Adaptive Low-Delay Video Streaming In Heterogeneous Wireless Networks Using Mprtp

2017 13TH INTERNATIONAL WIRELESS COMMUNICATIONS AND MOBILE COMPUTING CONFERENCE (IWCMC)(2017)

引用 8|浏览3
暂无评分
摘要
As computationally powerful mobile devices with high-resolution screens have become ubiquitous, network operators are facing the problem of coping with the huge traffic load caused by various video streaming applications. One approach to mitigate the problem is to dynamically offload parts of the traffic to Wi-Fi networks that offer higher data rates at lower costs. Offloading can be performed in the most seamless way by deploying multi-path transport protocols that incorporate mechanisms to distribute the load over the available network interfaces. In the present work, we develop such mechanisms that allow to efficiently transmit a low-delay live video stream over multiple paths. We propose a cross-layer solution that operates on the transport and application layers. It consists of two components. The Sender-Side Path Scheduling component allows to use multiple network interfaces for concurrent transmission of video data in order to achieve bandwidth aggregation, and packets loss reduction. The Sender-Side Video Adaptation component dynamically adapts the video bit rate to the network conditions. The joint operation of these two components allows to increase the Quality of Experience of the user by providing a smooth playback at the best possible video quality level with a minimum of playback interruptions and corrupted video frames. In order to efficiently adapt the video quality to the network conditions, our proposed solution leverages Scalable Video Coding. We prototyped the proposed solution using the Multipath Real-Time Transport Protocol, and evaluated it w.r.t. the achieved Quality of Experience. We observed that it allows to reduce the amount of playback interruptions by up to 50%, as compared to two baseline approaches.
更多
查看译文
关键词
multipath video streaming, SVC adaptation, heterogeneous wireless network, MPRTP
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要