Greedy Caching: An optimized content placement strategy for information-centric networks

Most content placement strategies in information-centric networks (ICN) primarily focus on pushing popular content to the network edge, fail to effectively utilize the caches in the network core and provide limited performance improvement. In this paper, we propose Greedy Caching, a content placement strategy that determines the set of content to be cached at each network node so as to maximize the network hit rate. Greedy Caching caches the most popular content at the network edge, recalculates the relative popularity of each piece of content based on the request miss stream from downstream caches and then determines the content to be cached in the network core. We perform exhaustive simulation in the Icarus simulator [1] using realistic Internet topologies (e.g., GARR, GEANT, WIDE, scale-free networks) as well as real-world request stream traces, and demonstrate that Greedy Caching provides significant improvement in content download delay (referred to as latency) over state-of-the-art dynamic caching and routing strategies for ICN for a wide range of simulation parameters. Our simulation results suggest an improvement of 5-28% in latency and 15-50% improvement in hit rate over state-of-the-art policies for synthetic traces.