IP编码网络及其优化算法

IP coding network and optimization algorithm

为了克服当前IP网络的带宽瓶颈问题,提出了一种基于网络编码的IP网络通信机制—IP编码网络。该文讨论了网络节点上包的代数运算与路由协议的结合方式,提出了IP编码网络中单、组播通信的实现方法。在此基础上,对应于现有最短路径、最小代价组播树的网络优化问题,提出了一种基于遗传算法的IP编码网络优化算法。该算法为单播和组播会话构造一个在最大流约束下通信代价优化的编码子网。算法的模拟实验结果表明对任意的单播和多源组播通信,都可以在有限的计算代价和计算时间内生成通信子网,有效减少了IP编码网络的编码开销和通信代价。该文的研究工作为网络编码的应用作了有益的实践探索。

A communication mechanism based on network coding, an IP coding network., was developed for current IP networks to relieve bandwidth bottlenecks. This paper presents algebraic models of the packets and routing protocols and designs of communication models for both unicast and multicast communications in IP coding networks. A genetic optimization algorithm was used to reduce the computations and bandwidth, existing optimization methods for the shortest path and minimum-cost multicast tree. The algorithm construct a packet coding sub-network for any unicast or multicast session for the max flow constraint. Simulations show a noticeable decrease in coding overhead and bandwidth consumption in the sub-network constructed at acceptable computation cost and time for any unicast or multi source muhicast communication.