Research On Overlay Multicast Delivery In Wireless Mesh Networks

The target of next generation wireless communication networks is to provide users with Intemet access and QoS-guaranteed multimedia service broadband connection in a flexible manner.Wireless Mesh Networks(WMNs),which efficiently extends the coverage of wireless access networks,is introduced for this purpose.However,to develop various applications,WMNs faces challenges on how to deliver efficiently real-time mlticast-demanding applications,such as demand of video,mobile e-learning etc.This dissertation is focus on how to improve the quality of flows delivery and the throughput of WMNs by combining overlay multicast with wireless mesh networks.Although overlay multicast is designed to carry out multicast in the application layer, many factors in other layers also affect the quality of delivery flows,such as physical topology,access control,routing etc.The conventional layered-protocol architecture screens those factors upward and lacks interaction between different protocol layers. WMNs will have a poor performance if the conventional layered protocol design methods and algorithms are employed.Thus,the methods of "optimization principle" and "cross-layer design method that up-layers achieve information from bottom-layers" are investigated in the literatures.However,these methods.usually lack the guidance of global optimization.In this dissertation,the cross-layer design using Network Utility Maximization(NUM) in WMNs and a distributing optimization algorithm based on utility to control delay,power consumption,link stress which take place in the process of delivery flows,are presented.In this dissertation,the background,and key issues of WMNs and overlay multicast are introduced.The necessity of Network Utility Maximization for the cross-layer design in WMNs is analyzed.The major works are as follows:Firstly,an approach for carrying out cross-layer design by joining congestion control and contention control aiming at end-to-end delay of the multicast flows in WMNs is proposed.It utilizes the network utility maximization framework to tackle this problem, by combining optimization flow rate allocation and channel access probabilities of links to decrease congestion and confliction so that the end-to-end delay of the delivery multicast flows in WMNs is decreased.Then,a distributed algorithm based on pricing scheme by using Lagrange dual decomposition technique for contention control and congestion control is presented.Secondly,to decrease mutual interference between links and improve the quality of flows delivery,the idea of net utility maximization is proposed by introducing the power consumption of links along flow's routing path to the NUM.The optimal design of WMNs is formulated as a nonlinear optimization problem.By using the dual decomposition method,the primal nonlinear problem is separated into several sub-problems which correspond to a protocol layer in the traditional network respectively.Based on this,the distributed power control and flows rate adjusting algorithm are designed to improve network performance by minimizing power consumption of the links and optimizing the rate of flows.Thirdly,in the application of overlay multicast,due to duplication of the data flows at end host,the problem of link stress is inevitable.When the link stress increases,this will introduce the issue of congestion and consume more network bandwidth.In this dissertation,an approach to combine optimizing rate allocation of flows and stress of delivering data flow links along flow's routing path,is proposed by a pricing scheme,in which each link adjusts its stress price according to the traffic load and each flow in turn collects the stress prices of all links along its routing path and calculates the overall network price.Basing on this,it adjusts the streaming rate and forward path.The works in this dissertation have been supported by the Natural Science Foundation of China program "Research on interactive media streaming distrioution technologies in wireless mesh network"(No.60773193) and the Natural Science Foundation of China program "Research on Wireless Sensor Networks based on Network Utility Maximization"(No.60772088).