Proposal of the Data Center-centric Flow …biblio.yamanaka.ics.keio.ac.jp/file/Imakiire-PN2017-25.pdfin DCN. Keywords Data center network, Elephant flow, Mice flooding, Online flow
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
一般社団法人 電子情報通信学会 信学技報 THE INSTITUTE OF ELECTRONICS, IEICE Technical Report INFORMATION AND COMMUNICATION ENGINEERS
This article is a technical report without peer review, and its polished and/or extended version may be published elsewhere.
Abstract Introduction of optical switching networks into Data Center networks (DCNs) attracts attention to reduce power consumption in Data Centers. We have proposed “HOLST” as a hybrid optical/electrical switching DCN architecture with optical TDM switching. In the “HOLST” architecture, an optical switching network is constructed with combining both MEMS and PLZT optical switches. To realize reducing power consumption in DCN, packet flow size detection method is required to classify flows into the electrical switching network, the optical TDM switching network, and the optical wavelength switching network. In addition, there are some requirements as the method of detecting flow size especially in DCN, such as high-speed detection for multi-characteristics of flows and overcoming traffic behavior e.g. mice flooding. In this paper, we will propose a flow classifying method using a hierarchical LRU (Least Recently Used) queues in a ToR (Top of Rack) switch. The proposed method takes in to account frequent flow arrival and traffic behavior in “HOLST”. We evaluate the accuracy of detecting flows and the effect of power reduction of the proposed method compared to conventional methods in DCN.
Keywords Data center network, Elephant flow, Mice flooding, Online flow detection, LRU
- 63 -
一般社団法人 電子情報通信学会 信学技報THE INSTITUTE OF ELECTRONICS, INFORMATION AND COMMUNICATION ENGINEERS
ている HOLST (High-speed optical layer 1 switch system for time slot switching based optical data center networks) [6, 13-17]への適用を目指したフロー判定手法として,階層型 LRU (Least Recently Used) キャッシュを用いた判定手法を提案する. 本論文の構成を以下に示す.2 章で光回線導入型ネ
エレファントフロー検出手法である.ALFE は既登録フローがキャッシュから追い出された回数 (Evict Times)とフロー長に高い相関があることを用いてフロー判定を行う.具体的には,Flow ID を登録するキャッシュを 3 つ(Elephant Eden,Rabbit Slam,Mice Tomb)設定し,Flow ID 到着時に Evict Times によって ID を挿入するキャッシュを選択する.エレファントフロー
は,マイスフラッディングによりキャッシュアウトさ
れても継続時間が長いため高 Evict Times を保有するため,キャッシュアウト後に到着したパケットによる
Flow ID の挿入箇所が Elephant Eden となり,マイスフラッディングの影響を受けずにエレファントフローの
題である.次に,ALFE では,マイスフラッディング対策のために全 Flow ID(一般に数万~数百万に達する )に対して Evict Times を保存するために Flow Records が必要となる.HOLST においては,256 台程度以上の ToR スイッチが DCN に接続されることが想定されており,巨大な Flow Records を ToR スイッチ
[3] D. Karthi and G. Das, “Post: A scalable optical data center network,” Proc. in 2013 IEEE International Conference on Advanced Networks and Telecommunications Systems (ANTS), Dec. 2013.
[4] K. Kitayama, Y. C. Huang, Y. Yoshida, R. Takahashi, and M. Hayashitani, “Optical packet and path switching intra-data center network: Enabling technologies and network performance with intelligent flow control,” Proc. in 2014 The European Conference on Optical Communication (ECOC), Sept. 2014.
- 67 -
[5] N. Farrington, G. Porter, S. Radhakrishnan, H. H. Bazzaz, V. Subramanya, Y. Fainman, G. Papen, and A. Vahdat,“Helios: a hybrid electrical/optical switch architecture for modular data centers,” Proc. in ACM SIGCOMM, pp. 339-350, Aug. 2010.
[6] M. Hirono, T. Sato, J. Matsumoto, S. okamoto, and N. Yamanaka, “HOLST: Architecture design of energy-efficient data center network based on ultra high-speed optical switch,” Proc. in IEEE LANMAN 2017, June 2017.
[7] A. Singla, A. Singh, and Y. Chen, “Osa: An optical switching architecture for data center networks with unprecedented flexibility,” in Presented as part of the 9th USENIX Symposium on Networked Systems Design and Implementation (NSDI 12). San Jose, CA: USENIX, pp. 239–252, 2012.
[8] A. Pal and K. Kant, “Roda: A reconfigurable optical data center network architecture,” in 2015 IEEE 40th Conference on Local Computer Networks (LCN), pp.561–569, Oct. 2015.
[9] M. C. Yuang, P. L. Tien, H. Y. Chen, W. Z. Ruan, T. K. Hsu, S. Zhong, J. Zhu, Y. Chen, and J. Chen, “Opmdc: Architecture design and implementation of a new optical pyramid data center network,” Journal of Lightwave Technology, vol. 33, no. 10, pp. 2019–2031, May 2015.
[10] Y. Lu, M. Wang, B. Prabhakar, and F. Bonomi, “Elephant Trap: A low cost device for identifying large flows,” Proc. in 15th IEEE Symposium on High-Performance Interconnects (HOTI), pp.99-105, Aug. 2007.
[11] T. Pan, X. Guo, C. Zhang, W. Meng and B. Liu, “ALFE: A replacement policy to cache elephant flows in the presence of mice flooding,” Proc. in 2012 IEEE International Conference on Communications (ICC), Ottawa, ON, 2012, pp. 2961-2965, June 2012.
[12] T. Benson, A. Akella, and D. A. Maltz, “Network traffic characteristics of data center in the Wild,” Proc. in ACM SIGCOMM Internet Measurement Conference (IMC), pp.267-280, Nov. 2010.
[16] A. Yamashita, W. Muro, M. Hirono, T. Sato, S. Okamoto, N. Yamanaka, and M. Veeraraghavan, “Hadoop Triggered Opt/Electrical Data-Center Orchestration Architecture for Reducing Power Consumption,” Proc. in 19th International Conference on Transparent Optical Networks (ICTON), Th.A6.4, July 2017.
[17] M. Hirono, W. Muro, S. Sekigawa, T. Sato, S. Okamoto, and N. Yamanaka, “Hadoop-based Application Triggered Automatic Flow Switching in Electrical/Optical Hybrid Data-Center Network,” to appear in 43rd European Conference on Optical
Communication (ECOC), Sept. 2017. [18] R. Pandya and J. Pandya, “C5.0 Algorithm to
Improved Decision Tree with Feature Selection and Reduced Error Pruning,” International Journal of Computer Application, Vol. 117, No.6, pp. 18-21, May 2015.
[19] The CAIDA anonymized 2014 Internet traces [Online]. Available: https://data.caida.org/datasets/passive-2014.