【作者】 常秋菊；

【导师】 杨金民； 吴昊；

【作者基本信息】 湖南大学 ， 软件工程， 2010， 硕士

【摘要】 移动自组网不依赖固定基础设施,具有自组织、可快速部署、环境适应力强、抗毁性强等特点,在军事战场通信、抢险救灾、紧急搜救等特殊环境中有着广泛的应用。移动自组网中的数据可用性非常低,数据复制技术作为数据管理的关键技术和提高数据可用性及系统性能的重要手段,在移动自组网中有着非常重要的应用。由于节点的移动性、能量有限性的问题,使数据复制技术面临节点邻居的不稳定和能量消耗不均的问题。因此,面向稳定预测和能量预测方面的复制算法的研究有着重要意义。本文的主要工作包括：第一、针对邻居节点的边缘移动和能量消耗的不公平性,提出了基于稳定邻居的数据复制(SNDR)算法。该算法对节点与邻居间的连接进行稳定性分析,计算一定时间间隔后的稳定率,通过稳定率获取节点的稳定邻居集。节点与稳定邻居共享数据资源,收集稳定邻居对数据的加权访问频度,并对访问频度总和降序排列。按照排列放置副本时,对访问频度相近的多个节点,综合考虑这些节点的剩余能量、稳定连接度和剩余存储空间,选择合适的节点放置副本。这样在减少节点边缘移动的影响同时,对节点的能量消耗进行平衡。在网络模拟器NS2中对算法进行验证,模拟实验结果表明,在移动自组网中SNDR算法能有效地提高数据的访问成功率,平衡节点间的能量消耗。第二、针对网络中节点快速耗尽而导致部分数据难以访问的情况,提出了基于能量预测的副本调整(EPRA)算法。该算法在节点的剩余能量较少时通过剩余能量值和能量消耗率对节点的剩余生存时间进行预测,若剩余生存时间较短,在节点的邻居范围内通过数据的副本数判断节点中数据的重要性。对重要数据的复制分为邻居中有剩余存储空间、有其它数据的重复副本和没有剩余存储空间及重复副本三种情况,并考虑邻居的访问频度、剩余能量和连接度以选择合适的节点进行放置。模拟实验结果表明,在移动自组网中,部分节点快速耗尽的情况下,EPRA算法能有效提高数据的访问成功率。更多还原

【Abstract】 Mobile ad hoc network does not rely on fixed infrastructure. Due to the features such as self-organizing, rapidly deploying, environmental adaptability and survivability, it has a wide application in military battlefield communications, disaster relief, emergency rescue and other special circumstances. The data availability in mobile ad hoc is very low. As a key technique of data management and an important method to improve data availability and system performance, data replication has an important application in mobile ad hoc network.However, because of the movement characteristics and limited energy of nodes, the data replication technology in mobile ad hoc network faces new problems such as unstable neighbor and unbalance energy consumption. Therefore, stability prediction and energy prediction play an important role in replication algorithm. Our main works can be summarized as follows:Firstly, according to the edge movement of neighbors and the unfairness energy consumption, a replication algorithm named SNDR is proposed based on stable neighbors. SNDR algorithm analysis the stability of the link between node and neighbor and calculates the stability ratio, then gains the stable neighbors of node. In SNDR algorithm, node shares data with stable neighbors; it collects the weighted access frequency of stable neighbors and arranges the sum of access frequency in descending order. In order to select a right node to allocate the replica, the nodes that have the similar access frequency in addition consider node’s residual energy, steady connection degree and free storage space in the placing process. In this way, SDNR can not only reduce the influence of the edge movement of neighbors but also balance the energy consumption between nodes. SNDR algorithm is implemented by network simulator NS2, the simulation experiment results show that in mobile ad hoc network, SNDR algorithm can effectively improve the rate of successful data access and balance the power consumption among nodes.Secondly, considering the situation that some data items become difficult to access after some nodes quickly exhaust the energy, a replica adjusting algorithm named EPRA is proposed based on energy prediction. When the residual energy of node is few, EPRA algorithm predicts the residual survival time of node by residual energy and the rate of energy consumption. If the residual survival time of node is short, EPRA algorithm judges the important data by the number of replicas within neighbors. In the copying process of important data, EPRA algorithm considers three situations that there are free storage spaces, there are other data’s duplication, there are no free storage space and other data’s duplication in neighbors, then select the right node to allocate the important data with the concern about access frequency, node’s residual energy and connection degree. Simulation experiment results show that when partial nodes quickly exhaust the energy, EPRA algorithm can effectively improve the rate of successful data access.更多还原