QoE-driven resource allocation for massive video distribution

Abstract

Massive video delivery systems employ the HTTP protocol and multiple Content Delivery Networks (CDNs), which serve the content to the end-users on behalf of the video providers and guarantee scalability and Quality of Experience (QoE). In this paper, a Video Control Plane (VCP) is presented which monitors the QoE delivered by any of the CDN belonging to its pool and selects the most performing one when a new video request is received. The VCP employs a continuously updated prediction of the CDNs performances based on the feedback sent by the video clients and computed through a k-NN regression algorithm. The proposed VCP has been evaluated through simulations and shows significant performance improvement in terms of QoE delivered to the user.