Enhancing the Crowdsourced Live Streaming

Rui-Xiao Zhang,Tianchi Huang,Ming Ma,Haitian Pang,Xin Yao,Chenglei Wu,Lifeng Sun
DOI: https://doi.org/10.1145/3304112.3325607
2019-01-01
Abstract:With the growing demand for crowdsourced live streaming (CLS), how to schedule the large-scale dynamic viewers effectively among different Content Delivery Network (CDN) providers has become one of the most significant challenges for CLS platforms. Although abundant algorithms have been proposed in recent years, they suffer from a critical limitation: due to their inaccurate feature engineering or naive rules, they cannot optimally schedule viewers. To address this concern, we propose LTS (Learn to schedule), a deep reinforcement learning (DRL) based scheduling approach that can dynamically adapt to the variation of both viewer traffics and CDN performance. After the extensive evaluation the real data from a leading CLS platform in China, we demonstrate that LTS improves the average quality of experience (QoE) over state-of-the-art approach by 8.71%-15.63%.
What problem does this paper attempt to address?