A New Content Popularity Probability Based Cache Placement and Replacement Plan in CCN
Qin Wang, Xinqi Zhu, Yiyang Ni, Li Wei Gu, Haitao Zhao, Hongbo Zhu · 2019
To improve the distribution efficiency and reduce the transmission redundancy for big-scale multimedia traffic, this paper proposes a new content cache placement and replacement strategy in Content Centric Networking (CCN) architecture. We first construct a content classification model by dividing interest packets into different types (such as subject, format). Then we build the user interest model by calculating the matching degrees of users' requests for subject and format content. Finally, we consider the routers' distances as weights to denote their probabilities to cache content, which is the caching probability model. Based on the three models, a novel content popularity probability (CPP) matrix is formulated for each router. Different from the current popular LRU and LFU replacement plans, the packets are placed/replaced dynamically at selected routers whose corresponding CPP matrix values are maximum. The feedback factor to indicate whether the content has been cached and the routers' storage capabilities are both taken into account. To further improve the system performance, a modified cache plan combined with the Leave Copy Down (LCD) scheme (Modified CPP LCD+LRU) is built as well. By comparing the proposed cache plan with current ones, such as LCE, the proposed plan reduces the traffic delivery distance significantly.