摘要
移动终端有限的处理能力和无线带宽制约了高清视频应用的进一步发展.针对这一问题,提出基于Docker的混合云视频直播体系,减少直播流传输时延、转码开销和传输成本.该直播体系利用私有云的可定制性降低直播流初始转码的时延.同时,设计转码通道工作量预测模型、延迟预测模型和直播流分发算法.利用公有云弹性可扩展、分布范围广的特点,考虑多个移动终端转码通道重叠的部分,减少直播流转码开销和在多个数据中心之间传输时的成本.实验结果表明,该直播体系有效减少了直播流传输时延,转码开销和传输成本.
The limitations in processing capability and wireless bandwidth of mobile devices is the main constraint on development of high-definition video applications. To address this challenge, this paper proposes a Docker-based hybrid cloud live streaming frame-work, focusing on reducing the transmission delay, transcoding overhead and transmission cost of live streaming. The framework utilizes the customizability of private cloud to reduce initial transcoding overhead of live streaming while decreasing transmission cost and delay among multiple datacenters, followed by the design of models in terms of transcoding pipeline workload prediction, delay prediction and live streaming delivery. This work takes advantage of elasticity and public cloud while considering the correlation among transcoding pipelines. Our experiment results show that this framework can significantly reduce the transmission delay,transcoding overhead and transmission cost of live streaming.
出处
《小型微型计算机系统》
CSCD
北大核心
2017年第10期2312-2317,共6页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(61502230
61073197)资助
江苏省自然科学基金项目(BK20150960)资助
江苏省普通高校自然科学研究项目(15KJB520015)资助
中美计算机科学研究中心开放课题项目(KJR16078)资助
