A Delay-Driven Congestion Control Method Based on AQM Algorithm

Abstract

The invention discloses a time-delay-driven AQM-based congestion control method. Firstly, reference parameters are set, and when a data packet arrives, the data packet enters the state of waiting to enter the queue, and the priority weight of the currently arriving data packet is obtained. w, the length of the current packet queue q l , the flow id of the currently arriving packet and the discharge rate of the current packet queue o r , then set the basic drawing factor of the current data package, and update the basic drawing factor, and then get the priority drawing factor of the current data package based on the updated basic drawing factor and priority weight, and then based on the priority drawing factor, the current data package The queuing delay q of the queue d and delay threshold q th Determine whether there is congestion in the current network, and determine whether the currently arriving data packet enters the current data packet queue according to the conclusion of whether there is congestion; the advantages are low delay in queuing, low flow completion time and balanced fairness.

Description

technical field [0001] The invention relates to a congestion control method, in particular to a delay-driven congestion control method based on an AQM algorithm. Background technique [0002] The AQM (Active Queue Management) algorithm is a method for managing buffer queues, and is usually used for network congestion control. A router deployed with a congestion control method based on the AQM algorithm has a relatively advanced scheduling and queue caching strategy. Instead of dropping packets when the queue is full, it discards packets before the router queue is full and provides a congestion control strategy. instruct. The congestion control method based on the AQM algorithm uses packet loss to indicate congestion, which can reduce the impact of packet loss and retransmission, and reduce network delay. [0003] The RED (Random Early Detection) algorithm, as an upgraded algorithm of the AQM algorithm, uses the average queue length as a congestion indicator to guide active...

AI Technical Summary

Problems solved by technology

However, when a large number of TCP flows and UDP flows coexist, they will compete with each other for bandwidth

If the arriving data packet has the same flow id as the extracted data packet, that is, both of them come from the same flow, both of them will be discarded. Due to this unreasonable data packet discarding, it will cause a large queuing delay and a large number of Packets are timed out and retransmitted, inevitably making the flow take longer to complete