深空DTN多属性决策拥塞控制的研究

Research on congestion control mechanism of multi-attribute decision making in deep-space DTN

针对深空时延容忍网络(DTN)因节点存储空间有限造成的拥塞问题,提出了一种改进的多属性决策选择下一跳节点的策略。其基本思想是根据节点的存储空间占用率将节点状态划分成普通、半拥塞和拥塞三个状态,结合节点接收数据包的历史记录,在此基础上采用多属性决策理论选择最优的下一跳节点,从而达到预防拥塞的目的。仿真结果表明,与原策略相比,提出的策略在单个节点队列长度、平均队列长度、数据包接收率和数据包平均分发时间上有明显的改善,有效地缓解了网络拥塞。

For congestion in deep-space DTN was mainly caused by the limited node memory space,this paper proposed an improved strategy for multi-attribute decision making to select the next hop node.The basic idea of the improved strategy was to divide the node status into normal,semi-congestion and congestion states according to the node storage occupancy,combined with the packet received history and then used multi-attribute decision theory to select the best next hop node,so as to achieve the purpose of prevention of congestion.Simulation results show that the improved strategy is more effective than the original one in the single node＇s queue length,average queue length,packet acceptance rate and the average packet distributed time.It alleviates the congestion effectively.