Considerations for selecting a topology

Selecting an appropriate topology for your deployment environment depends upon several factors.

When you select a topology pattern, consider the following factors:

Available hardware resources
Application invocation patterns
Types of business processes that you plan to implement (interruptible versus non-interruptible)
Individual scalability requirements
Administrative effort involved

The Application, Remote Messaging, and Remote Support topology pattern is the preferred topology for IBM® Business Process Manager Standard and IBM Business Process Manager Advanced, but the choice ultimately depends upon your individual requirements.

The IBM-supplied topologies can be applied to Process Server, Advanced-only Process Server, and Process Center topologies. Therefore, your Process Center, Advanced-only Process Server, and Process Server network deployment environments can be organized in a similar way. The procedures for creating environments for Process Server, Advanced-only Process Server, and Process Center based on IBM-supplied topologies are also similar. The only difference related to IBM-supplied patterns is the recommended patterns for a production environment, and the components configured on the clusters for those patterns.

For information on the components, features, and functionality available in each of the IBM Business Process Manager configurations, see IBM Business Process Manager overview

Condensed topology pattern selection criteria

Consider the information listed in the following table, which is a quick guide to selecting your production topology. This table provides a condensed list of the advantages and disadvantages of each of the topology patterns.

For information about which BPM products support the supplied topology patterns, see Topology patterns and supported product features.

Table 1. Considerations for selecting a topology for your deployment environment
Consideration	Topology Pattern
Consideration	Single cluster	Application, Remote Messaging, and Remote Support
Number of clusters to maintain	One cluster for all components	Three clusters: One cluster for applications One cluster for support infrastructure One cluster for messaging
Hardware requirements	Can be implemented on limited hardware	Most hardware intensive
Asynchronous interactions	Use should be minimal	Ideal environment for asynchronous interactions
Long-running processes, state machines, and human tasks	Use should be minimal	Ideal environment for interruptible processes, state machines, and human tasks
Administrative burden	Relatively small	Requires most administrative effort
Scalability	All components scaled at the same rate	Easiest to scale All functions separated Messaging cluster scalability still limited (benefit comes when other BPM products are introduced)