Client-driven network-level QoE fairness for encrypted 'DASH-S'

Abstract

Adaptive video streams, when competing behind a bottleneck link, generate flows that lead to instability, under-utilization, and unfairness. Recent studies suggest this negatively impacts users' perceived quality of experience (QoE). Two general classes of solution exist. Client-side bitrate adaptation improves stability and may achieve flow-rate equality. However, operating in isolation, bitrate adaptation has no ability to establish equal end-user experience, or QoE fairness, among competing clients. Conversely, network services can manage bottleneck resources to achieve QoE fairness, yet the widespread use of HTTPS renders them ineffective. In this paper we step back to evaluate the issue with a broad and fully inclusive view of fairness, or equality, by any definition. Our focal contribution is a constructive argument that makes clear, given ubiquitous use of encrypted HTTP, all notions of QoE fairness or equality can only be achieved when clients and network collaborate. Furthermore, operating boundaries are identified with their features and limits.We then architect and implement client-Driven Video Delivery (cDVD), a proof-ofconcept that provides a client-level API into the network, to explore these boundaries. cDVD measurements reinforce our argument and raise new opportunities for exploration.