Coded content caching and delivery over wireless networks

Abstract

Coded caching is a novel form of content caching for wireless networks to relieve the congestion during peak-traffic periods by exploiting the temporal variability of wireless traffic. During the off-peak traffic periods, popular contents are partially placed into users' local caches without the knowledge of user demands, and the rest of the contents are transmitted when the demands are revealed during the peak traffic periods. Compared to traditional caching, a significant reduction in the delivery rate is achieved due to coded multicasting opportunities created by the joint design of the placement and delivery phases even when users request distinct contents. In this dissertation, we investigate coded caching under four different network models. First, cache-aided coded content delivery is studied for devices with diverse quality-of-service (QoS) requirements. It is assumed that each user in the system has a distinct QoS requirement. The optimal coded caching scheme is proposed for the special cases of different two-user scenarios, while for the general cases, both centralized and decentralized coded caching schemes are proposed. Next, asynchronous user demands are considered, and both the nonuniform file popularity and the audience retention rates are taken into account. We propose a decentralized coded caching scheme, which shows a significant improvement over uncoded caching in terms of the average delivery rate, or the extension of other known delivery methods to the asynchronous scenario. Then, we study coded caching and delivery taking into account the correlations among the contents in the library. A correlation-aware coded caching scheme is proposed, which is shown to substantially outperform state-of-the-art correlation-ignorant solutions. Finally, together with a correlated library, we consider the situation when the content delivery is carried out over a Gaussian broadcast channel (BC). A lower bound and an upper bound on the minimum transmission power required to satisfy all possible demand combinations is characterized. Our results indicate that coded delivery, and particularly exploiting the correlations among the contents in a cache-aided Gaussain BC can provide significant energy savings.