Weighted Fair Queuing (thing)
Return to Weighted Fair Queuing (thing)
|I just can't believe that this node isn't widly popular.|
A flow-based queuing algorithm designed for predictable traffic throughput and response time on slow-speed links. It dynamically sorts traffic into conversations based on packet header addressing, then breaks up the "train" of packets within each conversation to ensure fair bandwidth allocation.
Okay, drink and snack break!
Welcome back. WFQ gives priority to low-volume traffic such as Telnet sessions (which you shouldn't be using anyway), and fairly allocates bandwidth between high-bandwidth flows. In other words, it would give two equal-sized file transfers an equal amount of bandwidth. Isn't it wonderful?
But just like people who chew with their mouths open or swear too much, WFQ has its disadvantages, one being its lack of scalability. If traffic on the network were to increase significantly, WFQ would be ineffective. Also, it's not available on high-speed connections over E1 speeds (2.048 Mbps). There is, however, class-based weighted fair queuing (CBWFQ) available on ATM.