If Preemption is enabled, higher-priority applications do not have to wait because lower priority applications have taken up the available capacity. Under-served queues can begin to claim their allocated cluster resources immediately, without having to wait for other queues' applications to finish running.