Multicast Algorithm In Multi-Channel Multi-Radio Wireless Mesh Networks

Multicast is a kind of "one to many" communication model that can optimize the usage of network resources, such as saving bandwidth resources, reducing network load, etc. For wireless mesh networks, the most important issue is to save network resources and improve network throughput. So the problem of multicast routing in wireless mesh networks has got more and more attention. Most past studies for the problem focused on the single-channel single-radio wireless mesh networks. The single channel single radio property severely hinders the network throughput. The most effective way to improve the network throughput is to equip each node with multiple channels and multiple radios, which can make the nodes utilize different channels for sending and receiving at the same time. For wireless networks, another important problem is channel interference. Serious channel interference can increase the end-to-end delay, thus affecting the quality of transmission. Recently, multimedia technologies have been developed rapidly. This kind of application require the network to provide a certain amount of Quality of Service (QoS), such as bandwidth, the end-to-end delay and packet loss rate. The recent research on QoS multicast routing problem focused on the cable network and single-channel single-radio wireless mesh network, and little is studied for multi-channel multi-radio (MCMR).First, this thesis reviews the multicast routing algorithm and Multi-Channel Multi-Radio Wireless Mesh Networks (MCMR WMNs). Sencond, the relevant unconstrained multicast routing algorithm on MCMR WMNs and traditional QoS multicast routing algorithm are introduced in detail. For the problem of some practical applications require high throughput, but some are sensitive to end-to-end delay, we propose a suitable robust multicast algorithm for MCMR WMNs. This algorithm can meet different application requirements by adjusting its parameters of the metric defined for link cost. This metric incorporates the bandwidth resource utilization and channel interference at the same time. We implement the RMRA in NS2and compare its performance with that of Multi-Channel Minimum Number of transmissions Trees (MCMNTs)[8]. The experimental results show our algorithm RMRA can effectively achieve better performance on network throughput and end-to-end delay.For MCMR WMNs, this paper sets delay-constrained multicast routing as the research object and proposes a algorithm called DCRMRA to save network bandwidth and satisfy the delay constraint on MCMR WMNs. This algorithm contains two steps: the first step uses RMRA to calculate the minimum cost multicast routing tree, and remove these paths and destinations violating the delay bound. The second step uses Lagrange algorithm [9] to find the delay constrained path with the minimum cost from the source node to the destination outside the multicast tree found at the first step. In order to show the advantages of DCRMRA, we define a algorithm called delay-constrained least-delay (DCLD) multicast algorithm. The second step of DCLD utilizes the shortest path to replace the paths violating the delay bound. We implement DCRMRA in NS2. The experimental results show that DCRMRA gain better performance compared with DCLD on multicast construction cost, end-to-end delay and network throughput.