Delivery Networks with Shared Caches and Correlated Content

The exponential growth of Internet traffic places a substantial strain on delivery networks. Despite recent technological advancements, delivery networks still struggle with high data rates and suffer from the imbalanced network load between off-peak and peak hours. Content caching effectively mitigates delivery challenges, promoting load balance and improving quality of service (QoS) by placing frequently accessed content closer to end users. This study concentrates on delivery networks with shared caches and correlated sources. This setting is highly beneficial for existing and emerging wireless networks as the library of a caching network could consist of content with a high degree of similarity in many practical applications. Therefore, exploiting the similarity among library content can be leveraged to reduce the delivery rate in such networks. This research investigates the caching network from an information-theoretic standpoint, formulating the caching problem as a form of distributed source coding (DSC) with side information at the decoder. From this perspective, several fundamental questions arise: Which content is most appropriate for caching as side information? Also, how should the chosen content be strategically placed? In addressing the posed questions, this study introduces an automatic clustering scheme that factors in both content similarity and popularity. This study investigates content placement and delivery strategies across varied network setups. The evaluation results clearly show that the proposed solution significantly enhances network efficiency compared to traditional caching methods. This enhancement is attributed to the innovative problem formulation and the effective extraction of side information in our study.