Low-cost traffic engineering for large-scale live streaming
TAI Jin1, LIU Chenyi2,4, YANG Yuan2,4, WANG Yangyang3, XU Mingwei2,3,4
1. Shenzhen International Graduate School, Tsinghua University, Shenzhen 518055, China; 2. Department of Computer Science and Technology, Tsinghua University, Beijing 100084, China; 3. Institute for Network Sciences and Cyberspace, Tsinghua University, Beijing 100084, China; 4. Beijing National Research Center for Information Science and Technology, Tsinghua University, Beijing 100084, China
Abstract：[Objective] Recent years have witnessed a remarkable rise in popularity among livestream-based network applications, prompting higher expectations for the quality of internet services. However, the public internet infrastructure's highest quality services and shared resources often fail to meet the demanding data transmission requirements of livestreaming applications. The objective of this research is to address the problem of cost-sensitive traffic engineering (TE) in Overlay networks for large-scale livestreaming, providing economically efficient livestreaming flow transmission services while minimizing costs and adhering to quality of service (QoS) requirements. By achieving these objectives, service providers can enhance network performance, optimize resource allocation, and deliver high-quality livestreaming to a wide user base.[Methods] To address the cost-sensitive traffic engineering problem, a comprehensive approach based on time-series optimization and approximate differentiable modeling is adopted. The scenario considered in this research is an application-layer transport network comprising forwarding nodes and virtual links. Business flows are transmitted along paths between forwarding nodes, and these paths may include multiple virtual paths to enhance performance. The problem is formulated as a time-series optimization problem, necessitating the decomposition into a series of time-series-based integer programming problems to simplify the solution process. To handle the nondifferentiable aspects of the problem, an innovative approximating differentiable model is proposed. Path selection is approximated with the Gumbel-Softmax function, a technique allowing for differentiable path selection. Moreover, differentiable functions are employed to approximate the cost and transmission delay functions, ensuring smooth optimization. The Lagrange multiplier method is utilized to transform the problem into an efficient optimization framework. An online routing algorithm (LiveTE) is developed to solve for the set of decision paths, utilizing a gradient descent algorithm to solve the optimization problem iteratively.[Results] The effectiveness of the LiveTE algorithm was evaluated via extensive experimentation and numerical simulations using real data obtained from a running-overlay livestreaming network. The results exhibited considerable cost reductions and improved transmission delay compared to existing methods. LiveTE achieved a remarkable total cost reduction of 52% while simultaneously lowering the average transmission delay by over 6%, highlighting its efficacy in enabling economically efficient livestreaming flow transmission services in overlay networks. The algorithm's ability to optimize resource allocation and improve QoS in large-scale livestreaming scenarios was evident, allowing service providers to enhance network performance and deliver high-quality livestreaming experiences to a diverse user base.[Conclusions] In summary, a comprehensive approach is presented to address the cost-sensitive traffic engineering problem in overlay networks for large-scale livestreaming applications. By formulating the problem as a time-series optimization problem and employing an approximating differentiable model, the proposed LiveTE algorithm achieves remarkable cost reductions while simultaneously improving transmission delay and QoS. The results contribute to the economically efficient delivery of high-quality livestreaming services, allowing service providers to optimize resource allocation and enhance network performance. Furthermore, the proposed LiveTE algorithm provides a valuable solution for service providers seeking to enhance user experience, mitigate costs, and maximize the utilization of overlay networks in livestreaming-based applications.
邰进, 刘辰屹, 杨芫, 王旸旸, 徐明伟. 低成本大规模直播流量工程[J]. 清华大学学报（自然科学版）, 2024, 64(3): 591-600.
TAI Jin, LIU Chenyi, YANG Yuan, WANG Yangyang, XU Mingwei. Low-cost traffic engineering for large-scale live streaming. Journal of Tsinghua University(Science and Technology), 2024, 64(3): 591-600.
[1] SECTOR S, ITU O F. Series G:Transmission systems and media, digital systems and networks international telephone connections and circuits-general definitions[R]. Genève:International Telecommunication Union, 2000. [2] CHOY S, WONG B, SIMON G, et al. The brewing storm in cloud gaming:A measurement study on cloud to end-user latency[C]//201211th Annual Workshop on Network and Systems Support for Games (NetGames). New York, USA:IEEE, 2012:1-6. [3] SALAMATIAN L, ANDERSON S, MATTHEWS J, et al. Curvature-based analysis of network connectivity in private backbone infrastructures[J]. Proceedings of the ACM on Measurement and Analysis of Computing Systems, 2022, 6(1):5. [4] GHOSH A, HA S, CRABBE E, et al. Scalable multi-class traffic management in data center backbone networks[J]. IEEE Journal on Selected Areas in Communications, 2013, 31(12):2673-2684. [5] LIAO X, JIN H, LIU Y, et al. AnySee:Peer-to-peer live streaming[C]//Proceedings of the IEEE INFOCOM 2006. 25th IEEE International Conference on Computer Communications. New York, USA:IEEE, 2006:1-10. [6] PIANESE F, PERINO D, KELLER J, et al. PULSE:An adaptive, incentive-based, unstructured P2P live streaming system[J]. IEEE Transactions on Multimedia, 2007, 9(8):1645-1660. [7] ANDERSEN D, BALAKRISHNAN H, KAASHOEK F, et al. Resilient overlay networks[C]//Proceedings of the Eighteenth ACM Symposium on Operating Systems Principles. Banff, Canada:Association for Computing Machinery, 2001:131-145. [8] SAVAGE S, COLLINS A, HOFFMAN E, et al. The end-to-end effects of Internet path selection[C]//Proceedings of Conference on Applications, Technologies, Architectures, and Protocols for Computer Communication. Cambridge, USA:Association for Computing Machinery, 1999:289-299. [9] MUKERJEE M K, NAYLOR D, JIANG J C, et al. Practical, real-time centralized control for CDN-based live video delivery[C]//Proceedings of 2015 ACM Conference on Special Interest Group on Data Communication. London, UK:Association for Computing Machinery, 2015:311-324. [10] ZHAO M C, ADITYA P, CHEN A, et al. Peer-assisted content distribution in akamai netsession[C]//Proceedings of 2013 Conference on Internet Measurement Conference. Barcelona, Spain:Association for Computing Machinery, 2013:31-42. [11] JIANG J C, DAS R, ANANTHANARAYANAN G, et al. Via:Improving internet telephony call quality using predictive relay selection[C]//Proceedings of 2016 ACM SIGCOMM Conference. Florianopolis, Brazil:Association for Computing Machinery, 2016:286-299. [12] QIAN L, LUO Z G, DU Y J, et al. Cloud computing:An overview[C]//First International Conference, CloudCom 2009 on Cloud Computing. Beijing, China:Springer, 2009:626-631. [13] YIN H, LIU X N, ZHAN T Y, et al. Design and deployment of a hybrid CDN-P2P system for live video streaming:Experiences with LiveSky[C]//Proceedings of the 17th ACM International Conference on Multimedia. Beijing, China:Association for Computing Machinery, 2009:25-34. [14] LI J Y, LI Z Y, LU R, et al. LiveNet:A low-latency video transport network for large-scale live streaming[C]//Proceedings of ACM SIGCOMM 2022 Conference. Amsterdam, Netherlands:Association for Computing Machinery, 2022:812-825. [15] PATHAN M, BUYYA R. A taxonomy of CDNs[M]//BUYYA R, PATHAN M, VAKALI A. Content Delivery Networks. Berlin:Springer, 2008:33-77. [16] LU Z H, WANG Y, YANG Y R. An analysis and comparison of CDN-P2P-hybrid content delivery system and model[J]. Journal of Communications, 2012, 7(3):232-245. [17] DAI J, CHANG Z Y, CHAN S H G. Delay optimization for multi-source multi-channel overlay live streaming[C]//2015 IEEE International Conference on Communications. London, UK:IEEE, 2015:6959-6964. [18] HE T, ZHU K X, CHEN Z P, et al. Popularity-guided cost optimization for live streaming in mobile edge computing[J]. Wireless Communications and Mobile Computing, 2022, 2022:5562995. [19] WANG B L, ZHANG X Y, WANG G, et al. Anatomy of a personalized livestreaming system[C]//Proceedings of 2016 Internet Measurement Conference. Santa Monica, USA:Association for Computing Machinery, 2016:485-498. [20] JAIN P, KUMAR S, WOODERS S, et al. Skyplane:Optimizing transfer cost and throughput using cloud-aware overlays[C]//20th USENIX Symposium on Networked Systems Design and Implementation. Boston, USA:USENIX Association, 2023:1375-1389. [21] SINGH R, AGARWAL S, CALDER M, et al. Cost-effective cloud edge traffic engineering with cascara[C]//Proceedings of the 18th USENIX Symposium on Networked Systems Design and Implementation. New York, USA:USENIX Association, 2021:201-216. [22] LUMEZANU C, BADEN R, SPRING N, et al. Triangle inequality variations in the internet[C]//Proceedings of the 9th ACM SIGCOMM Conference on Internet Measurement. Chicago, USA:Association for Computing Machinery, 2009:177-183. [23] POTAPCZYNSKI A, LOAIZA-GANEM G, CUNNIN- GHAM J P. Invertible Gaussian reparameterization:Revisiting the gumbel-softmax[C]//Proceedings of the 34th International Conference on Neural Information Processing Systems. Vancouver, Canada:Curran Associates Inc., 2020:1032. [24] 周静静, 杨家海, 杨扬, 等. 流量矩阵估算的研究[J]. 软件学报, 2007, 18(11):2669-2682. ZHOU J J, YANG J H, YANG Y, et al. Research on traffic matrix estimation[J]. Journal of Software, 2007, 18(11):2669-2682. (in Chinese) [25] APPLEGATE D, COHEN E. Making intra-domain routing robust to changing and uncertain traffic demands:Understanding fundamental tradeoffs[C]//Proceedings of 2003 Conference on Applications, Technologies, Architectures, and Protocols for Computer Communications. Karlsruhe, Germany:Association for Computing Machinery, 2003:313-324. [26] 廖虎. 高效内容分发网络中服务器优化部署及路由策略[D]. 成都:电子科技大学, 2018. LIAO H. Optimizing server deployment and routing strategy in efficient content distribution network[D]. Chengdu:University of Electronic Science and Technology of China, 2018. (in Chinese) [27] 田铭. 基于流量均衡的路由优化问题研究[D]. 北京:解放军信息工程大学, 2010. TIAN M. Research on routing optimization issues based on traffic balancing[D]. Beijing:Information Engineering University, 2010. (in Chinese) [28] 林徐, 张健. BGP与OSPF动态路由震荡及其解决方法[J]. 太原师范学院学报(自然科学版), 2018, 17(4):50-55. LIN X, ZHANG J. BGP and OSPF dynamic route flapping and its solution method[J]. Journal of Taiyuan Normal University (Natural Science Edition), 2018, 17(4):50-55. (in Chinese)
2024, Vol. 64 
