In this multilevel feedback queue this is the length of time a process from the second queue can spend in the CPU before being preempted.