【作者】 陶峰；

【导师】 吴国新；

【作者基本信息】 东南大学 ， 计算机应用技术， 2015， 硕士

【摘要】 机会网络是一种异于传统网络的网络体系结构,它以“存储-携带-转发”的工作模式,利用节点移动带来的通信机会实现数据传输。然而,由于网络拓扑变化快,使得机会网络存在消息传输延时较大、投递成功率较低等问题。这些问题限制了用户体验,制约了机会网络应用的进一步发展。随着应用场景的改变以及网络设施的改善,机会网络场景中加入了基站、热点和服务器等实体,以提高整个内容分发效率。近年来,随着手持智能设备的普及,机会网络的研究热点转向了结合社会网络理论研究手持设备组网、大数据转发与数据分流、与集中式网络结合的混合应用系统等。本文以大数据内容分发应用为场景,提出中心服务器内容分发模型,利用中心服务器的全局信息设计内容分发策略以提高分发效率。论文主要工作包括：设计中心服务器内容分发模型,该模型是基于机会网络的设备对等直连应用模型,利用中心服务器、蜂窝网络和机会网络的三层结构,中心服务器负责数据分析和分发控制,通过蜂窝网络进行小数据量的控制信息传输,机会网络实现大数据量的内容分发。用于中心服务器分析的数据来源于本校网络真实使用数据,从节点的移动反映人类活动规律,而人类活动存在固有的周期性与社会性的假设出发,通过对历史数据的统计分析,为每个物理位置选择连通性强和活跃性高的节点作为该位置上的辅助转发节点。实验证明,有效利用辅助转发节点能够在保证快速扩散的同时降低散发过程中的盲目性,提高内容分发效率。针对辅助转发节点数量和存储容量的限制,利用目标节点与辅助节点在位置上的重叠性,加强辅助节点选择时的限制条件,同时加入转发终止机制,当目标节点收到内容后,其它拥有该内容的节点在与服务器交互时及时获知并移除该内容,以达到缓解存储的目的。最后,通过ONE仿真平台进行仿真实验,实验结果表明优化后的内容分发策略在节点存储压力较大时有较好表现。更多还原

【Abstract】 The Opportunistic Networks is a new model of network architecture, which is different from the tradition. It completes data forwarding with "Store-Carry-Forward" work mode, using the communication opportunities brought by node mobility. However, since the network topology changes quickly, the Opportunistic Networks has some issues such as large delay of message transmission and low delivery rate. These issues affect the user experience, limiting the opportunities for further development of network applications. With the improvement of application scenarios and network infrastructure changes, in order to improve the efficiency of content delivery, Opportunistic Networks scenarios added base stations, hotspots, and server entity.In recent years, with the popularity of smart handheld devices, the research of Opportunistic Networks turned to the combination of social network theory research of handheld device networking, data transfer and large data streaming, mixed applications combined with centralized network and so on. In this paper, the scene is based on school smart mobile device networking application, presenting a central server content forwarding model, using the global information of central server to design content forwarding strategies to achieve improved the efficiency. The thesis mainly works are as follows:Design a content forwading model with central server, which is based on direct connection between devices, using a three-layer structure with central server, cellular network and Opportunistic Networks. Central server works for data analysis and forwarding control, and the important control information which is small amounts of data transferred over the cellular network, Opportunistic Networks responsible for implementing the content forwarding which is large amounts of data.The data used to analysis is from the real network data of SEU. According to the assumption that nodes’mobility is the reflection of human mobility and human activities are periodictable and sociable, selecting strong connectivity and highly active node for each physical location as an adjunct forwarding node that position through statistical analysis of historical data. Experimental results show that the effective use of auxiliary forwarding nodes can ensure the rapid dissemination of the process while reducing the spread of blindness, thus improving the efficiency of content forwading.For the auxiliary forwarding nodes’s numbers and storage capacity limits, strengthen the restrictions of auxiliary node selection by using the target node and a auxiliary node position overlap, while adding forwarded termination mechanism, when the target node receives content, other nodes which have this content informed promptly and remove the contents while interacting with the server to achieve remission storage purposes. Do Simulation experiment by the simulation platform:The ONE. The results of experiments show that the optimized content forwarding strategy has a better performance in large -node storage pressure.更多还原