延迟容忍网络中路由算法的研究

发布时间：2017-12-27 03:22

  本文关键词：延迟容忍网络中路由算法的研究　出处：《电子科技大学》2015年硕士论文　论文类型：学位论文

  更多相关文章： 延迟容忍网络 路由算法 条件接触间隔时间 接触持续时间

【摘要】：随着科技的进步,具有无线接口的通信设备越来越普及。在这个过程中出现一种特殊的无线移动网络——延迟容忍网络(Delay Tolerant Network,DTN)。与传统的无线网络不同,DTN不存在持续的网络连接并具有节点频繁移动、资源受限等特点。由于DTN的这些特性,DTN的路由变得非常具有挑战性,因为它必须处理网络中存在的网络划分、长延时和节点动态拓扑等问题。DTN最初被用于星际网络,目前存在多种不同的应用场景,包括无线传感器网络、军事自组织网络、车载网络和社交网络等等。在DTN中,不可能存在从源端到目的端的持续网络连接。因此,在DTN中,消息的传输依赖机会路由,其中节点使用存储-携带-转发的模式来进行消息路由。然而,基于有限的历史接触信息来制定有效的转发策略仍然是具有挑战性的。本文首先研究了DTN中已存在的一种依据历史相遇和传输信息的概率路由算法(Probabilistic Routing Protocol using History of Encounters and Transitivity,Prophet)。然后引入了一种附加测度——条件接触间隔时间,这一测度只使用本地的历史相遇信息来计算两个节点相对于第三个节点的相遇之间的平均接触间隔时间。我们还分析了引入的附加测度并说明了它在路由中的重要性。接着引入接触持续时间的概念,并分析其合理性。最后本文利用接触持续时间和条件接触间隔时间分别对Prophet的概率计算和转发策略进行了改进。实验结果表明,改进后的Prophet有效地降低了延时,提高了消息的投递率,减少了网络开销。然后在改进后的Prophet基础上,分析了散发与等待(Spray and Wait,SW)路由算法在散发阶段和等待阶段的不足,提出了一种改进的基于概率的散发与等待路由算法。在原有的散发与等待路由算法的散发阶段中,节点盲目地将消息副本散发给所有与它相遇的节点。基于这一点,本文对散发与等待路由算法的散发策略进行改进。考虑节点的所有邻居节点的消息递交概率值,应用公式计算出散发给各节点的消息副本数。另外在等待阶段,携带消息副本的节点不再是等待目的节点然后直接交付。当节点在移动的过程中遇到与目的节点相遇概率更大且条件接触间隔时间更小的节点的时候,将消息副本将转发给该节点。仿真结果表明,改进后的路由算法在消息的递交成功率、平均端到端时延、网络开销等方面有良好的性能。
[Abstract]:With the progress of science and technology, the communication equipment with wireless interface is becoming more and more popular. In this process, there is a special wireless mobile network, the Delay Tolerant Network (DTN). Unlike traditional wireless networks, DTN does not have a continuous network connection and has the characteristics of frequent mobile nodes and limited resources. Because of these characteristics of DTN, routing of DTN is very challenging, because it must deal with network partition, long delay and node dynamic topology. DTN was originally applied to interplanetary network. There are many different application scenarios at present, including wireless sensor network, military self organizing network, vehicular network and social network. In DTN, it is not possible to have a continuous network connection from the source to the destination. Therefore, in DTN, the transmission of messages depends on the opportunity routing, where the nodes use the storage - carry - forward mode for message routing. However, it is still challenging to make effective forwarding strategies based on limited historical contact information. In this paper, we first study a Probabilistic Routing Protocol using History of Encounters and Transitivity, which exists in DTN, which is based on historical encounter and transmission information. Then an additional measure -- the conditional contact interval time is introduced. The measure only uses the local historical encounter information to calculate the average contact interval between two nodes relative to the third nodes. We also analyzed the additional measure introduced and explained its importance in routing. Then the concept of contact duration is introduced and its rationality is analyzed. At last, the probability calculation and forwarding strategy of Prophet are improved by using contact duration and conditional contact interval time respectively. The experimental results show that the improved Prophet effectively reduces the delay, improves the delivery rate of the message, and reduces the network overhead. Then, on the basis of the improved Prophet, we analyze the shortcomings of the Spray and Wait (SW) routing algorithm in the sporadic stage and the waiting stage, and propose an improved probabilistic based distributed and wait routing algorithm. In the emanating phase of the original emission and waiting routing algorithm, nodes blindly distribute the copy of the message to all the nodes that meet it. Based on this, this paper improves the emission strategy of the emission and waiting routing algorithms. The number of messages sent to each node is calculated by using the application formula to calculate the probability value of the message submitted by all the neighbor nodes of the node. In addition, in the waiting phase, the nodes that carry a copy of the message are no longer waiting for the destination node and then delivering it directly. When nodes move in the process of encountering nodes that have greater probability of meeting with the destination node and smaller conditional contact intervals, the message copy will be forwarded to the node. The simulation results show that the improved routing algorithm has good performance in the success rate of message delivery, the average end to end delay, the network overhead and so on.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN929.5
，

本文编号：1340048