Content size-aware edge caching: a size-weighted popularity-based approach

In this paper, content caching is considered at the edge of the network with an objective of offloading recurrent traffic on the capacity-stringent backhaul links to the vicinity of end users. A radio access network equipped with edge servers is considered for caching contents of various sizes, based on which problems of maximizing the edge cache-hit-ratio and minimizing the average content-provisioning cost are respectively formulated. To solve the underlying 0-1 Knapsack problem, a size-weighted popularity (SWP)-based caching framework is proposed, where both content popularity and content size are taken into account when determining the contents to be cached. Depending on the available knowledge and the manner in which the contents are pre-fetched and cached at the edge servers, two algorithms: proactive and reactive, are proposed for the implementation of SWP-based caching. Simulation results are presented to evaluate the performance of our proposed algorithms. We observe a fundamental tradeoff between the average content-provisioning cost and the cache-hit-ratio, and the proactive algorithm outperforms the reactive algorithm.