Video quality as a driver for traffic management with multiple subscriber classes

Video services currently account for a very large portion of the total traffic on the Internet, and this portion is foreseen to keep rising. This trend, coupled with the resource-hungry nature of video services, poses significant problems for network management, if good perceptual quality levels are to be achieved. In mobile networks, in particular, this can be a problem when a cell contains several users streaming video concurrently. In this paper we present a short overview of a multi-faceted mechanism for cross-layer quality-driven traffic management for over-the-top video services at the last hop by considering different subscriber and application types and using them to perform admission control and traffic differentiation. The results obtained show a clear QoE improvement for over-the-top video streams when the proposed mechanisms are in place instead of a simple best-effort policy. Further work on this subject includes the extension and refinement of the traffic classification mechanism used to work on adaptive HTTP-based video streaming schemes, as well as the development of suitable parametric QoE models for them.