边缘计算中面向互动直播的用户分配策略

刘伟; 张骁宇; 杜薇; 彭若涛

doi:10.7544/issn1000-1239.202220113

边缘计算中面向互动直播的用户分配策略

刘伟^{1, 2, 3, 4,},
张骁宇^1,,
杜薇^{1, 2, ,},
彭若涛^1,

1.
武汉理工大学计算机与人工智能学院　武汉　430070
2.
交通物联网湖北重点实验室（武汉理工大学）　武汉　430070
3.
嵌入式系统与服务计算教育部重点实验室（同济大学）　上海　201804
4.
处理器芯片全国重点实验室（中国科学院计算技术研究所）　北京　100190

基金项目: 湖北省自然科学基金面上项目(2020CFB749)；同济大学嵌入式系统与服务计算教育部重点实验室开放基金项目(ESSCKF2018-05)；中科院计算技术研究所计算机体系结构国家重点实验室开放课题(CARCHB202015)

详细信息

作者简介:
刘伟: 1978年生. 博士，副教授. CCF会员. 主要研究方向为云计算与移动边缘计算

张骁宇: 1993年生. 硕士研究生. 主要研究方向为移动边缘计算

杜薇: 1978年生. 博士，副教授. CCF会员. 主要研究方向为服务计算、移动边缘计算

彭若涛: 1995年生. 硕士研究生. 主要研究方向为移动边缘计算

通讯作者:
杜薇（whutduwei@whut.edu.cn）

中图分类号: TP393
计量
- 文章访问数: 171
- HTML全文浏览量: 22
- PDF下载量: 85
出版历程
- 收稿日期: 2022-01-24
- 修回日期: 2022-10-09
- 网络出版日期: 2023-05-22
- 刊出日期: 2023-07-31

User Allocation Strategy for Interactive Live Streaming in Edge Computing

Liu Wei^{1, 2, 3, 4,},
Zhang Xiaoyu^1,,
Du Wei^{1, 2, ,},
Peng Ruotao^1,

1.
School of Computer and Artificial Intelligence, Wuhan University of Technology, Wuhan 430070
2.
Hubei Key Laboratory of Transportation Internet of Things ( Wuhan University of Technology ), Wuhan 430070
3.
Key Laboratory of Embedded System and Service Computing (Tongji University), Ministry of Education, Shanghai 201804
4.
State Key Lab of Processors ( Institute of Computing Technology, Chinese Academy of Sciences ), Beijing 100190

Funds: This work was supported by the General Program of Natural Science Foundation of Hubei Province (2020CFB749), the Open Fund of Key Laboratory of Embedded System and Service Computing (Tongji University) Ministry of Education (ESSCKF2018-05), and the Open Fund of State Key Laboratory of Computer Architecture (Institute of Computing Technology, Chinese Academy of Sciences) (CARCHB202015).

More Information

Author Bio:
Liu Wei: born in 1978. PhD, associate professor. Member of CCF. His main research interests include cloud computing and mobile edge computing

Zhang Xiaoyu: born in 1993. Master candidate. His main research interest includes mobile edge computing. (zxy626@whut.edu.cn)

Du Wei: born in 1978. PhD, associate professor. Member of CCF. Her main research interests include service computing and mobile edge computing

Peng Ruotao: born in 1995. Master condidate. His main research interest includes mobile edge computing. （ruotao923@whut.edu.cn）

摘要

摘要:
将互动直播部署在边缘计算环境中，可以在网络边缘对直播视频进行转码和传输，通过用户附近的边缘服务器提供低延迟的直播服务. 然而，在多边缘服务器、多用户场景下存在着直播用户分配问题，导致直播用户体验质量(quality of experience, QoE)无法得到保证. 为了提高直播用户QoE，需要根据用户的个性化需求合理地分配服务器资源. 首先分析真实数据集，发现大多数用户处于多基站重叠覆盖区域内，并且不同用户的互动需求存在差异；然后根据互动直播的特点提出一种适用于边缘计算场景的用户QoE模型，该模型综合考虑了直播用户的视频质量和互动体验；最后设计一种高效的直播用户分配算法，优化了多边缘服务器重叠覆盖区域内的直播用户QoE. 仿真实验表明，所提出的用户分配策略可为用户提供高码率和低延迟的直播视频，同时能有效降低边缘服务器切换次数和码率抖动，使直播用户QoE相较于其他策略提升超过19%.
- 互动直播 /
- 边缘计算 /
- 用户体验 /
- 重叠覆盖区域 /
- 用户分配
Abstract:
Deploying interactive live streaming in edge computing environment enable us to offload the transcoding and delivery cost to network edge and provide service with lower latency via the edge servers near the users. However, there is a problem of user allocation in the real complex multi-server and multi-user scenarios, leading to a poor quality of experience (QoE). In order to improve QoE of overlapping coverage areas, it is necessary to select edge servers for individual live streaming users according to their needs and to allocate server resources reasonably. First, the analysis of real-world data sets reveals that most users are in the overlapping coverage area of multiple base stations whose interaction needs vary. Then, a QoE model suitable for edge computing scenarios is proposed according to the analysis, which is based on the characteristics of interactive live streaming and comprehensively considers the interactive and video viewing experience of users. Finally, an efficient live streaming user allocation algorithm is designed to optimize the user QoE in the overlapping coverage area of multiple edge servers. Simulation experiments show that this strategy can provide users with high bit rate and low latency streaming while controlling the edge servers switching and bit rate jitter, thus improving the QoE of users by more than 19% compared with the other strategies.
- interactive live streaming /
- edge computing /
- user experience /
- overlapping coverage area /
- user allocation

HTML全文

图 1 可连接基站数对应的人数占比分布

Figure 1. Distribution of the proportion of people corresponding to the number of connectable base stations

下载: 全尺寸图片幻灯片

图 2 直播间数量和用户数量分布

Figure 2. The number of live streaming rooms and distribution of the number of users

下载: 全尺寸图片幻灯片

图 3 弹幕数量对应的人数累计占比分布

Figure 3. Distribution of cumulative proportion of the number of people corresponding to the number of barrage

下载: 全尺寸图片幻灯片

图 4 各直播间人均弹幕数分布

Figure 4. Distribution of the number of barrages per capita in each live streaming room

下载: 全尺寸图片幻灯片

图 5 基于边缘计算的直播视频分发架构

Figure 5. Live streaming distribution architecture based on edge computing

下载: 全尺寸图片幻灯片

图 6 LUA算法流程图

Figure 6. Flow chart of LUA algorithm

下载: 全尺寸图片幻灯片

图 7 用户数量对码率的影响

Figure 7. Effect of the number of users on bit rate

下载: 全尺寸图片幻灯片

图 8 用户数量对延迟的影响

Figure 8. Effect of the number of users on delay

下载: 全尺寸图片幻灯片

图 9 用户数量对服务器切换的影响

Figure 9. Effect of the number of users on server switching

下载: 全尺寸图片幻灯片

图 10 用户数量对码率抖动的影响

Figure 10. Effect of the number of users on bit rate jitter

下载: 全尺寸图片幻灯片

图 11 用户数量对互动体验的影响

Figure 11. Effect of the number of users on interactive experience

下载: 全尺寸图片幻灯片

图 12 用户数量对用户QoE的影响

Figure 12. Effect of the number of users on QoE

下载: 全尺寸图片幻灯片

图 13 边缘服务器数量对码率的影响

Figure 13. Effect of the number of edge servers on bit rate

下载: 全尺寸图片幻灯片

图 14 边缘服务器数量对延迟的影响

Figure 14. Effect of the number of edge servers on delay

下载: 全尺寸图片幻灯片

图 15 边缘服务器数量对服务器切换的影响

Figure 15. Effect of the number of edge servers on server switching

下载: 全尺寸图片幻灯片

图 16 边缘服务器数量对码率抖动的影响

Figure 16. Effect of the number of edge servers on bit rate jitter

下载: 全尺寸图片幻灯片

图 17 边缘服务器数量对互动体验的影响

Figure 17. Effect of the number of edge servers on interactive experience

下载: 全尺寸图片幻灯片

图 18 边缘服务器数量对用户QoE的影响

Figure 18. Effect of the number of edge servers on user QoE

下载: 全尺寸图片幻灯片

图 19 用户互动需求对延迟的影响

Figure 19. Effect of user interaction demand on delay

下载: 全尺寸图片幻灯片

图 20 用户互动需求对互动体验的影响

Figure 20. Effect of user interaction demand on interactive experience

下载: 全尺寸图片幻灯片

图 21 权重a对延迟和互动体验的影响

Figure 21. Effect of the values of weight a on delay and interactive experience

下载: 全尺寸图片幻灯片

图 22 权重a对服务器切换次数和码率的影响

Figure 22. Effect of the values of weight a on the number of server switching and bit rate

下载: 全尺寸图片幻灯片

表 1 主要符号含义

Table 1 Key Notations Meanings

符号	含义
${b}_{v}$	用户 $v$ 所观看的主播
${m}_{{b}_{v}}$	主播 ${b}_{v}$ 连接的边缘服务器
${R}_{v}$	用户 $v$ 当前时段观看直播的平均码率
$M\left(v\right)$	用户 $v$ 所在位置附近可连接到的边缘服务器集合
${x}_{v,m}$	用户 $v$ 和边缘服务器 $m$ 之间的连接关系
${T}_{{b}_{v},{m}_{{b}_{v}}}$	主播 ${b}_{v}$ 到边缘服务器 ${m}_{{b}_{v}}$ 之间的直播视频传播延迟
${T}_{{m}_{{b}_{v}},m}$	边缘服务器 ${m}_{{b}_{v}}$ 到边缘服务器 $m$ 之间的直播视频传播延迟
${T}_{m,v}$	边缘服务器 $m$ 到用户 $v$ 之间的传播延迟
${F}_{v}$	用户 $v$ 的互动频率
${F}_{{b}_{v}}$	用户 $v$ 当前所在直播间内主播的平均互动频率
${I}_{v}$	用户 $v$ 的互动需求值
${R}_{v}^{*}$	用户 $v$ 所期望的码率
${R}_{v}^{-}$	用户 $v$ 前一时间段的平均码率

下载: 导出CSV

表 2 视频码率参数

Table 2 Video Bit Rate Parameters

分辨率/Px	码率范围/Mbps	平均码率/Mbps
360	0.4~1	0.7
480	0.5~2	1.2
720	1.5~4	1.8
1080	3~6	3.5

下载: 导出CSV

表 3 主要参数设置

Table 3 Main Parameters Setting

参数	取值
边缘服务器覆盖半径/m	300
边缘服务器直播转码总计算资源/vCPUs	15
边缘服务器直播传输总带宽资源/Mbps	40
边缘服务器转码延迟G/(ms·Mbps⁻¹)	50
边缘服务器转码计算开销H/(vCPU·Mbps⁻¹)	0.5
边缘服务器间传输延迟/ms	[100,300]
边缘服务器到用户传输延迟/ms	[5,15]
服务切换用户体验损失值 $s$	2.5
码率下降用户体验损失值k	1.5
赠礼金额影响因素 ${Y}_{v}$	[1,3]

下载: 导出CSV

表 4 实验设置

Table 4 Experimental Setting

边缘服务器数量	用户数量	互动需求值范围
6	50~150	0~2
4~12	100	0~2
6	100	0~1，1~2，2~3

下载: 导出CSV

表 5 实验结果对比

Table 5 Comparison of Experimental Results

算法	QoE	平均码率/Mbps	平均延迟/ms	平均互动体验
最优解	37.00	3.5	212.6	2.97
LUA	31.82	2.45	223.9	2.49

下载: 导出CSV

参考文献(25)

[1]	Cisco V. Cisco visual networking index: Forecast and trends, 2017–2022[EB/OL]. White Paper , 2018[2021-07-10].https://www.docin.com/p-2219756448.html
[2]	Jia Qingmin, Xie Renchao, Huang Tao, et al. The collaboration for content delivery and network infrastructures: A survey[J]. IEEE Access, 2017, 5: 18088−18106 doi: 10.1109/ACCESS.2017.2715824
[3]	Deng J, Tyson G, Cuadrado F, et al. Internet scale user-generated live video streaming: The twitch case[C] //Proc of Int Conf on Passive and Active Network Measurement. Berlin: Springer, 2017: 60−71
[4]	Yan Huan, Lin T, Gao Chuhan, et al. On the understanding of video streaming viewing behaviors across different content providers[J]. IEEE Transactions on Network and Service Management, 2018, 15(99): 444−457
[5]	贾庆民,李子姝,李诚成,等. 基于MEC的移动网络直播视频分发机制研究[J]. 信息通信技术,2018(5):40−44 doi: 10.3969/j.issn.1674-1285.2018.05.006 Jia Qingmin, Li Zishu, Li Chengcheng, et al. Research on live video distribution mechanism of mobile network based on MEC[J]. Information and Communication Technology, 2018(5): 40−44 (in Chinese) doi: 10.3969/j.issn.1674-1285.2018.05.006
[6]	Shi Weisong, Cao Jie, Zhang Quan, et al. Edge computing: Vision and challenges[J]. IEEE Internet of Things Journal, 2016, 3(5): 637−646 doi: 10.1109/JIOT.2016.2579198
[7]	施巍松,张星洲,王一帆,等. 边缘计算:现状与展望[J]. 计算机研究与发展,2019,56(1):69−89 doi: 10.7544/issn1000-1239.2019.20180760 Shi Weisong, Zhang Xingzhou, Wang Yifan, et al. Edge computing: State-of-the-art and future directions[J]. Journal of Computer Research and Development, 2019, 56(1): 69−89 (in Chinese) doi: 10.7544/issn1000-1239.2019.20180760
[8]	张秋平,孙胜,刘敏,等. 面向多边缘设备协作的任务卸载和服务缓存在线联合优化机制[J]. 计算机研究与发展,2021,58(6):1318−1339 doi: 10.7544/issn1000-1239.2021.20201088 Zhang Qiuping, Sun Sheng, Liu Min, et al. Online joint optimization mechanism of task offloading and service caching for multi-edge device collaboration[J]. Journal of Computer Research and Development, 2021, 58(6): 1318−1339 (in Chinese) doi: 10.7544/issn1000-1239.2021.20201088
[9]	李智勇,王琦,陈一凡,等. 车辆边缘计算环境下任务卸载研究综述[J]. 计算机学报,2021,44(5):963−982 doi: 10.11897/SP.J.1016.2021.00963 Li Zhiyong, Wang Qi, Chen Yifan, et al. A survey on task offloading research in vehicular edge computing[J]. Chinese Journal of Computers, 2021, 44(5): 963−982 (in Chinese) doi: 10.11897/SP.J.1016.2021.00963
[10]	张展,张宪琦,左德承,等. 面向边缘计算的目标追踪应用部署策略研究[J]. 软件学报,2020,31(9):2691−2708 doi: 10.13328/j.cnki.jos.005938 Zhang Zhan, Zhang Xianqi, Zuo Decheng, et al. Reserch on target tracking application deployment strategy for edge computing[J]. Journal of Software, 2020, 31(9): 2691−2708 (in Chinese) doi: 10.13328/j.cnki.jos.005938
[11]	Pang Haitian, Zhang Cong, Wang Fangxin, et al. Optimizing personalized interaction experience in crowd-interactive livecast: A cloud-edge approach[C] //Proc of the 26th ACM Multimedia Conf. New York: ACM, 2018: 1217−1225
[12]	Zhang Zhicai, Wang Ru, Yu F R, et al. QoS aware transcoding for live streaming in edge-clouds aided hetNets: An enhanced actor-critic approach[J]. IEEE Transactions on Vehicular Technology, 2019, 68(11): 11295−11308 doi: 10.1109/TVT.2019.2942629
[13]	Ge Chang, Wang Ning, Chai W, et al. QoE-assured 4K HTTP live streaming via transient segment holding at mobile edge[J]. IEEE Journal on Selected Areas in Communications, 2018, 36(8): 1816−1830 doi: 10.1109/JSAC.2018.2845000
[14]	Hu Xinjue, Quan Wei, Guo Tao, et al. Mobile edge assisted live streaming system for omnidirectional video[J]. Mobile Information Systems, 2019, 2019(9): 1−15
[15]	Hu Yunchao, Patel M, Sabella D, et al. Mobile edge computing—A key technology towards 5G[J]. ETSI White Paper, 2015, 11: 1−16
[16]	He Qiang, Cui Guangming, Zhang Xuyun, et al. A game-theoretical approach for user allocation in edge computing environment[J]. IEEE Transactions on Parallel and Distributed Systems, 2019, 31(3): 515−529
[17]	Peng Qinglan, Xia Yunni, Feng Zeng, et al. Mobility-aware and migration-enabled online edge user allocation in mobile edge computing[C] //Proc of 2019 IEEE Int Conf on Web Services. Piscataway, NJ: IEEE, 2019: 91−98
[18]	Wang Ziyi, Cui Yong, Hu Xiaoyu, et al. MultiLive: Adaptive bitrate control for low-delay multi-party interactive live streaming[C] //Proc of IEEE Conf on Computer Communications. Piscataway, NJ: IEEE, 2020: 1093−1102
[19]	Liu Junquan, Zhang Weizhan, Huang Shouqin, et al. QoE-driven HAS live video channel placement in the media cloud[J]. IEEE Transactions on Multimedia, 2020, 23: 1530−1541
[20]	Gao Bin, Zhou Zhi, Liu Fangming, et al. Winning at the starting line: Joint network selection and service placement for mobile edge computing[C] //Proc of IEEE Conf on Computer Communications. Piscataway, NJ: IEEE, 2019: 1459−1467
[21]	Jiang Junchen, Das R, Ananthanarayanan G, et al. Via: Improving internet telephony call quality using predictive relay selection[C] //Proc of the 29th ACM SIGCOMM Conf. New York: ACM, 2016: 286−299
[22]	博客园. 实时音视频互动系列: 又拍云UTUN网络详解[EB/OL]. 2017[2021-07-14].https://www.cnblogs.com/upyun/p/7169392.html Cnblogs. Real time audio and video interactive series: Detailed explanation of cloud UTUN network [EB/OL]. 2017[2021-07-14].https://www.cnblogs.com/upyun/p/7169392.html (in Chinese)
[23]	Zhang Weiwen, Wen Yonggang, Chen Zhenzhong , et al. QoE-driven cache management for HTTP adaptive bit rate (ABR) streaming over wireless networks[C] //Proc of 2012 IEEE Global Communications Conf. Piscataway, NJ: IEEE, 2012: 1951−1956
[24]	Wang Xiaodong, Tian Ye, Lan Rongheng, et al. Beyond the watching: Understanding viewer interactions in crowdsourced live video broadcasting services[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2019, 29(11): 3454−3468 doi: 10.1109/TCSVT.2018.2877694
[25]	He Qiyun, Liu Jiangchuan, Wang Chonggang, et al. Coping with heterogeneous video contributors and viewers in crowdsourced live streaming: A cloud-based approach[J]. IEEE Transactions on Multimedia, 2016, 18(5): 916−928 doi: 10.1109/TMM.2016.2544698

施引文献(17)

期刊类型引用(7)

1.	张淑芬，张宏扬，任志强，陈学斌. 联邦学习的公平性综述. 计算机应用. 2025(01): 1-14 . 百度学术
2.	朱智韬，司世景，王健宗，程宁，孔令炜，黄章成，肖京. 联邦学习的公平性研究综述. 大数据. 2024(01): 62-85 . 百度学术
3.	李锦辉，吴毓峰，余涛，潘振宁. 数据孤岛下基于联邦学习的用户电价响应刻画及其应用. 电力系统保护与控制. 2024(06): 164-176 . 百度学术
4.	刘新，刘冬兰，付婷，王勇，常英贤，姚洪磊，罗昕，王睿，张昊. 基于联邦学习的时间序列预测算法. 山东大学学报(工学版). 2024(03): 55-63 . 百度学术
5.	赵泽华，梁美玉，薛哲，李昂，张珉. 基于数据质量评估的高效强化联邦学习节点动态采样优化. 智能系统学报. 2024(06): 1552-1561 . 百度学术
6.	杨秀清，彭长根，刘海，丁红发，汤寒林. 基于数据质量评估的公平联邦学习方案. 计算机与数字工程. 2022(06): 1278-1285 . 百度学术
7.	黎志鹏. 高可靠的联邦学习在图神经网络上的聚合方法. 工业控制计算机. 2022(10): 85-87+90 . 百度学术