As video delivery systems are tasked with larger and larger content libraries, significantly larger streaming loads across increasingly broader geographies, and a wide diversity of client devices, the need for more dynamic scaling and virtualized provisioning of large on-demand resource pools is apparent.
This paper examines strategies and policies for enabling a logically-shared video delivery load across multiple independent but networked video servers, and provides insight into practical applications such as efficient and scalable models for the origin, mid-tier, or edge caching points of a Content Delivery Network (CDN).