Client-driven Adaptive Bitrate Techniques for Media Streaming over HTTP: Initial Findings

Recently, we have seen a huge interest from academia and industry in HTTP adaptive streaming (HAS) solutions. To improve viewer experience, content providers deploy various HAS solutions that include a key element at the client side, the adaptive bitrate (ABR) logic. This client-driven technique aims to dynamically select an suitable bitrate and adapt to the available network conditions. In literature, many ABR techniques have been proposed. However, how the ABR works depends on the choice of content providers and network characteristics. In this paper, we review a set of heuristic-based client-driven ABR techniques for on-demand video delivery in HAS. We conducted a systematic mapping study to identify the primary studies related to bitrate selection in HAS. Based on the heuristic that ABR implements, we classified the client-driven ABR techniques into two categories: throughput-based, and buffer-based. A detailed analysis is given that will help researchers and practitioners for considering these ABR techniques in HAS.