The round-robin scheduling policy was designed specifically for fairness in time-sharing systems.

A small unit of time constant, called a time quantum (or quanta), is pre-defined. It normally lies in the range of 10 to 100 milliseconds. The queue of processes is essentially a circular queue, and the scheduler goes round the queue, allocating the CPU to each process for a time interval or up to 1 quanta. After a process has used up its time quantum, a timer interrupt will occur and the scheduler will pick the next process in the queue to run. If a process completes in less than 1 quantum of time, the process itself will release the CPU. If there is only one process, then that process may get a series of time quanta in which it is allowed to run.

Fig. RR-1

The RR method of scheduling is preemptive, and reduces the total efficiency of CPU because there is a switch time delay when switching from one process to another.