This image shows a configuration that splits the request processing workload between two servers (that is, S1 and S2). A single JADE node is running in each server. Although only two copies of the Web application are shown for each node, it is more usual to run two or three copies for each logical CPU, depending on the resource usage for each request. Each node (and therefore each server) can have up to 200 outstanding requests.
You could extend this configuration by running multiple nodes in one or more of the servers.
