Prioritized Probabilistic Caching Algorithm in Content Centric Networks

Abstract

We propose a new prioritized probabilistic caching scheme for content-centric networks (CCN) to selectively cache more important data more than the others by using heterogeneous caching probabilities. The proposed caching algorithm is used along with Least Recently Used (LRU) cache replacement policy. The caching algorithm allows each CCN node to cache data based on popularity and priority of the data. The popularity of data is determined based on the number of corresponding interest packets sent by content requesters. The priority of data is dictated by the influence of the data to the quality of content reconstructed from the data. With these popularity and priority, the caching probability can be computed by weighing summation. We conduct computer simulations to evaluate the performance of the proposed caching algorithm. Results show that the proposed caching algorithm outperforms a universal caching scheme in terms of content quality at a requester and caching performance such as cache-hit percentage and server load reduction.