24.0.0.0 process system by configuring the<ibmPfs_bpdIndexer>element in Process Federation Server configuration is deprecated. However, you can upgrade your Business Automation Workflow federated system to 24.0.0.0 as usual, and then disable the indexer in Process Federation Server to replace it with the new Federated Data Repository process indexing as documented in Enabling the Federated Data Repository process indexing.
- To federate a system hosting BPEL process instances and tasks, you must configure the indexer in
Process Federation Server
configuration through a
<ibmPfs_bpelIndexer>element. For more details, see Enabling indexing of BPEL-related data in a federated environment. - To federate a Case Manager system, you must configure the indexer in Case Manager configuration as described in Indexing case instances.
- Retrievers
- Retrievers perform requests to the federated systems, and store into the Federated Data Repository cache data such as:
- the system status
- the list of launchable entities
- the teams memberships of users
The retriever is always configured as part of Process Federation Server:- To federate the BPMN (BPD) process instances and tasks hosted by a Business Automation Workflow system, you must
configure the retriever through a
<ibmPfs_bpdRetriever>element. For more details, see The ibmPfs_bpdRetriever element. - To federate the BPEL process instances and tasks hosted by a Business Automation Workflow system, you must
configure the retriever through a
<ibmPfs_bpelRetriever>element. For more details, see The ibmPfs_bpelRetriever element. - To federate a Case Manager system, you must
configure the retriever through a
<ibmPfs_caseRetriever>element. For more information, see The ibmPfs_caseRetriever element.
External components
- The Federated Data Repository. The
following Federated Data Repository services
are supported:
- Elasticsearch 7.x - all versions from 7.17.0. (note that support for Elasticsearch 7.x is deprecated.)
- Elasticsearch 8.x - all versions.
- Elasticsearch 9.0 to 9.1.x.
- OpenSearch 2.x - all versions from 2.5.0.
- Opensearch 3.0 to 3.1.x.
- Two additional indexes are used to store resuable queries and saved searches.
- (Deprecated) Relational database to store saved searches. All Process Federation Server instances share the same relational database.