Popularity based probabilistic caching strategy design for named data networking

Named Data Networking (NDN) is a promising future network architecture. In NDN, caching is widely deployed to improve content delivery. A lot of caching strategies have been proposed to improve caching performance, such as complicated cooperative caching and non-cooperative caching. No-cooperative Probability caching is proposed to increase the content variety across caching nodes, although it enjoys the feasibility due to its simplicity, it basically caches content randomly without special intention and usually caches content which would never be used again. To solve this problem, Popularity based Probabilistic Caching (PPC) is proposed in this paper. PPC decides whether content would be reused based on their popularity and hence cache them with different possibility. Content popularity consists of two aspects, the immediate local popularity and the potential deduced popularity. PPC considers both factors to make caching decisions. Simulation is carried out to compare PPC and other existing caching strategies to prove the improvement, and the results show that the general traffic is reduced, while the hit ratio of caching is enhanced.