Popularity-Aware Content Caching for Distributed Wireless Helper Nodes

Content caching enables end users to obtain contents in a short time. To further reduce the time to access a mobile wireless network, we consider a device-to-device communication scenario where contents are distributed and users may obtain contents directly from neighbors instead of accessing the base station. For maximizing the cache hit rate per unit of bandwidth consumed, we derive a theoretical model based on the probability distribution of content popularity. To make the problem tractable, we propose a popularity-aware content caching mechanism using the modified weighted Zipf distribution. Through our simulation with a real YouTube trace, the proposed algorithm is compared with the existing well-known least-recently-used algorithm in terms of the achieved hit rate, caching time, cache eviction rate, and consumed bandwidth. The results show that our approach is able to achieve near-optimal performances and at the same time approximate the optimal cache time a cached content resides at the device.