BurstBalancer: Do Less, Better Balance for Large-scale Data Center Traffic

被引:9
作者
Liu, Zirui [1 ,2 ]
Zhao, Yikai [1 ,2 ]
Fan, Zhuochen [1 ,2 ]
Yang, Tong [1 ,2 ,3 ]
Li, Xiaodong [1 ,2 ]
Zhang, Ruwen [1 ,2 ]
Yang, Kaicheng [1 ,2 ]
Zhong, Zheng [1 ,2 ]
Huang, Yi [4 ]
Liu, Cong [4 ]
Hu, Jing [4 ]
Xie, Gaogang [5 ]
Cui, Bin [1 ,2 ]
机构
[1] Peking Univ, Sch Comp Sci, Beijing, Peoples R China
[2] Peking Univ, Natl Engn Lab Big Data Anal Technol & Applicat, Beijing, Peoples R China
[3] Peng Cheng Lab, Shenzhen, Peoples R China
[4] Huawei Technol, Shenzhen, Peoples R China
[5] Chinese Acad Sci, CNIC, Beijing, Peoples R China
来源
2022 IEEE 30TH INTERNATIONAL CONFERENCE ON NETWORK PROTOCOLS (ICNP 2022) | 2022年
基金
中国国家自然科学基金;
关键词
FREQUENT; TIME;
D O I
10.1109/ICNP55882.2022.9940372
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
Layer-3 load balancing is a key topic in the networking field. It is well acknowledged that flowlet is the most promising solution because of its good trade-off between load balance and packet reordering. However, we find its one significant limitation: it makes the forwarding paths of flows unpredictable. To address this limitation, this paper presents BurstBalancer, a simple yet efficient load balancing system with a sketch, named BalanceSketch. Our design philosophy is doing less changes to keep the forwarding path of most flows fixed, which guides the design of BalanceSketch and balance operations. We have fully implemented BurstBalancer in a small-scale testbed built with Tofino switches, and conducted large-scale NS-2 simulations. Our results show that BurstBalancer achieves 5%similar to 35% smaller FCT than LetFlow in symmetric topology and up to 30 x smaller FCT in asymmetric topology, while 58 x fewer flows suffer from path changing. All related codes are open-sourced at Github(2).
引用
收藏
页数:13
相关论文
共 93 条
[1]  
Al-Fares Mohammad, 2010, Nsdi, V10, P89
[2]  
Alizadeh M, 2014, ACM SIGCOMM COMP COM, V44, P503, DOI [10.1145/2619239.2626316, 10.1145/2740070.2626316]
[3]   pFabric: Minimal Near-Optimal Datacenter Transport [J].
Alizadeh, Mohammad ;
Yang, Shuang ;
Sharif, Milad ;
Katti, Sachin ;
McKeown, Nick ;
Prabhakar, Balaji ;
Shenker, Scott .
ACM SIGCOMM COMPUTER COMMUNICATION REVIEW, 2013, 43 (04) :435-446
[4]   Data Center TCP (DCTCP) [J].
Alizadeh, Mohammad ;
Greenberg, Albert ;
Maltz, David A. ;
Padhye, Jitendra ;
Patel, Parveen ;
Prabhakar, Balaji ;
Sengupta, Sudipta ;
Sridharan, Murari .
ACM SIGCOMM COMPUTER COMMUNICATION REVIEW, 2010, 40 (04) :63-74
[5]  
Alizadeh M, 2012, PROCEEDINGS OF THE 11TH ACM WORKSHOP ON HOT TOPICS IN NETWORKS (HOTNETS-XI), P133
[6]  
Alvarez-Horcajo J, 2017, IEEE INT CONF CL NET, P65
[7]  
[Anonymous], P4-16 Language Specification
[8]  
Arzani B, 2018, PROCEEDINGS OF THE 15TH USENIX SYMPOSIUM ON NETWORKED SYSTEMS DESIGN AND IMPLEMENTATION (NSDI'18), P419
[9]   Taking the Blame Game out of Data Centers Operations with NetPoirot [J].
Arzani, Behnaz ;
Ciraci, Selim ;
Loo, Boon Thau ;
Schuster, Assaf ;
Outhred, Geoff .
PROCEEDINGS OF THE 2016 ACM CONFERENCE ON SPECIAL INTEREST GROUP ON DATA COMMUNICATION (SIGCOMM '16), 2016, :440-453
[10]  
Bai W, 2016, 13TH USENIX SYMPOSIUM ON NETWORKED SYSTEMS DESIGN AND IMPLEMENTATION (NSDI '16), P537