Cognitive Cellular Content Delivery Networks: Cross-Layer Design and Analysis

Cellular communication networks are plagued with redundant capacity, which results in low infrastructure utilization and low cost-effectiveness of network capital investments. This paper proposes a context-aware system that recycles the redundant capacity to provide cognitive secondary content delivery services. The proposed system applies bottom- up cross-layer design to provide best-effort access to recommended contents tailored to the real-time traffic condition. To characterize the secondary service performance of the proposed system, a discrete time Markov model (DTMC) is adopted for a novel cross-layer analysis that incorporates physical layer, data link layer and application layer. Unlike existing literature, our analysis reveals the impact of content file size (session length) on various quality-of- service (QoS) metrics such as blocking probability, average throughput, and delay distribution. Our analyses provide useful guidelines for cross-layer optimization of cognitive content delivery systems.