Hello Jamal and others, I was working on new faster and scalable HTB algorithm. I'm happy to share my last test results: I created hierarchy with 1000 leaves borrowing from common bounded parent. I measured with total 30 Mbit of data flow over 5 seconds. Here are total times in ms spend in qdisc dequeue routine: CBQ 4450 ms Old HTB 4630 ms New HTB 455 ms Times are varying for CBQ with flow's character, lowest times were close to new HTB but I have to say that at these tests CBQ's output was totaly useless (no bounding, large errors in rate ..). I'm going to sleep just now but I'll release new HTB code for wider testing tomorrow. It will work for 2.4 kernels only until someone backport it ;) It is because it uses 2.4 rb_tree functions. New code has guaranted dequeue and enqueue complexity O(log(N)) where N is number of active classes. I hope I haven't done error in measurement ;) devik
