Round Robin (Computer Science)

from Wikipedia, the free encyclopedia

The term round-robin method , English Round-Robin , refers to a scheduling process, a process executing the queue.

It is used, for example, as a process scheduler , where it allocates limited execution resources to several competing processes. The round robin procedure grants all processes successively for a short period of time during a time slot allocation to an executing CPU ; this is also called arbitration .

Round-robin is also used for load sharing ( load balancing used). The aim of load distribution is to use several resources of the same type as evenly as possible.

realization

Example process scheduler

The processes are managed in a queue . The first process gets access to the resources for a time slot, then it joins the end of the queue and all processes move up one position. The next process is selected according to the FIFO principle. The process can also voluntarily return the resources earlier. Even if a process is completed before the end of its time slot, the resources are reallocated immediately.

Example scheduling multitasking operating system

In operating systems with preemptive multitasking , the scheduler creates an execution plan for the active processes using the round robin method. After each time slot, it then uses a queue to determine which process comes next. The dispatcher then assigns the processor to this process for a time slot.

Example load sharing domain server

Round robin z. B. used in the Domain Name System , where a name server can deliver several IP addresses on request . For load balancing on large websites or IRC networks , this is done on multiple physical servers .

Example of load distribution routing

Routing protocols such as B. Routing Information Protocol (RIP) use the round robin method for load distribution on different lines (routes). Routes with the same node metrics and the same destination network are sequentially loaded for packet delivery. This happens either for each forwarded packet individually (per packet) or for each new destination host (per destination) .

rating

For the criteria on which this evaluation is based, see Scheduling, “Criteria” section .

Round robin treats all processes equally, so that on the one hand no process is treated unfairly or even starves to death, but on the other hand it is also not possible to preferentially process processes with higher urgency . The throughput of this scheduling method is generally neither particularly low nor particularly high. The use of fixed length time slots makes round robin inflexible.