We propose to add a configuration property in the JPPF driver, to specify that it should route jobs to other drivers only when its number of nodes is at most a given threshold. For example, if we define:

jppf.peers.load.balance.threshold = 3

then the driver will only route jobs to peer drivers when it has less than 3 nodes. The default value would be infinite (say Integer.MAX_VALUE) to preserve the current default behavior. A value of 0 or less would mean that jobs are never routed to a peer driver.