Taming the Aggressiveness of Heterogeneous TCP Traffic in Data Center Networks
Authors: 
Tao Zhang, Ran Huang, Jiawei Huang, Shaojun Zou, Chang Ruan, Kai Chen, Jianxin Wang, Geyong MinAuthors Info & Claims
Tao Zhang, Ran Huang, Jiawei Huang, Shaojun Zou, Chang Ruan, Kai Chen, Jianxin Wang, Geyong MinAuthors Info & ClaimsPages 2253 - 2268
Abstract
To achieve low latency and high link utilization, ECN-based transport protocols (i.e., DCTCP) are widely deployed in data center networks (DCN). In multi-tenant environment, however, the newly introduced ECN-enabled TCP greatly impairs the performance of applications with out-dated and misconfigured TCP stacks. The reason is that the ECN-enabled switch fails to treat the mixed TCP traffic fairly, resulting in the distinguished performance gap between the ECN-enabled and ECN-disabled TCPs. This paper proposes DDT (Dual Dynamic Thresholds), an active queue management algorithm (AQM) to achieve the flow-level fairness for coexisting heterogeneous TCP traffic. DDT monitors the switch queue in real time, and dynamically tunes the distance between ECN-marking and packet-dropping thresholds to mitigate the aggressiveness difference between the ECN-enabled and ECN-disabled TCPs. The results of real implementations and large-scaled simulations show that DDT elegantly fills the aggressiveness gap of heterogeneous TCP traffic without disturbing their own control loops, while only introducing acceptable deployment overhead at switch.
References
[1]
T. Zhanget al., “DDT: Mitigating the competitiveness difference of data center TCPs,” in Proc. 3rd Asia–Pacific Workshop Netw., Aug. 2019, pp. 8–14.
[2]
M. Alizadehet al., “Data center TCP (DCTCP),” in Proc. ACM SIGCOMM Conf., Aug. 2010, pp. 63–74.
[3]
K. Heet al., “AC/DC TCP: Virtual congestion control enforcement for datacenter networks,” in Proc. ACM SIGCOMM Conf., Aug. 2016, pp. 244–257.
[4]
B. Cronkite-Ratcliffet al., “Virtualized congestion control,” in Proc. ACM SIGCOMM Conf., Aug. 2016, pp. 230–243.
[5]
G. Judd, “Attaining the promise and avoiding the pitfalls of TCP in the datacenter,” in Proc. NSDI, 2015, pp. 145–157.
[6]
M. Alizadeh, A. Kabbani, T. Edsall, B. Prabhakar, A. Vahdat, and M. Yasuda, “Less is more: Trading a little bandwidth for ultra-low latency in the data center,” in Proc. NSDI, 2012, pp. 253–266.
[7]
A. Muniret al., “Minimizing flow completion times in data centers,” in Proc. IEEE INFOCOM, Apr. 2013, pp. 2157–2165.
[8]
B. Vamanan, J. Hasan, and T. N. Vijaykumar, “Deadline-aware datacenter TCP (D2TCP),” ACM SIGCOMM Comput. Commun. Rev., vol. 42, no. 4, pp. 115–126, Sep. 2012.
[9]
H. Wu, J. Ju, G. Lu, C. Guo, Y. Xiong, and Y. Zhang, “Tuning ECN for data center networks,” in Proc. 8th Int. Conf. Emerg. Netw. Experiments Technol., Dec. 2012, pp. 25–36.
[10]
R. Mittalet al., “TIMELY: RTT-based congestion control for the datacenter,” in Proc. SIGCOMM, 2015, pp. 537–550.
[11]
C. Lee, C. Park, K. Jang, S. Moon, and D. Han, “Accurate latency-based congestion feedback for datacenters,” in Proc. ATC, 2015, pp. 403–415.
[12]
G. Kumaret al., “Swift: Delay is simple and effective for congestion control in the datacenter,” in Proc. Annu. Conf. ACM Special Interest Group Data Commun. Appl., Technol., Architectures, Protocols Comput. Commun., Jul. 2020, pp. 514–528.
[13]
Y. Liet al., “HPCC: High precision congestion control,” in Proc. ACM Special Interest Group Data Commun., Aug. 2019, pp. 44–58.
[14]
V. Addanki, O. Michel, and S. Schmid, “PowerTCP: Pushing the performance limits of datacenter networks,” in Proc. NSDI, 2022, pp. 1–16.
[15]
(2014). The Network Simulator-NS2. [Online]. Available: https://www.isi.edu/nsnam/ns
[16]
D.-M. Chiu and R. Jain, “Analysis of the increase and decrease algorithms for congestion avoidance in computer networks,” Comput. Netw. ISDN Syst., vol. 17, no. 1, pp. 1–14, Jun. 1989.
[17]
M. Kato, L. Eggert, A. Zimmermann, R. Meter, and H. Tokuda, “Extensions to FreeBSD datacenter TCP for incremental deployment support,” BSDCan, Ottawa, ON, Canada, Tech. Rep. 559, 2015.
[18]
M. Kühlewind, D. P. Wagner, J. M. R. Espinosa, and B. Briscoe, “Using data center TCP (DCTCP) in the internet,” in Proc. IEEE Globecom Workshops (GC Wkshps), Dec. 2014, pp. 583–588.
[19]
P. Cheng, F. Ren, R. Shu, and C. Lin, “Catch the whole lot in an action: Rapid precise packet loss notification in data centers,” in Proc. NSDI, 2014, pp. 17–28.
[20]
W. Bai, K. Chen, H. Wu, W. Lan, and Y. Zhao, “PAC: Taming TCP incast congestion using proactive ACK control,” in Proc. IEEE 22nd Int. Conf. Netw. Protocols, Oct. 2014, pp. 385–396.
[21]
T. Benson, A. Akella, and D. A. Maltz, “Network traffic characteristics of data centers in the wild,” in Proc. 10th ACM SIGCOMM Conf. Internet Meas., Nov. 2010, pp. 267–280.
[22]
I. Cho, K. Jang, and D. Han, “Credit-scheduled delay-bounded congestion control for datacenters,” in Proc. Conf. ACM Special Interest Group Data Commun., Aug. 2017, pp. 239–252.
[23]
M. Alizadeh, A. Javanmard, and B. Prabhakar, “Analysis of DCTCP: Stability, convergence, and fairness,” ACM SIGMETRICS Perform. Eval. Rev., vol. 39, no. 1, pp. 73–84, 2011.
[24]
P. Goyal, A. Narayan, F. Cangialosi, S. Narayana, M. Alizadeh, and H. Balakrishnan, “Elasticity detection: A building block for internet congestion control,” in Proc. ACM SIGCOMM Conf., Aug. 2022, pp. 158–176.
[25]
Z. Yuet al., “Programmable packet scheduling with a single queue,” in Proc. ACM SIGCOMM Conf., Aug. 2021, pp. 179–193.
[26]
M. Wurtzler, “Analysis and simulation of weighted random early detection (WRED) queues,” 2002. [Online]. Available: https://www.ittc.ku.edu/research/thesis/documents/mark_wurtzler_thesis.pdf
[27]
V. Olteanuet al., “An edge-queued datagram service for all datacenter traffic,” in Proc. NSDI, 2022, pp. 761–777.
[28]
W. Bai, K. Chen, L. Chen, C. Kim, and H. Wu, “Enabling ECN over generic packet scheduling,” in Proc. 12th Int. Conf. Emerg. Netw. Experiments Technol., Dec. 2016, pp. 191–204.
[29]
W. Bai, L. Chen, K. Chen, and H. Wu, “Enabling ECN in multi-service multi-queue data centers,” in Proc. NSDI, 2016, pp. 537–549.
[30]
S. Liu, A. Ghalayini, M. Alizadeh, B. Prabhakar, M. Rosenblum, and A. Sivaraman, “Breaking the transience-equilibrium nexus: A new approach to datacenter packet transport,” in Proc. NSDI, 2021, pp. 47–63.
[31]
W. Bai, L. Chen, K. Chen, D. Han, C. Tian, and H. Wang, “Information-agnostic flow scheduling for commodity data centers,” in Proc. NSDI, 2015, pp. 455–468.
[32]
N. K. Sharma, M. Liu, K. Atreya, and A. Krishnamurthy, “Approximating fair queueing on reconfigurable switches,” in Proc. NSDI, 2018, pp. 1–16.
[33]
A. K. Choudhury and E. L. Hahne, “Dynamic queue length thresholds in a shared memory ATM switch,” in Proc. IEEE INFOCOM, Mar. 1996, pp. 679–687.
[34]
Intel Tofino P4-Programmable Ethernet Switch Asic That Delivers Better Performance at Lower Power. Accessed: 2023. [Online]. Available: https://www.intel.com/content/www/us/en/products/network-io/ethernet/programmable-ethernet-switch.html?wapkw=programmable-ethernet-switch
[35]
S. Yan, X. Wang, X. Zheng, Y. Xia, D. Liu, and W. Deng, “ACC: Automatic ECN tuning for high-speed datacenter networks,” in Proc. ACM SIGCOMM Conf., Aug. 2021, pp. 384–397.
[36]
L. Yu, J. Sonchack, and V. Liu, “Cebinae: Scalable in-network fairness augmentation,” in Proc. ACM SIGCOMM Conf., Aug. 2022, pp. 219–232.
[37]
N. K. Sharmaet al., “Programmable calendar queues for high-speed packet scheduling,” in Proc. NSDI, 2020, pp. 685–699.
[38]
Y. Lei, L. Yu, V. Liu, and M. Xu, “PrintQueue: Performance diagnosis via queue measurement in the data plane,” in Proc. ACM SIGCOMM Conf., Aug. 2022, pp. 516–529.
[39]
A. Tirumala and L. Cottrell, “IPERF quick mode,” 2002. [Online]. Available: https://www-iepm.slac.stanford.edu/bw/iperf/
[40]
J. Huang, Y. Huang, J. Wang, and T. He, “Adjusting packet size to mitigate TCP incast in data center networks with COTS switches,” IEEE Trans. Cloud Comput., vol. 8, no. 3, pp. 749–763, Jul. 2020.
[41]
S. Zou, J. Huang, J. Wang, and T. He, “Flow-aware adaptive pacing to mitigate TCP incast in data center networks,” IEEE/ACM Trans. Netw., vol. 29, no. 1, pp. 134–147, Feb. 2021.
[42]
T. Zhanget al., “Rethinking fast and friendly transport in data center networks,” IEEE/ACM Trans. Netw., vol. 28, no. 5, pp. 2364–2377, Oct. 2020.
[43]
D. Shan, F. Ren, P. Cheng, R. Shu, and C. Guo, “Observing and mitigating micro-burst traffic in data center networks,” IEEE/ACM Trans. Netw., vol. 28, no. 1, pp. 98–111, Feb. 2020.
[44]
D. Shan and F. Ren, “ECN marking with micro-burst traffic: Problem, analysis, and improvement,” IEEE/ACM Trans. Netw., vol. 26, no. 4, pp. 1533–1546, Aug. 2018.
[45]
P. X. Gao, A. Narayan, G. Kumar, R. Agarwal, S. Ratnasamy, and S. Shenker, “PHost: Distributed near-optimal datacenter transport over commodity network fabric,” in Proc. 11th ACM Conf. Emerg. Netw. Experiments Technol., Dec. 2015, pp. 1–12.
[46]
M. Handleyet al., “Re-architecting datacenter networks and stacks for low latency and high performance,” in Proc. Conf. ACM Special Interest Group Data Commun., Aug. 2017, pp. 29–42.
[47]
B. Montazeri, Y. Li, M. Alizadeh, and J. Ousterhout, “Homa: A receiver-driven low-latency transport protocol using network priorities,” in Proc. Conf. ACM Special Interest Group Data Commun., Aug. 2018, pp. 221–235.
[48]
S. Huet al., “Aeolus: A building block for proactive transport in datacenters,” in Proc. Annu. Conf. ACM Special Interest Group Data Commun. Appl., Technol., Architectures, Protocols Comput. Commun., Jul. 2020, pp. 422–434.
[49]
A. Saeedet al., “Annulus: A dual congestion control loop for datacenter and WAN traffic aggregates,” in Proc. Annu. Conf. ACM Special Interest Group Data Commun. Appl., Technol., Architectures, Protocols Comput. Commun., Jul. 2020, pp. 735–749.
[50]
V. Addanki, M. Apostolaki, M. Ghobadi, S. Schmid, and L. Vanbever, “ABM: Active buffer management in datacenters,” in Proc. ACM SIGCOMM Conf., Aug. 2022, pp. 36–52.
[51]
V. Arun, M. Alizadeh, and H. Balakrishnan, “Starvation in end-to-end congestion control,” in Proc. ACM SIGCOMM Conf., Aug. 2022, pp. 177–192.
[52]
S. S. Chen, W. Wang, C. Canel, S. Seshan, A. C. Snoeren, and P. Steenkiste, “Time-division TCP for reconfigurable data center networks,” in Proc. ACM SIGCOMM Conf., Aug. 2022, pp. 19–35.
Recommendations
Analysis for TCP in data center networks
The unfairness caused by bandwidth sharing via TCP in data center networks is called TCP Outcast problem. Some researchers show that the throughput of a flow with small Round Trip Time (RTT) is less than that with large RTT which is completely contrary ...
Randomizing TCP payload size for TCP fairness in data center networks
As many-to-one traffic patterns prevail in data center networks, TCP flows often suffer from severe unfairness in sharing bottleneck bandwidth, which is known as the TCP outcast problem. The cause of the TCP outcast problem is the bursty packet losses ...
TCP incast solutions in data center networks: A classification and survey
AbstractIn recent years, Data Centers Networks (DCNs) have been deployed to serve as the backbone to support the extensive variety of services offered through the Internet like social networking, web hosting, and e-commerce. The Transmission ...
Comments
Information & Contributors
Information
Published In
1063-6692 © 2023 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
Publisher
IEEE Press
Publication History
Published: 28 December 2023
Published in TON Volume 32, Issue 3
Qualifiers
- Research-article
Contributors
Other Metrics
Bibliometrics & Citations
Bibliometrics
Article Metrics
- 0Total Citations
- 20Total Downloads
- Downloads (Last 12 months)20
- Downloads (Last 6 weeks)4
Reflects downloads up to 27 Dec 2024
Other Metrics
Citations
View Options
Login options
Check if you have access through your login credentials or your institution to get full access on this article.
Sign in