An adaptation aware hybrid client-cache approach for video delivery with dynamic adaptive streaming over HTTP

Abstract

The push from online video content providers to replace traditional broadcast television comes with the added expectation of comparable Quality of Experience (QoE) of video streaming. In-network caching, combined with prefetching, can be useful to improve DASH video delivery latency. In this paper, we propose an Adaptation Aware Hybrid Client-Cache (AAHCC) framework, a cache pre-fetching scheme that prefetch bitrate using an adaptation algorithm based on throughput measurements from the client and the forecasted throughput at the cache. For evaluation, we ran the experiments in a network with and without background traffic on GENI distributed testbed. Our results show that AAHCC by predicting the correct bitrate for the next segment reduces the number of unused pre-fetches and improve the byte-hit rate compare to a basic client-driven pre-fetch scheme.