A Probability Caching Decision Policy with Evicted Copy Up in Content Centric Networking

As a candidate of Next Generation Internet architectures, Content centric networking (CCN) has attracted researchers' attention widely. Adopting a distributed content caching mechanism, CCN works in content oriented mode instead of the traditional host-to host delivery mode. Caching decision policy is the key to the performance of CCN. A novel caching decision policy named Probability Caching with Evicted Copy Up (PCECU) is proposed to meet three targets as far as possible, the dispersion of contents in the network, the network survivable time of contents and the caching ratio of higher popularity contents in the nodes nearby user. Closed-form expressions of the average hit probability in different network levels are derived in an L-level tandem network scenario. The performance of PCECU is compared with that of traditional policies LCE and LCP for UGC and VOD services through theoretical analysis and simulation. Furthermore, two important factors affecting policy performance, cache size and cache probability, are analyzed. The simulation and theoretical analysis results have shown that PCECU can improve the average hit probability and in the meantime reduce the average hit distance of content request.