Design and evaluation of a utility-based caching mechanism for information-centric networks

Information-centric networking (ICN) is one promising direction for the future Internet, and how to manage the in-network caching resources is a fundamental problem in ICN. In this paper, we address the problem by proposing a utility-based caching mechanism. In the mechanism, network nodes track the utilities of the contents that they have ever cached, and en-route nodes cooperate to make the caching decisions. For enabling the utility tracking, we introduce a novel component named Tracking Store in ICN routers, and develop two methodologies for implementing this component based on dynamic LRU queue and time-decaying Bloom filter (TBF). Through analysis and extensive simulations using real-world topologies, we show that at a sustainable router overhead, our proposed mechanism, with both of its implementations, achieves a superior caching performance than existing solutions under various content popularity scenarios.We also explore the inherent tradeoff of the mechanism, and provide guideline for its realworld deployment.