容延网络中一种改进的概率路由协议

An Improved Probabilistic Routing Protocol in Delay Tolerant Network

容延网络大多基于历史路由协议和粗粒度信息做出消息转发决策,然而这些粗粒度信息无法较好地描述网络节点间的接触模式,降低了转发决策的准确性。为此,提出基于细粒度接触表征的概率路由协议FGPRo PHET,利用时隙滑动窗口机制维护细粒度实时统计数据,通过调整滑动窗口尺寸控制历史数据粒度,进而描述网络节点间的接触模式。根据细粒度接触统计数据及相遇节点间所有接触事件的持续时间,计算节点接触概率。实验结果表明,与现有概率路由协议相比,FG-PRo PHET具有较高的报文投递率和网络缓存资源利用率,以及较低的通信开销。

Most existing history-based routing protocols for Delay Tolerant Network（DTN） make message forwarding decisions according to coarse-grained encounter information. However, these coarse-grained information can not give precise expression of the contact patterns between nodes in the network, thereby leading to inaccurate forwarding decisions. This paper presents a Fine-grained Probabilistic Routing Protocol Using History of Encounters and Transitivity （FG-PRoPHET） ,it uses slotted sliding window mechanism for maintaining fine-grained real-time statistics, and controls the historical data granularity by adjusting the size of sliding window, further describes the contact patterns between network nodes. Based on the statistics of fine-grained contact and the overall duration of all contact events occurring during sliding window of the encountering peer to peer nodes, it can compute the probability of node contact. Experimental results show that, compared with existing probabilistic routing protocols, FG-PRoPHET can significantly enhance message delivery rate and resource utilization of network caching with lower communication overhead.