Pervasive wireless CDN for greening video streaming to mobile devices

Content Delivery Networks (CDNs) play an important role in today's video distribution solutions. While the need for efficient content delivery is well understood, the energy consumption of various video CDN topologies deserves extra attention since energy considerations are increasingly becoming a key performance factor for system designers and operators. This paper investigates the energy efficiency aspect of a special class of video CDN system, namely pervasive wireless CDN based on Mesh WiFi implementation. We start by analyzing and defining the energy efficiency of a generic video CDN system [1], followed by mathematical analysis and simulation model description for wireless CDNs exploring the idea of intelligent coordination of content placement among wireless access points which serve as local video servers. Computer simulation analysis, based on mixed integer linear programming (MILP), is used to derive optimal content placement and caching strategies while minimizing the energy consumption subject to a constraint in terms of the storage size at each mesh router. The objective function of the MILP model captures the desire to optimize multiple performance factors, in this particular case we are focusing on the QoS/latency time and energy consumption, and the results demonstrate trade-off dependency between QoS and power dissipation. The content placement policy at the most coarse granularity can be described as: chunks of media streaming files are placed across the wireless nodes in such way that on average each chunk is closest to the users most interested in the content, and contextual data (i.e. distribution of the users, popularity of the content, quality of the wireless links between the nodes, etc) is modeled too and consequently the placement of media files/chunks across the nodes is dynamically changed in order to ensure that condition (i) holds. In summary, the paper offers additional insight into the content diffusion strategies as related to a pervasive wireless CDN, and implications of the placement strategies relative to its energy efficiency.