Research On Load Balancing In WLAN Based On Historical Traffic Data And Table-lookingup Method

With the mass deployment of wireless local area network (WLAN) in airports, schools, hotels and other public hotspots, people can access the network and get network services anytime, anywhere. Due to the layout of the access points (AP) is unreasonable, terminal distribution is not controllable and other reasons, it often leads that load distribution in WLAN is not uniform. That the uneven distribution of network load in the WLAN can cause network congestion, which can lead that the network performance decrease, the quality of the service cannot be guaranteed, the user experience is not good and other issues. Thus, it is very important to research on load balancing technology in WLAN.Existing load balancing technology in WLAN has many problems. Some algorithm is simple and easy to implement, but its load balancing effect is not ideal. Although load balancing effect of other algorithm is good, but the algorithm is complex, it is difficult to realize and it takes much network resource and computing resource. Based on the simplest and most common periodic law of things, it uses historical traffic data of the internet in the decision making process of load balancing, so that the load balancing algorithm has advantages that it is simple, the effect of the algorithm is satisfactory and the cost of network resource and computing resource is less.This paper mainly includes the following aspects:(1) Research on algorithm reliability. Natural phenomena are periodic, which has inspired the author of the paper to realize that the flow of people is generally cyclical changes and so the people’s demands for network services is also cyclical changes correspondingly. On this basis, this paper presents a load balancing scheme which is based on historical traffic data and table-lookingup method in WLAN.(2) Research on specific steps of the algorithm. The load-balancing scheme requires to gather network traffic data over a period of time, and process these data to generate tables. The system predicts the network traffic in any time period by the data in the tables. The system calculates the threshold and set for each AP with the predicted value of the network traffic. The new terminal allows access to AP whose network flow is less than the threshold and the new terminal is not allowed access to AP whose network flow is more than the threshold.(3) Research on Algorithm Improvement. The algorithm is improved from three aspects, such as the generation of table data, the correction of table data and the selection of AP. The improved algorithm has the advantages that the data redundancy is low, the prediction is more accurate, and the load balancing effect is better.(4)Experimental simulation and Research on algorithm performance. The basic type and the improved type algorithms are simulated and analyzed with different load-balancing measures. The results of the simulation show that the algorithms can obtain very satisfactory load-balancing effect with taking a small amount of network resources and computing resources when the network traffic forecasts is accurate. Compared with the basic type of algorithm, the improved algorithm can achieve better load-balancing effect.