Single Cluster topology pattern
The Single Cluster topology pattern is an IBM-supplied topology pattern. In a Single Cluster topology pattern, all the functions of the deployment environment are combined into a single cluster.
A Single Cluster topology pattern is ideal for limited hardware. Because all the components are installed in the same cluster, fewer physical machines are required. However, because each server instance must run the supporting applications and your integration applications, you need more memory for the individual Java Virtual Machines (JVMs). In addition, one or more members of the cluster must also run the messaging engines required for asynchronous interactions. Thus, the Single Cluster topology pattern is typically used for proof of concept, development, and testing environments.
- Because asynchronous interactions (involving JMS and MQ/JMS bindings), human tasks, state machines, and long-running business processes can make extensive use of the messaging infrastructure, a single cluster environment is not ideal for applications with these components.
- Any messaging requirements must be kept to a minimum with this topology pattern (except for z/OS).
- Service Component Architecture (SCA) internal asynchronous invocations, the Java™ Message Service (JMS), and MQ messaging bindings do not support multiple messaging engines in the same cluster.
The Single Cluster topology pattern is suitable for scenarios that are focused on running applications and on synchronous invocations.
From an administrative and scalability perspective, the Single Cluster topology pattern has advantages. A single cluster where each member runs all the IBM Business Process Manager components are simpler to administer. Instead of several server instances in multiple clusters, you have a single cluster with fewer members. If the needs of your environment grow, scaling the infrastructure is a simple matter of adding additional nodes and cluster members. Thus, the process of adding capability is simple, but all components are scaled at the same rate. For example, if the messaging engines spread across server members use policies, there could be some additional administrative effort in creating and maintaining the policies.
In a Single Cluster topology pattern, all deployment environment functions and functional groups of components run on a single cluster:
- The applications:
- Business Process Choreographer (BPC) container
- Process Server
- Business Rules manager
- Business Space
- Process Portal
- REST API Services
The messaging infrastructure cluster hosts a single service integration bus and single messaging engine that use the same database schema as the product database by default. Each deployment environment has its own bus. The single bus is called BPM.deployment_environment_name.Bus.
- The support infrastructure applications:
- Performance Data Warehouse
- Business Process Choreographer Explorer
