Expanding your topology
There are three ways to expand your topology: adding members to a cluster, adding cells, or adding deployment environments.
Adding cluster members
The easiest way to expand your infrastructure is to add more cluster members to your existing clusters. You can add cluster members to each cluster independently or in combination, depending on where you see the need for growth. In a single-cluster topology, you can expand the cluster (which provides the application, support, and messaging capabilities). In a three-cluster topology with separate application cluster, support cluster, and messaging cluster, you can add new cluster members for any one, any two, or all three clusters. By expanding your clusters in this way, you can improve your application throughput.
- You plan to deploy new applications to your existing environment.
- You anticipate increased volume requirements for your existing applications.
- You need more capacity for operational purposes, such as failover.
- You have applications or sets of applications that serve different business purposes and you want to keep them distinct. If you deploy these distinct applications to the same deployment environment, you might introduce dependencies between otherwise unrelated business domains. Dependencies can affect such things as maintenance schedules and application availability when, for example, unrelated applications are less reliable.
- After you analyze your performance characteristics, you realize that planned deployments might push your system beyond the limits of the current deployment target. They might introduce too many modules from new applications or new versions of applications to run in the existing memory space, or use your database tables too heavily.
Adding cells
If you decide that expanding your existing clusters is not an appropriate solution, consider creating another deployment environment in another cell. This approach gives you the most room for growth, the most flexibility for expanded functional requirements, and complete isolation for your applications.
To add another cell, you duplicate common cell-level configurations, such as global security settings. You use multiple consoles to manage your applications, such as separate administrative consoles and failed event managers.
Adding deployment environments
If the previous two options do not satisfy your requirements, there is a third alternative for expanding your topology. You can create multiple independent deployment environments for your applications in the same cell.
- You want to reuse complex configurations at the cell level. For example, you want to set up multiple test environments quickly without having to configure security or nodes each time.
- You want the convenience of a single IBM® Process Center and a single administrative console to manage multiple IBM Process Server servers.
- You want to avoid the larger footprint associated with adding cells.
- You must expand the capacity of your environment but adding cluster members is not an appropriate solution.
- You want to isolate process applications into separate Java virtual machines (JVMs), associating each application group with a specific set of JVMs.
The first deployment environment in the cell is typically for Process Center. You need one Process Center per cell or else you lose the single repository experience. Subsequent deployment environments are for Process Server only. When you create additional deployment environments after creating the first deployment environment, the cell-scoped configuration artifacts already exist and they will be reused because they are shared. You create the additional deployment environments in the same way that you created the first deployment environment.
Running multiple deployment environments in the same cell is an advanced topology that requires research and planning.