WLAN中基于“小区呼吸”的多约束负载均衡

Multi-Constraint Load Balancing Based on Cell Breathing in WLAN

在无线局域网(WLAN)中,负载不均衡会降低网络吞吐量、增加分组传输时延.由于无需修改客户端的优点,研究者们将码分多址(CDMA)网络中的"小区呼吸"概念引入WLAN以实现负载均衡."小区呼吸"方法通过改变接入点(AP)的信标功率,从而改变AP的覆盖范围,进而控制AP的用户数.现有的同类方案在使用"小区呼吸"方法时,大多没有考虑两个比较重要的问题:AP的负载均衡与用户数据功率损失的矛盾以及AP的业务可用性(包括AP业务欺骗和业务漏洞).文中将这两个问题作为使用"小区呼吸"进行负载均衡的约束,首先对多约束负载均衡问题进行形式化定义,然后提出分析模型来求解关键参数,如相邻AP的负载差、平均数据功率损失代价以及AP全覆盖的条件.最后,采用遗传算法(GA)求解多约束负载均衡问题,并测试在不同用户密度下,所提方案和同类方案的AP负载、吞吐量以及平均分组传输成功率.实验表明,所提方案实现了优化.

In WLAN, load imbalance incurs two problems： lower network throughput and longer transmission delay. To realize load balancing, researchers introduced the concept of cell breathing in Code Division Multiple Access （CDMA） networks into wireless local area networks （WLANs） due to requiring no special modification of clients. Cell breathing technique adjusts APs＇ coverage area through adjusting their beacon power to control the load of APs. Most of existing load bal- ancing methods based on ceil breathing did not care two key problems： the tradeoff between load balancing on APs and data power loss of users, as well as the service availability of APs （inclu- ding AP service cheating and AP service loophole）. In this paper, these two problems are viewed as constraints when using cell breathing method to realize load balancing. Firstly the problem of multi-constraint load balancing is formulated, and then some analytical models are proposed to solve key parameters such as the load gap of two neighboring APs, the average data power loss cost and the condition of AP full coverage. At last, the genetic algorithm is employed to solve the problem of multi-constraint load balancing, and test the proposed scheme and the similar scheme in terms of AP load, throughput and the average successful transmission probability under differ- ent user density. The simulation shows the proposed scheme realizes optimization.