QoS traffic scheduling is a scheduling methodology of network traffic based upon QoS (Quality of Service). Here, the frames or packets are mapped to internal forwarding queues based on its QoS information, which are then services according to a queuing scheme.
Typically, multiple queues are present each with different priority levels. The scheduler decides the type of treatment to be given to the traffic in each queue. When traffic is available, the scheduler maps it to the appropriate queue. For example, video and voice traffic are kept is queued with higher priority than background traffic.
Weighted Round Robin (WRR) − In this method, frames or packets of all the queues in the scheduler are serviced in each cycle. Priority among the different queues is maintained by forwarding a specific number of frames in each queue per cycle in a rotational manner. For example in a system with four queues 𝑄0, 𝑄1, 𝑄2, 𝑄3, 5 frames of 𝑄0, 3 frames of 𝑄1, 1 frame each of 𝑄2 and 𝑄3 may be sent in each cycle. This assigns the highest priority and consequently better traffic flow to 𝑄0 while not starving low priority traffic infinitely.
Strict Priority (SP) − This method imparts the highest service to high priority traffic. Here, the queuing mechanism forwards as many frames as possible in a higher priority frame before moving to the queue with the next priority level. For example in a system with four queues 𝑄0, 𝑄1, 𝑄2, 𝑄3, with decreasing priority, strict priority method with send as many frames as permissible of 𝑄0 before moving on to 𝑄1.
Combination of WRR and SP − This queuing method is configurable in nature and it combines both WRR and SP. In this method, strict priority is given to time-sensitive or real-time traffic like voice and video, while WRR is adopted for other traffics.