• Please send your full manuscript to: jcm@etpub.com

A Popularity-Based Collaborative Caching Algorithm for Content-Centric Networking

Xiaodong Zhu 1,2, Weining Qi 1, Jinlin Wang 1, and Yiqiang Sheng 1
1. National Network New Media Engineering Research Center, Institute of Acoustics, Chinese Academy of Sciences, Beijing, China
2. University of Chinese Academy of Sciences, Beijing, China
Abstract—The key aspect of Content-Centric Networking (CCN) is that the contents are cached by in-network nodes, and hence the caching scheme is important for the efficient use of cache and content delivery. Although the caching scheme in CCN has been widely researched, it is still a great challenge to better utilize in-network caching of CCN with an appropriate caching scheme. In this paper, we propose an effective popularity-based collaborative caching algorithm with the idea to reduce content redundancy within network and make popular contents near clients. The algorithm is lightweight and implicit collaborative with real-time analysis local popularity and little change to the pipeline of CCN. We also analyze that the proposed algorithm is suitable for solving the lack of contents and high external bandwidth charge problems of China Mobile Communications Corporation. The algorithm is simulated with many factors that may have impacts on the performance, and our algorithm is effective, scalable and stable for various topologies, popularity estimate methods and features. The results of simulation show that our algorithm outperforms the existing classical algorithm in access hops and server hit proportion.

Index Terms—Content-Centric networking, China mobile network, popularity-based, collaborative caching

Cite: Xiaodong Zhu, Weining Qi, Jinlin Wang, and Yiqiang Sheng, “A Popularity-Based Collaborative Caching Algorithm for Content-Centric Networking," Journal of Communications, vol. 11, no. 10, pp. 886-895, 2016. Doi: 10.12720/jcm.11.10.886-895
Copyright © 2013-2017 Journal of Communications, All Rights Reserved
E-mail: jcm@etpub.com