The Research About Adaptive Active Queue Management Algorithm

With the advent of technology, Internet has got a huge development, which plays a vital part in people’s daily life and work. While network resources are increasingly abundant, the structure of the network becomes more and more complex. Now, the rapid development of Internet technology brings various features to the IP services, which causes the network congestion. Network congestion dramatically brings big trouble to people, which maybe increase the packet loss rate, increase transmission delay, degrade the throughput,affect the overall stability of the network seriously and cause the drop of the Quality of Service. So the Active Queue Management algorithm is becoming a hot point in the field of network congestion, which becomes one of the best tools to improve the network congestion and the quality of network service.Some classical active queue management algorithms are studied systematically in this thesis. In order to reduce packet loss probability of router algorithm and improve the stability of queue length, the adaptive active queue management algorithm H-ARED which based on fuzzy logic, is put forward. First, this thesis optimizes the calculation function which makes full use of the buffer size of router. Second, according to the current network conditions, the new fuzzy logic controller, which adjusts the biggest packet loss probability dynamically, can achieve the purpose that the queue length is stabilized, the network throughput is improved, packet loss probability of network is reduced and the link utilization is raised. Finally, a large number of simulations between H-ARED and ARED are made, using the NS2 tools, which shows that H-ARED has better stability of queue length, reduces the packet loss rate to some extent,improves network throughput. So H-ARED can better adapt to different network environments, control network congestion effectively and H-ARED has better stability and robustness.