基于多因素分析的信息中心网络缓存放置策略研究

发布时间：2018-11-05 08:57

【摘要】：因特网经过六十多年的发展,基于TCP/IP协议的核心网络架构也出现了许多弊端,难以满足用户的需求,因此传统细腰型协议体系亟待改变。2020年及未来,万物互联的时代到来,网络中爆炸式激增的内容数据流量,给传统的网络带来了巨大的冲击和不同挑战。信息中心网络架构创新性地提出用内容名称来替代原来的网络协议地址,从用户对内容需求角度出发,不再关注内容的位置,而是更多的关注网络和用户之间交互的内容。在第二章中,我们主要介绍了信息中心网络的几个关键技术,命名规则,路由机理以及缓存策略等关键技术,并且主要以内容中心网络为例,详细介绍了其数据包的通信流程。此外我们对已有的缓存策略进行了较为全面的了解与研究,对于网络中存在的多种缓存影响因素,主要分为以下三个方面:拓扑相关的缓存影响因子,主要从节点的缓存容量大小和拓扑重要性来考虑的;内容相关的缓存影响因子从节点对内容的感知度来决定的;其他类型的缓存影响因子主要有节点的状态和距离内容命中节点的跳数等,以上介绍的所有因素都会影响网络内缓存系统效率。因此本文提出了基于多因素分析的缓存放置策略,如何综合性评估各缓存因素对于节点缓存效率的影响成为本文研究的关键。层次分析法是一种科学并且准确性较高的决策方法,广泛应用于多目标的复杂问题的决策分析。首先对各节点具有代表性的缓存影响因子进行筛选,再根据层次分析法构建系统评价模型,在命中节点处对选出的缓存影响因子做出权重评价,并结合网络中实时统计的仿真数据,从兴趣包传输路径上选出最佳缓存放置节点,做出最终缓存决策。最后一章我们对缓存放置策略进行建模仿真,完善节点模型并部署策略,仿真结果显示,本文提出的基于多因素分析的缓存放置策略与内容中心网络中默认的LCE策略,以及概率缓存的Probcache策略相比,在缓存命中率,命中平均跳数和服务器负载等指标方面的性能均有提升,得到可行性的分析结果,验证了本文策略的可行性与正确性。
[Abstract]:After more than 60 years of development, the core network architecture based on TCP/IP protocol also has many drawbacks, so it is difficult to meet the needs of users. Therefore, the traditional fine-waisted protocol system needs to be changed urgently in 2020 and in the future. With the coming of the interconnect of all things, the explosive increase of the content data flow in the network brings great impact and different challenges to the traditional network. The information center network architecture innovatively proposes to replace the original network protocol address with the content name. From the point of view of the user's content requirement, the information center network architecture pays more attention to the interactive content between the network and the user instead of paying attention to the location of the content. In the second chapter, we mainly introduce several key technologies of the information center network, such as naming rules, routing mechanism and cache policy, etc. Taking the content center network as an example, we introduce the communication flow of the data packets in detail. In addition, we have a more comprehensive understanding and research on the existing cache policies. There are three main factors affecting caching in the network: topologically dependent cache influencing factors. Considering the cache size and topology importance of the node; The content dependent cache influence factor is determined by the node's perception of the content. Other types of cache factors mainly include the state of nodes and the number of hops from which the content hits the node. All the factors mentioned above will affect the efficiency of the cache system in the network. Therefore, this paper proposes a cache placement strategy based on multi-factor analysis. How to comprehensively evaluate the impact of each cache factor on node cache efficiency is the key of this paper. The Analytic hierarchy process (AHP) is a scientific and accurate decision making method, which is widely used in the decision analysis of multi-objective complex problems. Firstly, the representative cache impact factors of each node are screened, and then the system evaluation model is constructed according to the analytic hierarchy process (AHP), and the weight of the selected cache impact factor is evaluated at the hit node. Combined with the real-time statistical simulation data in the network, the best cache placement node is selected from the transmission path of interest packets, and the final cache decision is made. In the last chapter, we model and simulate the cache placement strategy, perfect the node model and deploy the policy. The simulation results show that the cache placement strategy based on multi-factor analysis and the default LCE policy in the content-centric network are proposed in this paper. Compared with the probabilistic cache Probcache strategy, the performance of the cache hit ratio, hit average hops and server load is improved, and the feasibility analysis results are obtained, which verify the feasibility and correctness of this strategy.
【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】