Per-chunk caching for video streaming from a vehicular cloud

Caching content at the edge of mobile networks is considered as a promising way to deal with the data tsunami. In addition to caching at fixed base stations or user devices, it has been recently proposed that an architecture with public or private transportation acting as mobile relays and caches might be a promising middle ground. In previous work, we have assumed users are streaming video files, and analyzed how many replicas of each video file to cache in such a vehicular fleet, towards minimizing the amount of bits per file downloaded from (expensive) infrastructure links. However, this work has been assuming that a vehicle will store the entire content, or none of it. In practice, later chunks have an inherent "delay tolerance" as there is more time to find them before they must be played out. What is more, numerous studies as well as everyday experience suggest that most files (e.g. YouTube) ones are not entirely watched. This makes the previous policies suboptimal, as fewer (or no) replicas could be allocated to late chunks of a file and more to the most popular chunks. In this work, we formulate an optimization problem to compute the optimal allocation per chunk, to minimize the load on the cellular infrastructure, and we show that significant performance gains can be achieved compared to per content allocation policies.