This image shows a configuration that splits the request processing workload between multiple copies of a Web application running in separate JADE nodes within the same machine. This configuration has less resource contention than a configuration that has all Web application copies running in the same node.
You can use this configuration when a large number or requests must be processed in parallel but the local resource usage (for example, CPU) per request is low, in which case the number of copies for each node will be much higher (for example, 30 to 40). In addition, the outstanding request limit will be high (in this example, 500 for each node).
