Although you have the option of assigning workload to individual agents, assigning workload to clusters of agents is often preferable. Clusters are groups of agents that can share the same capabilities, such as a Java Development Kit (JDK) installation, an ODBC client, or a Citrix client, or can consist of agents that possess entirely heterogeneous capabilities, such as the capability to divide a lab's machines into separate agent groups that can support different teams. With this approach, you select a cluster of agents capable of executing the required workload when you configure your test's workload. Silk Performer then handles the dynamic assignment of workload to specific agents. Only those agents that are required to deliver the workload are actually used.
Agent clusters are managed centrally in Silk Central and are retrieved by Silk Performer when you start a load test—you do not need to configure the availability of agents yourself and conflicts with concurrently scheduled load tests are resolved automatically.