Two general optimization techniques for low-latency services

GUO Yaning; XU Mingwei

doi:10.16511/j.cnki.qhdxxb.2024.26.016

Journal of Tsinghua University(Science and Technology) >

2024 , Vol. 64 >Issue 8: 1306 - 1318

DOI: https://doi.org/10.16511/j.cnki.qhdxxb.2024.26.016

COMPUTER SCIENCE AND TECHNOLOGY

Two general optimization techniques for low-latency services

GUO Yaning ,
XU Mingwei

Expand

1. Institute for Network Sciences and Cyberspace, Tsinghua University, Beijing 100084, China;
2. Zhongguancun National Laboratory, Beijing 100097, China

Received date: 2023-07-15

Online published: 2024-07-19

Fold

Abstract

[Significance] Low-latency services have become indispensable in people's work and daily life. Various low-latency services exist, including video conferences for online communication and cloud games for entertainment, which can meet different requirements of users. These services make it convenient for users to interact anywhere in real time, thereby overcoming the limitations of traditional applications. Therefore, these services have recently attracted numerous users, and their deployment scale has rapidly expanded. With the development of 5G technology, the coverage of low-latency services will spread further; these services have broad development prospects. Therefore, performance optimization of low-latency services is a hotspot in academia and industry. The most critical performance indicator for low-latency services is end-to-end latency. In addition to maintaining low latency, achieving high throughput and link usage to improve service quality and attract more users is also necessary. Therefore, performance optimization is key in the further development of low-latency services. [Progress] Various performance optimization schemes used at different positions of the transmission path were proposed by researchers. Among them, the two most widely used schemes were the low-latency congestion control algorithm (CCA) deployed on the server side and the active queue management algorithm (AQM) deployed in the network. Their design tried to avoid queuing as much as possible and to reduce end-to-end delay. The CCA and AQM constantly updated their design to solve the limitations of previous algorithms, improved their performance, and enhanced the practicality of the algorithms for large-scale deployment. Specifically, CCA improved the estimation strategy of congestion signals to make them more accurate, completed the logic of the adjustment of the sending rate and incorporated consideration for fairness into the design. AQM focused on queuing delay and minimized the amount of parameters, trying to implement a more lightweight algorithm. Although CCA and AQM shared similar goals, they were researched in parallel and independently. Being in the same control loop and both affecting the quality of low-latency services, the synergistic effect between CCA and AQM attracted considerable academic attention. Existing evaluations indicated a potential mismatch, resulting in poor performance when they were used together. To achieve superior collaboration between CCA and AQM, various optimization solutions were proposed. Among them, general CCA and AQM based on machine learning and cross-layer joint optimization were representative schemes. Although these solutions aimed to solve the mismatch problem and proposed general algorithms, they faced challenges in real-world deployment. [Conclusions and Prospects] This paper summarizes the main design ideas and performance evaluation of important low-latency CCA and AQM, sorts the performance and theoretical analysis of the combined use of the two algorithm types, analyzes the potential coordination problems between them, and further elaborates on the research work on the collaborative optimization of CCA and AQM. Through the summary, we propose that future research on low-latency services performance optimization must emphasize versatility and practical deployment and believe that cross-layer joint optimization is a practical idea to solve the existing mismatch, make CCA and AQM work well together and further improve the performance of low-latency services, which can be the focus of future research.

Key words： congestion control algorithm; active queue management; low-latency services; cross-layer joint optimization; machine learning; performance optimization

Cite this article

GUO Yaning , XU Mingwei . Two general optimization techniques for low-latency services[J]. Journal of Tsinghua University(Science and Technology), 2024 , 64(8) : 1306 -1318 . DOI: 10.16511/j.cnki.qhdxxb.2024.26.016

References

[1] LIN X S, MA Y F, ZHANG J S, et al. GSO-simulcast:Global stream orchestration in simulcast video conferencing systems[C]//Proceedings of the ACM SIGCOMM 2022 Conference. Amsterdam, the Netherlands:ACM, 2022:826-839.
[2] KARL K A, PELUCHETTE J V, AGH AKHANI N. Virtual work meetings during the COVID-19 pandemic:The good, bad, and ugly[J]. Small Group Research, 2022, 53(3):343-365.
[3] KÄMÄRÄINEN T, SIEKKINEN M, YLÄ-JÄÄSKI A, et al. A measurement study on achieving imperceptible latency in mobile cloud gaming[C]//Proceedings of the 8th ACM on Multimedia Systems Conference. Taibei, China:ACM, 2017:88-99.
[4] BRAKMO L S, O'MALLEY S W, PETERSON L L. TCP Vegas:New techniques for congestion detection and avoidance[J]. ACM SIGCOMM Computer Communication Review, 1994, 24(4):24-35.
[5] HENGARTNER U, BOLLIGER J, GROSS T. TCP Vegas revisited[C]//Proceedings IEEE INFOCOM 2000. Tel Aviv, Israel:IEEE, 2000:1546-1555.
[6] JIN C, WEI D, LOW S H, et al. FAST TCP:From theory to experiments[J]. IEEE Network, 2005, 19(1):4-11.
[7] ARAMANI A, KOKKU R, DAHLIN M. TCP Nice:A mechanism for background transfers[J]. ACM SIGOPS Operating Systems Review, 2002, 36(SI):329-343.
[8] KUEHLEWIND M, HAZEL G, SHALUNOV S, et al. Low extra delay background transport (ledbat)[R/OL].(2012-12-19)[2023-7-15]. https://www.rfc-editor.org/rfc/rfc6817.
[9] MENG T, SCHIFF N R, GODFREY P B, et al. PCC proteus:Scavenger transport and beyond[C]//Proceedings of the Annual Conference of the ACM Special Interest Group on Data Communication on the Applications, Technologies, Architectures, and Protocols for Computer Communication. New York, USA:ACM, 2020:615-631.
[10] ZAKI Y, PÖTSCH T, CHEN J, et al. Adaptive congestion control for unpredictable cellular networks[C]//Proceedings of the 2015 ACM Conference on Special Interest Group on Data Communication. London, UK:ACM, 2015:509-522.
[11] ABBASLOO S, XU Y, CHAO H J. C2TCP:A flexible cellular TCP to meet stringent delay requirements[J]. IEEE Journal on Selected Areas in Communications, 2019, 37(4):918-932.
[12] ARUN V, BALAKRISHNAN H. Copa:Practical delay-based congestion control for the internet[C]//15th USENIX Symposium on Networked Systems Design and Implementation (NSDI 18). Renton, USA:USENIX Association, 2018:329-342.
[13] LIU S, BAŞAR T, SRIKANT R. TCP-Illinois:A loss and delay-based congestion control algorithm for high-speed networks[C]//Proceedings of the 1st International Conference on Performance Evaluation Methodolgies and Tools. Pisa, Italy:ACM, 2006:55-es.
[14] TAN K, SONG J, ZHANG Q, et al. A compound TCP approach for high-speed and long distance networks[C]//Proceedings of the 25th IEEE International Conference on Computer Communications. Barcelona, Spain:IEEE, 2006:1-12.
[15] CARLUCCI G, DE CICCO L, HOLMER S, et al. Congestion control for web real-time communication[J]. IEEE/ACM Transactions on Networking, 2017, 25(5):2629-2642.
[16] ZHU X Q, PAN R. NADA:A unified congestion control scheme for low-latency interactive video[C]//2013 20th International Packet Video Workshop. San Jose, USA:IEEE, 2013:1-8.
[17] JOHANSSON I. Self-clocked rate adaptation for conversational video in LTE[C]//Proceedings of the 2014 ACM SIGCOMM Workshop on Capacity Sharing Workshop. Chicago, USA:ACM, 2014:51-56.
[18] CARDWELL N, CHENG Y C, GUNN C S, et al. BBR:Congestion-based congestion control:Measuring bottleneck bandwidth and round-trip propagation time[J]. Queue, 2016, 14(5):20-53.
[19] HOCK M, BLESS R, ZITTERBART M. Experimental evaluation of BBR congestion control[C]//2017 IEEE 25th International Conference on Network Protocols (ICNP). Toronto, Canada:IEEE, 2017:1-10.
[20] WINSTEIN K, SIVARAMAN A, BALAKRISHNAN H. Stochastic forecasts achieve high throughput and low delay over cellular networks[C]//10th USENIX Symposium on Networked Systems Design and Implementation. Lombard, USA:USENIX Association, 2013:459-472.
[21] POLESE M, CHIARIOTTI F, BONETTO E, et al. A survey on recent advances in transport layer protocols[J]. IEEE Communications Surveys&Tutorials, 2019, 21(4):3584-3608.
[22] AL-SAADI R, ARMITAGE G, BUT J, et al. A survey of delay-based and hybrid TCP congestion control algorithms[J]. IEEE Communications Surveys&Tutorials, 2019, 21(4):3609-3638.
[23] HAILE H, GRINNEMO K J, FERLIN S, et al. End-to-end congestion control approaches for high throughput and low delay in 4G/5G cellular networks[J]. Computer Networks, 2021, 186:107692.
[24] LORINCZ J, KLARIN Z, OŻEGOVIĆ J. A comprehensive overview of TCP congestion control in 5G networks:Research challenges and future perspectives[J]. Sensors, 2021, 21(13):4510.
[25] ABBASLOO S, YEN C Y, CHAO H J. Classic meets modern:A pragmatic learning-based congestion control for the internet[C]//Proceedings of the Annual Conference of the ACM Special Interest Group on Data Communication on the Applications, Technologies, Architectures, and Protocols for Computer Communication. New York, USA:ACM, 2020:632-647.
[26] GOYAL P, AGARWAL A, NETRAVALI R, et al. ABC:A simple explicit congestion controller for wireless networks[C]//17th USENIX Symposium on Networked Systems Design and Implementation (NSDI 20). Santa Clara, USA:USENIX Association, 2020:353-372.
[27] DU Z, ZHENG J, YU H, et al. A unified congestion control framework for diverse application preferences and network conditions[C]//Proceedings of the 17th International Conference on emerging Networking Experiments and Technologies. New York, USA:ACM, 2021:282-296.
[28] DONG M, MENG T, ZARCHY D, et al. PCC vivace:Online-learning congestion control[C]//15th USENIX Symposium on Networked Systems Design and Implementation (NSDI 18). Renton, USA:USENIX Assosiation, 2018:343-356.
[29] FLOYD S, JACOBSON V. Random early detection gateways for congestion avoidance[J]. IEEE/ACM Transactions on Networking, 1993, 1(4):397-413.
[30] FLOYD S, GUMMADI R, SHENKER S J. Adaptive RED:An algorithm for increasing the robustness of RED's active queue management[R]. Berkeley, USA:International Computer Science Institute, 2001.
[31] WYDROWSKI B, ZUKERMAN M. GREEN:An active queue management algorithm for a self managed internet[C]//2002 IEEE International Conference on Commu-nications. Conference Proceedings. New York, USA:IEEE, 2002:2368-2372.
[32] NICHOLS K, JACOBSON V. Controlling queue delay[J]. Communications of the ACM, 2012, 55(7):42-50.
[33] PAN R, NATARAJAN P, PIGLIONE C, et al. PIE:A lightweight control scheme to address the bufferbloat problem[C]//2013 IEEE 14th International Conference on High Performance Switching and Routing (HPSR). Taipei, China:IEEE, 2013:148-155.
[34] FENG W C, SHIN K G, KANDLUR D D, et al. The BLUE active queue management algorithms[J]. IEEE/ACM Transactions on Networking, 2002, 10(4):513-528.
[35] PALMEI J, GUPTA S, IMPUTATO P, et al. Design and evaluation of COBALT queue discipline[C]//2019 IEEE International Symposium on Local and Metropolitan Area Networks (LANMAN). Paris, France:IEEE, 2019:1-6.
[36] KHADEMI N, ROS D, WELZL M. The new AQM kids on the block:An experimental evaluation of CoDel and PIE[C]//2014 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS). Toronto, Canada:IEEE, 2014:85-90.
[37] MUHAMMAD S, CHAUDHERY T J, NOH Y. Study on performance of AQM schemes over TCP variants in different network environments[J]. IET Communications, 2021, 15(1):93-111.
[38] GRAZIA C A, PATRICIELLO N, KLAPEZ M, et al. A cross-comparison between TCP and AQM algorithms:Which is the best couple for congestion control?[C]//2017 14th International Conference on Telecommunications (ConTEL). Zagreb, Croatia:IEEE, 2017:75-82.
[39] PIOTROWSKA A. On cross-layer interactions for congestion control in the internet[J]. Applied Sciences, 2021, 11(17):7808.
[40] HA S, RHEE I, XU L S. CUBIC:A new TCP-friendly high-speed TCP variant[J]. ACM SIGOPS Operating Systems Review, 2008, 42(5):64-74.
[41] CARLUCCI G, DE CICCO L, MASCOLO S. Controlling queuing delays for real-time communication:The interplay of E2E and AQM algorithms[J]. ACM SIGCOMM Computer Communication Review, 2016, 46(3):1.
[42] ZHANG H, ZHU H T, XIA Y, et al. Performance analysis of BBR congestion control protocol based on NS3[C]//2019 Seventh International Conference on Advanced Cloud and Big Data (CBD). Suzhou, China:IEEE, 2019:363-368.
[43] ABBASLOO S, CHAO H J. Bounding queue delay in cellular networks to support ultra-low latency applications[J/OL]. arXiv.(2019-08-02)[2023-07-15]. https://arxiv.org/pdf/1908.00953.pdf.
[44] DONG M, LI Q X, ZARCHY D, et al. PCC:Re-architecting congestion control for consistent high performance[C]//12th USENIX Symposium on Networked Systems Design and Implementation (NSDI 15). Oakland, USA:USENIX Association, 2015:395-408.
[45] BACHL M, FABINI J, ZSEBY T. LFQ:Online learning of per-flow queuing policies using deep reinforcement learning[C]//2020 IEEE 45th Conference on Local Computer Networks (LCN). Sydney, Australia:IEEE, 2020:417-420.
[46] NÁDAS S, GOMBOS G, HUDOBA P, et al. Towards a congestion control-independent core-stateless AQM[C]//Proceedings of the Applied Networking Research Workshop. Montreal, Canada:ACM, 2018:84-90.
[47] MA H H, XU D, DAI Y Y, et al. An intelligent scheme for congestion control:When active queue management meets deep reinforcement learning[J]. Computer Networks, 2021, 200:108515.
[48] BAI J N, ZHANG T H, XIE G M. MACC:Cross-layer multi-agent congestion control with deep reinforcement learning[J/OL]. arXiv.(2022-06-04)[2023-07-15]. https://arxiv.org/pdf/2206.01972.pdf.
[49] KATABI D, HANDLEY M, ROHRS C. Congestion control for high bandwidth-delay product networks[J]. ACM SIGCOMM Computer Communication Review, 2002, 32(4):89-102.
[50] TAI C H, ZHU J, DUKKIPATI N. Making large scale deployment of RCP practical for real networks[C]//IEEE INFOCOM 2008-The 27th Conference on Computer Communications. Phoenix, USA:IEEE, 2008:2180-2188.
[51] XIA Y, SUBRAMANIAN L, STOICA I, et al. One more bit is enough[C]//Proceedings of the 2005 Conference on Applications, Technologies, Architectures, and Protocols for Computer Communications. Philadelphia, USA:ACM, 2005:37-48.

Options

Outlines

模态框（Modal）标题

Abstract

Cite this article

References

Visited