隨著越來越多的網路服務遷移到資料中心，導致資料中心網路中交換機之間傳輸的流量快速增長。資料中心網路(DCN)通常設計為在任兩台host之間有多條路徑，而在DCN中通常把流量分為兩種，分別為大象流與老鼠流。SDN（軟體定義網路）為優化資源管理提供了有效的技術方法。然而，最近許多研究只專注於提升網路整體吞吐量並未考慮老鼠流的完成時間(FCT)，因此小鼠流可能會與大象流爭奪頻寬而導致FCT增加。在本文中，我們提出了一種流量分類方法，將大象流與小鼠流彼此區隔，並提出大象流調度方法，網路中若沒有小鼠流時，大象流可以暫時調度到小鼠流的路徑。經過實驗證明，與等價多路徑 (ECMP) 和Hedera 相比，我們的方法有效的改善小鼠流FCT，並保持了大象流的吞吐量。

As more and more network services migrate to the data center, resulting in a rapid increase in the traffic transmitted between switches in the data center network(DCN), the data center network is usually designed with multiple paths between any two hosts. In DCN, the traffic is classified into two types, namely elephant flows and mice flows. Software Defined Networking(SDN) provides an effective technical approach for optimizing resource management. However, many recent studies only focus on improving the overall throughput of the network and do not consider the completion time of the mice flows(FCT), so the mice flows may compete with the elephant flow for bandwidth and lead to an increase in FCT. In this paper, we propose a flow classification method to separate the elephant flows from the mice flows, and propose an elephant flow scheduling method. If there are no mice flows in the network, the elephant flows can be temporarily scheduled to paths that belong to mice flows. Experiments show that our method effectively improves FCT of mice flows and maintains the throughput of elephant flow compared to equivalent ECMP and Sieve.

[1] C. E. Leiserson, “Fat-trees: Universal networks for hardware-efficient supercomputing”, in IEEE Transactions on Computers, vol. C-34, no. 10, pp. 892-901, Oct. 1985.
[2] Ankit Singla, Chi-Yao Hong, Lucian Popa, P. Brighten Godfrey, “Jellyfish: Networking data centers randomly”, 9th USENIX Symposium on Networked Systems Design and Implementation. pp.1-14, 2012.
[3] Albert Greenberg, James R. Hamilton, Navendu Jain, Srikanth Kandula, Changhoon Kim, Parantap Lahiri, David A. Maltz, Parveen Patel, and Sudipta Sengupta, “VL2: a scalable and flexible data center network,” SIGCOMM Comput. Commun. Rev. 39, pp. 51–62, 2009.
[4] Chuanxiong Guo, Guohan Lu, Dan Li, Haitao Wu, Xuan Zhang, Yunfeng Shi, Chen Tian, “BCube: a high performance, server-centric network architecture for modular data centers”, ACM SIGCOMM Computer Communication Review 39.4, 2009, pp.63-74.
[5] Chuanxiong Guo, Haitao Wu, Kun Tan, Lei Shi, Yongguang Zhanget, Songwu Lu, “Dcell: a scalable and fault-tolerant network structure for data centers”, ACM SIGCOMM Computer Communication Review. Vol. 38. No.4. ACM, 2008, pp. 75-86.
[6] Niranjan Mysore, Radhika, Andreas Pamboris, Nathan Farrington, Nelson Huang, Pardis Miri, Sivasankar Radhakrishnan, “PortLand: a scalable fault-tolerant layer 2 data center network fabric”, ACM SIGCOMM Computer Communication Review. Vol. 39. No. 4. ACM, 2009, pp. 39-50.
[7] Farrington, Nathan, George Porter, Sivasankar Radhakrishnan, Hamid Hajabdolali Bazzaz, “Helios: a hybrid electrical/optical switch architecture for modular data centers”, ACM SIGCOMM Computer Communication Review. Vol. 41. No. 4. ACM, pp.339-350, 2011.
[8] Hailong Zhang, Xiao Guo, Jinyao Yan, Bo Liu and Qianjun Shuai, “SDN-based ECMP algorithm for data center networks”, 2014 IEEE Computers, Communications and IT Applications Conference, 2014, pp. 13-18.
[9] B. Huang and S. Dong, “An Enhanced Scheduling Framework for Elephant Flows in SDN-Based Data Center Networks”, 2020 IEEE Symposium on Computers and Communications (ISCC), 2020, pp. 1-7.
[10] Maiass Zaher, Aymen Hasan Alawadi, S ́andor Moln ́ar, “Sieve: A flow scheduling framework in SDN based data center networks”, Computer Communications, Volume 171, Pages 99-111, 2021.
[11] T. Benson, A. Akella, and D. A. Maltz, “Network Traffic Characteristics of Data Centers in the Wild”, Proceedings of the 10th ACM SIGCOMM Conference on Internet Measurement Conference, Melbourne. Australia, 2010, pp. 267-280.
[12] C. Hopps. “Analysis of an Equal-Cost Multi-Path Algorithm”, RFC Editor, USA, pp1-8, 2000.
[13] J. Cao , R. Xia , P. Yang , C. Guo , G. Lu , L. Yuan , Y. Zheng , H. Wu , Y. Xiong , D. Maltz, “Per-packet load-balanced, low-latency routing for clos-based data center networks”, in: Proceedings of the Ninth ACM Conference on Emerging Networking Experiments and Technologies, ACM, 2013, pp. 49–60.
[14] D. Zats , T. Das , P. Mohan , D. Borthakur , R. Katz , “Detail: reducing the flow completion time tail in datacenter networks”, Comput. Commun. ACM SIGCOMM, pp.139-150, 2012.
[15] M. Alizadeh , A. Greenberg , D.A. Maltz , J. Padhye , P. Patel , B. Prabhakar , S. Sengupta , M. Sridharan , “Data center tcp (dctcp)”, Comput. Commun. ACM SIGCOMM pp. 63–74, 2011.
[16] A. Dixit , P. Prakash , Y.C. Hu , R.R. Kompella , “On the impact of packet spraying in data center networks”, in: INFOCOM, 2013 Proceedings IEEE, IEEE, 2013, pp. 2130–2138 .
[17] D. Kreutz, F. Ramos, P. Ver ́ıssimo, C. Esteve Rothenberg, S. Azodolmolky, and S. Uhlig, “Software-Defined Networking: A Comprehensive Survey”, Proceedings of the IEEE, vol. 103, pp. 14-76, 2015.
[18] N. Mckeown, T. Anderson, and H. Balakrishnan, “OpenFlow: Enabling innovation in campus networks”, ACM SIGCOMM Computer Communication Review, vol. 38, pp. 69-74, 2008.
[19] M. Al-Fares, S. Radhakrishnan, B. Raghavan, N. Huang, and A.Vahdat, “Hedera: Dynamic Flow Scheduling for Data Center Networks”, Proceedings of the 7th USENIX Symposium on Networked Systems Design and Implementation, San Jose. CA. USA, 2010, pp. 281-296.
[20] S. TAO, Y. LIU, Y. Wang, R. Ma, “Ashman: A Bandwidth Fragmentation-Based Dynamic Flow Scheduling for Data Center Networks”, Computer Journal, vol. 60, pp. 1498-1509, 2017.
[21] L. Liu, Y. Jiang, G. Shen, “An SDN-based Hybrid Strategy for Load Balancing in Data Center Networks”, 2019 IEEE Symposium on Computers and Communications (ISCC), 2019, pp. 1-6
[22] Y. Zhang, L. Cui, and Y. Zhang, “A stable matching based elephant flow scheduling algorithm in data center networks”, Computer Networks, vol. 120, pp. 186-197, 2017.
[23] M. Noormohammadpour, C.S. Raghavendra, “Datacenter traffic control: Understanding techniques and tradeoffs”, IEEE Commun. Surv. Tutor. 20 (2) (2017), pp. 1492–1525, 2017
[24] C. Wang, G. Zhang, H. Chen and H. Xu, “An ACO-based elephant and mice flow scheduling system in SDN”, 2017 IEEE 2nd International Conference on Big Data Analysis (ICBDA), pp. 859-863, 2017.
[25] W. Wang, Y. Sun, K. Zheng, M. A. Kaafar, D. Li and Z. Li, “Freeway: Adaptively Isolating the Elephant and Mice Flows on Different Transmission Paths”, 2014 IEEE 22nd International Conference on Network Protocols, pp. 362-367, 2014.
[26] OpenFlow protocol API. https://ryu.readthedocs.io/en/latest/ofproto ref.html
[27] Alshahrani, Reem Abdullah. “Delay modeling in data center networks: A taxonomy and performance analysis”, Diss. Kent State University, pp. 63-76 2013, http://rave.ohiolink.edu/etdc/view?acc num=kent1375712266
[28] iperf. https://iperf.fr/iperf-doc.php.
[29] Python SimpleHTTPServer. https://pypi.org/project/simple-http-server/.
[30] S. Alexander, D. Zuikov, D. Zimarina, V. Pashkov and R. Smeliansky, “Advanced study of SDN/OpenFlow controllers”, Proc. 9th Central Eastern Eur. Softw. Eng. Conf. Russia, pp. 1-6, 2013.
[31] N. Handigol, B. Heller, V. Jeyakumar, B. Lantz and N. McKeown, “Reproducible network experiments using container-based emulation”, Proc. ACM 8th Int. Conf. Emerging Netw. Exp. Technol. pp. 253-264, 2012. 42