Content delivery network with shared caches

What is it about?

The increasing number of internet users and IoT devices, in addition to the widespread use of social media and streaming services, has put a significant strain on current delivery networks. Content caching has emerged as a viable solution to combat the high delivery data rate and improve the quality of services (QoS) for the end-users in such networks. In this paper, we propose a cache-aided delivery network with correlated sources in a shared cache framework where a different group of users connects to each cache. We address the caching strategy and examine the trade-off between the delivery rate and the memory size from an information-theoretic perspective. We propose a correlation-aware clustering scheme to extract the most efficient side information for the entire library during the placement phase considering the similarity among sources and the maximum distortion constraint in the system. We also formulate the expected delivery rate by joint consideration of the rate-distortion function and caching strategy, where the limit for the maximum allowable distortion in the system is determined based on the Lagrange multipliers technique and reverse water-filling algorithm. Moreover, we introduce the optimum library partitioning formulated to minimize the worst-case delivery rate in the system. We also study the proposed solution in a special case where only one shared cache is available in the network. Our extensive simulations validate the proposed scheme provides a considerable boost in network efficiency compared to legacy caching schemes.

Featured Image