On Performance of Cache Policies in Named Data Networking

Abstract

Named Data Network (NDN) is gaining increasingly concerns, as an important direction of the future Internet architecture research centered on content. Content caching has played a key role in NDN. Existing cache replacement policies like Least Frequently Used (LFU) and Least Recently Used (LRU) have failed to make full use of the popularity of contents, which leads to a low cache efficiency in the dynamic network. In order to make the best use of content popularity in the cache strategy, this paper has proposed a cache replacement policy based on content popularity (CCP), and designed the data structure format and replacement algorithm. For fully studying and analyzing the performance of different cache policies in NDN in terms of network throughput, server load and cache hit ratio, we have done a lot of simulations to show how they will improve the network. The simulation results show that our proposed CCP can significantly decrease the server load with a higher cache hit ratio and increase the network capacity at the same time compared with LRU and LFU. And the average throughput is reduced significantly by nearly 47% in comparison to that of the case without in-networking caching. Moreover, it also shows the performance under different sizes of content store. The effectiveness of the CCP strategy is proved during the simulation.