A Data Caching Scheme Based on Node Classification in Named Data Networking
-
摘要: 缓存是命名数据网络(named data networking, NDN)有别于传统网络最突出的特性之一，NDN中默认所有节点都具有缓存所有经过数据的功能.这种“处处缓存”策略导致网内大量冗余数据的产生，使网内缓存被严重浪费.针对上述问题，首次提出了一种基于节点分类(based on node classification, BNC)的数据存储策略.基于节点位置的不同，将数据返回客户端所经过的节点分为“边缘”类节点与“核心”类节点.当数据经过“核心”类节点时，通过权衡该类节点的位置与数据在不同节点的流行度分布，将数据存储在对其他节点最有利的节点中；当数据经过“边缘”类节点时，通过该数据流行度来选择最有利于客户端的位置.仿真结果表明，提出的策略将有效提高数据命中率，减少数据请求时延和距离.
-
关键词:
- 命名数据网络 /
- 节点分类数据存储策略 /
- 网内存储 /
- 冗余数据 /
- 内容中心网络
Abstract: Compared with the traditional Internet, in-networking caching is one of the most distinguishable features in named data networking (NDN). In NDN, a node caches every passing data packet as a default model. The caching scheme generates a large number of redundant data in in-networking. As a consequence, the networking cache resource is wasted seriously. To solve the problem, a caching scheme based on node classification (BNC) is proposed firstly in this paper. Based on different node positions, the nodes that data packet passes through are divided into two types: “edge” type and “core” type. When data packet passes through the “core” type nodes, by considering location and data popularity distribution at different nodes, it is cached in a node which is beneficial to other nodes. When the data packet passes through the “edge” nodes, a node is selected through data popularity to be beneficial to the client. The simulation results show that the proposed scheme can efficiently improve the in-network hit ratio and reduce the delay and hops of getting the data packet. -
计量
- 文章访问数:
- HTML全文浏览量: 0
- PDF下载量:
下载:
