A Study Of Network Capacity Based On Coding Strategies

In this paper we study how to improve the network capacity of wireline and wireless networks with coding techniques.For coding techniques,we consider network coding,which is initiated to improve the capacity of wireline multicast networks and decode-and-forward cooperation,which is studied to improve the capacity of relay channels.However,there is limitation in the existing literature that neither of these two coding techniques improves network capacity effectively in real practice.For wireline networks,this paper presents Coded IP,which is the first protocol to implement IP-layer network coding.In the existing work,the application of network coding is mainly concentrated on the P2P network coding which is application-layer oriented.In P2P network coding,the network routers do not participate in coding.Thus it can never achieve the network capacity.In this paper we give the specification and detailed implementation of Coded IP, and test its performance in a real multicast testbed.Experimental results show that IP-layer can take a great advantage in terms of throughput improvement and user downloading time,compared with P2P network coding and IP multicast. In addition,IP-layer network coding lowers the bandwidth consumption at the routers and manages to provide a more balanced traffic load.For wireless networks,this paper proposes to improve the end-to-end network throughput with physical layer cooperation.In particular,we consider a multi-channel network,and specify the physical layer cooperative model with decode-and-forward strategy and rateless codes.Based on this model,we re-formulate the routing discovery and channel assignment problem in multi-channel networks,and give centralized and distributive algorithms to solve the problem.Simulation confirms that the end-to-end throughput could be effectively improved with the physical layer cooperation.