I understand what WRED is and what goals it aims to achieve during congestion (drop packets, avoid Taildrop and so on).

However, the actual principle of it is quite confusing to me. Consider this graph from NetworkLessons.com enter image description here

Let's say that this is a queue for the "TCP" class and it has a certain amount of bandwidth reserved and everything necessary for it to work.

Why are we dropping everything once we reach an average of 45 packets in the queue, despite the queue not even being full yet? That doesn't seem very efficient to me. To me, it looks like we just have a bunch of space in the queue that will never be used because of this.

1 Answer 1


The basic algorithm is called random early detection - it is designed to drop before the queue completely fills up. Since RED is often only applied to part of the traffic, the non-RED (i.e. in this case, non-TCP) traffic can use the remainder of the interface queue.

With weighted RED, the remainder can also be used with traffic in higher QoS classes (which may have a higher queue threshold and/or a lower drop probability).


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.