Adaptive Coded Caching Scheme for Multi-layer Videos in Heterogeneous Broadcast Networks

Coded caching provides an opportunity to reduce data traffic load in peak hours via coded multicast transmissions. This paper studies the coded caching problem for video contents in heterogeneous packet erasure channels. The video contents are encoded in a scalable manner, where each user's decoding quality depends on the number of layers it receives. Two main challenges lie in this problem. (1) In heterogeneous erasure broadcast channels, users may receive varying numbers of packets. Thus, the performance of the system will be limited by the worst channel condition of users, which is referred to as the “worst-user effect” in existing works. (2) In a scenario where users require different video qualities, different numbers of layers may be transmitted to the users, which indicates a reduction in the multicast opportunities. In this paper, we propose a novel coded caching scheme that enables video quality adaptation, addressing these two challenges at once. The insight is to align the number of layers each user requests with its channel condition, ensuring that the amount of data each user attempts to receive is the same, thus maximizing the multicast opportunities. Analyses show that the transmission rate is determined by the average erasure probability among users, thus alleviating the worst-user effect. Simulation results illustrate the superior performance of the proposed scheme compared to state-of-the-art methods.