基于社会网络属性的容忍延迟网络关键技术研究

发布时间：2018-04-25 23:04

本文选题：容忍延迟网络 + 社会网络　；参考：《北京邮电大学》2014年博士论文

【摘要】：随着无线通信需求的增加,无线网络逐渐延伸到拓扑更多变、信道更恶劣的应用环境中,如星际通信、野生动物监控、移动社交和灾难救助等。这些应用场合的共同特点是：节点间的稳定通信不能够得到保证,源宿之间的端到端路径不复存在。容忍延迟网络,即容迟网络,作为应对此类环境的最佳解决方案,牺牲时延与缓存空间换取数据的成功投递,利用节点有规律的移动、借助“存储—携带—转发”的通信方式变天堑为通途。近几年来,相关交叉学科的发展将容忍延迟网络关键技术的研究推向了新的阶段。网络的泛在化与信息采集技术的普及让研究者意识到,通信节点往往附着在具有自我意识的个体之上,其移动呈现出复杂的统计规律；复杂网络与社交网络的发展让研究者从全新的角度审视个体间的社会关系对网络性能的影响；数据挖掘与语义网的兴起揭示了比特数据的内容属性,让研究者从更高的维度来看待信息的传递过程。随着以上新特性的引入,具有社会网络属性的容迟网络成为研究的热点与难点。在具有社会网络属性的容迟网络中,个体对地点的偏好、个体之间的社会关系、群组与社区、个体对数据内容的偏好等因素都可以看作是社会属性的体现,本文立足于研究社会属性的引入对传统容迟网络的影响,将社会属性归纳为：移动模型的社会性、网络拓扑的异构性、传输数据的内容性和节点转发的自私性,从四个不同的侧面对容迟网络的关键技术展开研究。在移动模型的社会性方面,本文从进化的角度深入研究节点移动产生复杂规律的原因,将节点移动的动力和阻力分别建模为社会收益与移动开销,将收益开销比作为衡量个体适应环境能力的准则,使复杂的社会行为在进化过程中涌现出来。本文提出了基于遗传算法的移动模型GAMM,通过分析生成场景的空时规律,证明了个体朝着收益开销比最大化方向进化的过程是熵减过程,其统计规律符合真实场景的空时分布规律。此外,本文还通过社区模型、探索者模型和交通工具模型验证了移动模型的可扩展性,为后文的研究打下了基础。在网络拓扑的异构性方面,本文首先从理论的角度研究高级节点的引入对同构容迟网络性能的影响,将消息传播的过程模拟为传染病扩散的过程,利用复杂网络中的传染病模型得到不同路由机制下性能指标的闭式表达,并将模型扩展到异构场景,分析不同路由机制对信息站系统的适应能力。本文提出了具有高兼容性的容迟网络信息站架构,将传统的路由策略扩展为支持信息站的路由机制,并配合以相应的缓存控制机制,旨在发现并充分利用信息站的高容量缓存和信息站之间的高速链路。此外借助信息站高速的信息扩散能力,设计了抑制冗余消息的副本控制策略,并从理论分析和仿真测试两个方面证明了信息站系统与改进机制的有效性。在传输数据的内容性方面,本文重点研究如何利用数据的内容属性从宏观的角度控制消息的转发过程,提出了以内容为中心的容迟网络架构CCDTN。该架构融合了下一代网络技术中的以信息为中心网络架构和容迟网络自身的特点,用描述矢量将地理位置、用户喜好和数据内容三者的社会属性紧密地联系在一起,应用层支持社交网络中文件分享、发布订阅等新型业务,并预留了丰富的社会网络属性感知接口以提高架构的可扩展性。基于CCDTN架构,本文引入节点的亲密度、消息的相似度、节点对消息的关注度等概念,综合利用容迟网络的社会属性提高路由机制的性能。节点转发的自私性是将社会属性引入容迟网络而产生的副作用,本文发现由于个体对地点和数据内容的偏好,加上缓存空间的限制,容迟网络中的自私行为呈现出内容选择性。本文借助理论模型分析了内容选择性自私行为产生的原因,并从缓存控制的角度给出了解决方案,提出了互补消息补位的优先级排队策略CCDTN-SR,提高了内容转发的公平性。本文通过仿真的手段分析了CCDTN网络架构的性能,以及不同路由机制对该架构和新型业务的适应能力,并验证了自私行为抑制机制的有效性。
[Abstract]:As the demand for wireless communication is increasing, the wireless network has gradually extended to more topology changes, such as interstellar communication, wildlife monitoring, mobile social networking and disaster relief in worse channel applications. The common features of these applications are that the stable communication between nodes can not be guaranteed and the end to end path between the source and the host is gone. There is a delay network, that is, the delay network, as the best solution to this environment, sacrificing the time delay and the cache space for the successful delivery of the data, using the regular movement of the node, and using the "storage - carry - forward" communication mode to change the road.
In recent years, the development of interrelated disciplines has pushed the research on the key technologies of tolerance delayed networks to a new stage. The ubiquitous and information acquisition technology of the network makes researchers realize that communication nodes are often attached to individuals with self-consciousness, and their movement presents complex statistical laws; complex networks and social networks The development of the network allows researchers to examine the impact of social relationships among individuals from a new perspective. The rise of data mining and semantic networks reveals the content attributes of bit data and lets researchers look at the transmission process of information from a higher dimension. With the introduction of the new characteristics above, the capacity of the social network is delayed. The network has become a hot and difficult point in the research.
In the late network with social network attributes, the preference of the individual to the place, the social relationship between individuals, the groups and communities, and the preference of the individual to the data content can be regarded as the embodiment of the social attributes. This article is based on the study of the influence of the introduction of social attributes on the traditional late network, and the social attributes are summarized as: The sociality of the dynamic model, the heterogeneity of the network topology, the content of the data transmission and the selfishness of the node forwarding are studied from four different sides to the key technologies of the delayed network.
In the sociality of the mobile model, this paper studies the reasons for the complex rules of node movement from the evolutionary point of view, modeling the dynamic and resistance of the mobile node as the social income and the moving overhead, and taking the income ratio as the criterion to measure the ability of the individual to adapt to the environment, so that the complex social behavior emerges in the evolutionary process. In this paper, a mobile model GAMM based on genetic algorithm is proposed. By analyzing the space-time law of generating the scene, it is proved that the process that the individual evolves in the direction of maximizing the revenue cost is entropy reduction, and the statistical rules conform to the space time distribution law of the real scene. In addition, this article also uses the community model, the Explorer model and the traffic worker. The model verifies the scalability of the mobile model, which lays the foundation for further research.
In the aspect of the heterogeneity of network topology, this paper first studies the influence of the introduction of advanced nodes on the performance of isomorphic delayed network from the theoretical point of view. The process of message propagation is simulated as the process of infectious disease diffusion, and a closed expression of performance indexes under different routing mechanisms is obtained by using the infectious disease model in the complex network, and the model is extended. In this paper, the adaptability of different routing mechanisms to the information station system is analyzed. In this paper, a high compatibility delay tolerant network information station architecture is proposed. The traditional routing strategy is extended to support the routing mechanism of the information station, and the corresponding caching mechanism is used to find and make full use of the high capacity cache of the information station. The high speed link between information stations. In addition, with the help of information station's high speed information diffusion ability, a replica control strategy for suppressing redundant messages is designed, and the effectiveness of the information station system and the improved mechanism is proved from two aspects of theoretical analysis and simulation test.
In terms of the content of data transmission, this paper focuses on how to use the content attributes of the data to control the forwarding process from the macro point of view, and proposes a content centric late network architecture CCDTN. which combines the features of the information centric network architecture and the capacity of the late network in the next generation network technology. Vector is closely linked to the social attributes of geographical location, user preferences and data content three. Application layer supports file sharing, publishing and subscriptions and other new services in social networks, and has reserved a rich social network attribute perception interface to improve the scalability of the architecture. Based on the CCDTN architecture, this paper introduces the degree of intimacy of nodes. The concept of similarity of messages, degree of concern of nodes to messages, and so on, comprehensively utilize the social attributes of Delay Tolerant Networks to improve the performance of routing mechanisms.
The selfishness of node forwarding is the side effect of introducing social attributes into the delayed network. This paper finds that the selfish behavior in the late network presents a content selectivity due to the preference of the individual to the location and data content and the restriction of the cache space, and the reasons for the selective selfish behavior of the content are analyzed by the theoretical model. The solution is given from the perspective of caching control, and the priority queuing strategy of complementary message complementarity is proposed, CCDTN-SR, which improves the fairness of content forwarding. This paper analyzes the performance of the CCDTN network architecture by means of simulation, and the adaptability of different routing mechanisms to the architecture and new type of services, and verifies the selfish behavior. The effectiveness of the inhibition mechanism.

【学位授予单位】：北京邮电大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN929.5

【相似文献】