DTN中基于区域划分和节点活跃度的路由研究

发布时间：2018-08-05 10:21

【摘要】：传统的因特网(Internet)是基于TCP/IP协议的,TCP/IP协议之所以能够迅速发展,是因为它适应了世界范围内数据通信的需要。TCP/IP协议的平稳运行需要满足持续的、双向的、端到端的路径,并且它所依赖的物理链路需要往返延时短、对称的数据速率、误码率低等特性。但随着网络的不断发展,一些网络的物理链路却不具备这些特性中的一个或多个。在这种情况下,2002年Kevin Fall 等人在 ICIR 会议上提出了容滞网络(Delay Tolerant Networks,DTN)的概念,这是一种面向消息的覆盖层体系网络结构。自此,研究者们对DTN的各个方面展开了研究,研究的领域包括DTN的路由、拥塞控制、安全等。DTN最初是由星际通信网络发展而来,它具有低传输率、间歇型连接、节点频繁移动、延迟容忍、错误容忍、有限的存储以及通信环境恶劣等特点,能够适应网络中由于节点稀疏分布、无规律移动以及资源有限等条件造成的节点与节点之间的链路不稳定性。在这样的网络环境下,决定了在DTN中应采用其独特的路由策略。研究者们在路由方面做了大量的工作,提出了一系列行之有效的路由算法,主要有单拷贝和多拷贝两大类。因为DTN独特的网络环境,如何快速可靠地将数据传输到目的地,需要研究者们设计出合理的数据传输方案。本文首先介绍了 DTN的研究背景和特点,然后对DTN体系结构、节点的移动模型以及比较经典的路由协议分别进行了相关介绍,最后深入研究散发等待路由,在散发等待路由的基础上,提出本文的路由算法。本文的主要研究内容如下:(1)结合网络区域的概念,提出划分网络区域的方案,并在区域上设置中继节点,在区域间设置摆渡节点。考虑到每个区域的节点疏密程度、区域间的距离远近以及节点的移动性等特点,提出了节点活跃度的概念。再根据节点活跃度动态地分配报文的副本数。(2)针对网络区域划分的特性以及结合节点活跃度,提出了基于区域划分和节点活跃度的散发等待路由算法。通过区域划分,可以将一个大的网络区域划分为多个小的网络区域,并在这些区域上设置中继节点和摆渡节点。通过节点活跃度,节点可以动态地分配报文副本数,还可以根据活跃度选择最适合的下一跳节点。(3)本文使用THE ONE仿真器对提出的算法、二分散发等待路由、基于报文剩余价值的二分散发等待路由和基于节点能力的自适应散发等待路由进行了仿真,并对结果进行了分析比较。
[Abstract]:The traditional Internet (Internet) is the rapid development of the TCP / IP protocol based on the TCP/IP protocol, because it meets the needs of worldwide data communication. The smooth operation of the TCP / IP protocol needs to meet the continuous, two-way, end-to-end path. And the physical link it depends on needs short round trip delay, symmetric data rate, low bit error rate and so on. However, with the development of network, some network physical links do not have one or more of these characteristics. In this case, Kevin Fall et al put forward the concept of (Delay Tolerant Networks (DTN) at the ICIR conference in 2002, which is a message-oriented overlay architecture. Since then, researchers have studied various aspects of DTN, including routing, congestion control, security, and so on. DTNs were originally developed by interstellar communication networks with low transmission rates, intermittent connections, and frequent node movements. Delay tolerance, error tolerance, limited storage and poor communication environment can adapt to the link instability between nodes caused by the sparse distribution of nodes, irregular movement and limited resources. In such a network environment, it is decided to adopt its unique routing strategy in DTN. Researchers have done a lot of work in routing, and put forward a series of effective routing algorithms, including single-copy and multi-copy. Because of the unique network environment of DTN, how to transfer data to the destination quickly and reliably requires researchers to design a reasonable data transmission scheme. This paper first introduces the research background and characteristics of DTN, then introduces the architecture of DTN, the mobility model of nodes and the classical routing protocol. Based on the distributed waiting routing, this paper proposes a routing algorithm. The main contents of this paper are as follows: (1) combined with the concept of network region, a scheme of dividing network area is proposed, and relay nodes are set up in the area, and ferry nodes are set in the region. Considering the characteristics of node density in each region, distance between regions and mobility of nodes, the concept of node activity is proposed. Then dynamically allocate the number of copies of the packet according to the node activity. (2) aiming at the characteristics of the network area division and combining the node activity, a routing algorithm based on the region partition and node activity is proposed. A large network area can be divided into a number of small network areas, and relay nodes and ferry nodes can be set in these areas. Through the activity of nodes, nodes can dynamically allocate the number of copies of packets, and can also select the most suitable next hop node according to the activity degree. (3) in this paper, we use THE ONE emulator to set the proposed algorithm, and divide the distributed waiting route. The binary distribution waiting routing based on the residual value of packets and the adaptive routing based on node capability are simulated, and the results are analyzed and compared.
【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】：TP393.0

【相似文献】