Optimizing Media Streaming in Edge-Cloud Systems: Leveraging Geography and User Behavior

With the burgeoning prevalence of short videos on the Internet, the design of video caching mechanisms has become a critical foundation for efficient short video distribution. However, traditional caching strategies with fixed durations and sizes need help to balance storage space utilization and transmission speed. This paper proposes a dynamic approach that utilizes popularity evolution modeling to forecast short video storage requirements, thereby enabling adaptive caching scheme selection. The experimental results demonstrate that the proposed method reduces backhaul traffic, bolsters cache utilization, and diminishes initial buffering delays, underscoring its significant impact on content delivery networks.