DTN中基于拥塞程度的概率路由研究

发布时间：2018-05-20 13:34

本文选题：容滞网络 + 节点拥塞程度　；参考：《安徽大学》2017年硕士论文

【摘要】：伴随着移动无线通信技术的快速发展,许多新兴网络不断涌现出来,如星际网络、陆地移动Ad hoc网络、军事网络、无线传感器网络,它们不同于传统的基于TCP/IP协议的网络,无端到端的路径存在,且间歇性连接、网络拓扑多变、长的可变延迟、非对称数据速率等。为了实现这类特殊的网络通信,在2003年SIGCOMM国际会议上,一种新的网络体系结构被KevinFall等DTN研究者提出,称为容滞网络(DTN:Delay Tolerant Network),这种网络体系结构是在传统的网络层与应用层之间加了一层聚束层(Bundle层),通过此层中的"存储-携带-转发"方式来完成报文的转发以实现最终递交。DTN中的路由算法主要有两类,一类为单拷贝路由算法,另外一类为多拷贝路由算法。多拷贝路由算法主要是通过增加报文的副本数目来实现的,即允许同一报文被多次复制转发给中继节点,通过多条路径最终递交到信宿节点。在递交率方面多拷贝路由比单拷贝路由要高且路由算法研究的重点也在多拷贝路由算法上,因此在受限的网络环境下,有效的多拷贝路由算法具有重要意义。本文的主要工作如下:1.首先,概率路由算法是一种基于历史信息和可传递概率的路由算法,选择与信宿节点相遇概率高的节点作为中继节点,但未考虑历史相遇时每次的连接时间,如果历史相遇概率高但每次连接时间较短,也有可能导致报文无法顺利递交过去。因此本文提出了新的递交预期值,即考虑了相遇的概率,又增加了每次相遇时的连接时间。其次,当中继节点到信宿节点的相遇概率大于当前节点到信宿节点的相遇概率时,报文递交给中继节点,但此时并未考虑中继节点的拥塞程度,当中继节点拥塞时,即使中继节点到信宿节点的相遇概率大,中继节点接收报文也有可能导致更大的拥塞发生或者直接将接收的报文丢弃。因此本文提出了节点拥塞程度的概念,增加了转发时对中继节点拥塞程度的判断。综合了新的递交预期值和中继节点拥塞程度判断,提出了一种改进的基于节点拥塞程度的概率路由算法,改善了报文的递交率和平均延迟等网络性能。2.本文在基于节点拥塞程度的概率路由算法基础上,增加了相应的拥塞检测机制和拥塞控制机制;当转发报文给中继节点时,优先依次转发丢弃策略因子D值较大的报文,当中继节点拥塞时,优先依次丢弃转发策略因子D值较小的报文,直到为新报文腾出足够的缓存空间。与未加入拥塞控制的基于节点拥塞程度的概率路由算法相比,提高了报文的递交率和降低了网络的开销率。3.本文使用The ONE仿真软件对所提出的一种改进的基于节点拥塞程度的概率路由算法及拥塞控制策略分别进行了仿真验证,并与其它的路由算法进行比较分析。
[Abstract]:With the rapid development of mobile wireless communication technology, many new networks are emerging, such as interstellar networks, terrestrial mobile Ad hoc networks, military networks, wireless sensor networks, which are different from the traditional networks based on TCP/IP protocol. End-to-end paths exist with intermittent connections, network topologies vary, long variable delays, asymmetric data rates, and so on. In order to realize this kind of special network communication, at the 2003 SIGCOMM International Conference, a new network architecture was proposed by DTN researchers such as KevinFall. Called DTN: delay Tolerant Network, this network architecture is to add a bundle layer between the traditional network layer and the application layer. In this layer, the "storage-carry-forward" mode is used to complete the forwarding of messages. There are two kinds of routing algorithms in the final delivery. DTN. One is a single copy routing algorithm, the other is a multi-copy routing algorithm. Multi-copy routing algorithm is mainly implemented by increasing the number of copies of the message, that is to say, the same message is allowed to be copied and forwarded to the relay node many times, and finally submitted to the host node through multiple paths. In the aspect of delivery rate, multi-copy routing is higher than single-copy routing, and the research focus of routing algorithm is also on multi-copy routing algorithm. Therefore, in restricted network environment, effective multi-copy routing algorithm is of great significance. The main work of this paper is as follows: 1. Firstly, the probabilistic routing algorithm is a routing algorithm based on historical information and transferable probability. The node with high probability of meeting with the host node is chosen as the relay node, but the connection time of each time when the historical encounter is not considered. If the probability of historical encounter is high but the time of each connection is short, it may also lead to the failure to submit the message smoothly. In this paper, a new expected value is proposed, that is, the probability of meeting is considered, and the connection time of each encounter is increased. Secondly, when the encounter probability between the relay node and the host node is greater than that between the current node and the host node, the message is submitted to the relay node, but the degree of congestion of the relay node is not considered at this time, when the relay node is congested, Even if the encounter probability between relay node and host node is high, the receiving message of relay node may lead to more congestion or discard the received message directly. Therefore, the concept of node congestion degree is proposed in this paper, which increases the judgment of relay node congestion degree when forwarding. A new probabilistic routing algorithm based on node congestion degree is proposed to improve the network performance of packet delivery rate and average delay. Based on the probabilistic routing algorithm based on node congestion degree, this paper adds the corresponding congestion detection mechanism and congestion control mechanism. When the relay node is congested, the packets with a smaller forwarding policy factor D are discarded first and then until sufficient buffer space is made available for the new message. Compared with the probabilistic routing algorithm based on the degree of congestion without congestion control, the packet delivery rate and the overhead rate of the network are improved. In this paper, an improved probabilistic routing algorithm based on node congestion degree and congestion control strategy are simulated and verified by using The ONE simulation software, and compared with other routing algorithms.
【学位授予单位】：安徽大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN915.0

【相似文献】