基于内容预测的信息中心网络缓存技术研究

发布时间：2018-05-23 21:43

本文选题：内容中心网络 + 缓存　；参考：《北京邮电大学》2017年硕士论文

【摘要】：随着互联网不断发展,实际网络中用户的需求更加多样化、复杂化,每天产生的网络流量也不断增加,使得网络的负载日渐加大,这对现行的网络架构提出了新的挑战。传统网络采用了以主机地址为核心的通信模式,然后在未来互联网生态中,用户更关注的是信息本身而非信息的归宿地。在这样的背景下,学术界提出了信息中心网络(Information-Centric Networking, ICN)模型,这种新架构的核心是内容,所有的新型路由缓存都基于此。而网内缓存是这种新型网络模式的显著特征之一,通过网内缓存技术,ICN可以实现更快的内容访问速率,提高资源利用率,减少用户等待时间。ICN的实现形式有多种,内容中心网络(Content-Centric Networking,CCN)就是其中一种实现架构。传统CCN中采用了处处缓存的策略,这种策略不能很好利用有限的缓存空间,并且无法区分不同内容之间的访问频率差异,造成了严重的资源浪费和内容冗余,不利于提高缓存空间内容的多样性,同时,也对网络的整体负载带来了诸多压力。内容中心网络中以内容本身为核心,缓存系统的设计也应该着重考虑内容本身的特性。针对这个问题,本文提出了一套基于内容流行度预测的BEACON缓存策略系统。首先,定义了新型的内容流行度概念来衡量用户对内容的偏好程度,并设计了内容流行度收集和统计模块,实现了 CCN架构下的内容流行度监测功能;其次,基于历史数据,使用灰色预测理论进行流行度的预测,并做了详细的数学建模分析。然后,添加了拓扑重要性的评价模块,采用介数来评价节点在网络中的重要程度。在此基础上,完成BEACON缓存策略的设计。最后,本文在仿真环境中,对比了该策略与其它几种常用策略的性能差异,实验表明,BEACON策略能够有效地提高网络的缓存命中率,提高缓存资源利用率,降低网络负载,进而减小用户获取资源的时间消耗,提升了网络的整体性能。
[Abstract]:With the continuous development of the Internet, the needs of users in the actual network become more diversified and complex, and the network traffic generated every day is increasing, which makes the load of the network increasing day by day, which poses a new challenge to the current network architecture. The traditional network uses the host address as the core communication mode, then in the future Internet ecology, the user pays more attention to the information itself rather than the destination of the information. In this context, the information center network Information-Centric networking (ICN) model is proposed. The core of this new architecture is content, based on which all new routing caching is based. Intra-network caching is one of the prominent features of this new network mode. ICN can achieve faster content access rate, improve resource utilization, and reduce user waiting time. ICN can be implemented in many ways. Content-Centric networking is one of the implementation architectures. In traditional CCN, the policy of everywhere cache is adopted, which can not make good use of limited cache space, and can not distinguish the difference of access frequency between different contents, which results in serious waste of resources and redundancy of content. It is not conducive to improving the diversity of cache content, but also brings a lot of pressure to the overall load of the network. The content itself is the core in the content-centric network, and the design of the cache system should also focus on the characteristics of the content itself. To solve this problem, this paper proposes a BEACON cache policy system based on content popularity prediction. Firstly, a new concept of content popularity is defined to measure the user's preference for content, and a content popularity collection and statistics module is designed to realize the function of content popularity monitoring based on CCN architecture. The grey prediction theory is used to predict the popularity, and a detailed mathematical modeling analysis is made. Then, an evaluation module of topological importance is added to evaluate the importance of nodes in the network. On this basis, the design of BEACON cache policy is completed. Finally, in the simulation environment, this paper compares the performance difference between this strategy and several other commonly used strategies. Experiments show that the BEACON strategy can effectively improve the cache hit rate of the network, improve the utilization of cache resources, and reduce the network load. Furthermore, the time consumption of users to acquire resources is reduced, and the overall performance of the network is improved.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP393.0

【相似文献】