[1]张登科,王兴伟,黄敏.2-Tier:共生抗倾斜可演化部署数据中心网络体系结构[J].东南大学学报(自然科学版),2020,50(2):402-408.[doi:10.3969/j.issn.1001-0505.2020.02.026]
 Zhang Dengke,Wang Xingwei,Huang Min.2-Tier: Data center network architecture with symbiotic traffic-skew mitigated capability and evolutionary deployment[J].Journal of Southeast University (Natural Science Edition),2020,50(2):402-408.[doi:10.3969/j.issn.1001-0505.2020.02.026]
点击复制

2-Tier:共生抗倾斜可演化部署数据中心网络体系结构()
分享到:

《东南大学学报(自然科学版)》[ISSN:1001-0505/CN:32-1178/N]

卷:
50
期数:
2020年第2期
页码:
402-408
栏目:
计算机科学与工程
出版日期:
2020-03-20

文章信息/Info

Title:
2-Tier: Data center network architecture with symbiotic traffic-skew mitigated capability and evolutionary deployment
作者:
张登科1王兴伟1黄敏2
1东北大学计算机科学与工程学院, 沈阳 110004; 2东北大学信息科学与工程学院, 沈阳 110004
Author(s):
Zhang Dengke1 Wang Xingwei1 Huang Min2
1 College of Computer Science and Engineering, Northeastern University, Shenyang 110004, China
2 College of Information Science and Engineering, Northeastern University, Shenyang 110004, China
关键词:
数据中心网络 流量倾斜 演化部署 可重构 共生
Keywords:
data center network traffic skew evolutionary deployment reconfiguration symbioses
分类号:
TP393
DOI:
10.3969/j.issn.1001-0505.2020.02.026
摘要:
为了解决超大规模数据中心网络中的流量倾斜问题,提出了一种新型数据中心网络体系结构2-Tier. 首先,考虑到超大规模数据中心网络存在结构对称性和用户应用数据倾斜性之间的天然矛盾,设计了具有共生抗倾斜能力且可演化部署的2层网络体系结构,并给出其网络模型,上层基于覆盖网络技术实现按需移动式拓扑,下层采用当前数据中心最佳实践的对称式网络拓扑. 进而,基于多重图设计了该网络体系结构的数学模型. 随后基于可重构底层通信技术给出技术实现方案,并设计了3种可演化部署技术路线. 从算法实现角度探讨了2类基于2-Tier架构的流量热区应对方式. 理论分析证明2-Tier是有效的,能够降低最大边介数值,提高网络传输容量,改善网络流量倾斜性. 仿真实验结果表明,相比于现有可重构架构,2-Tier架构带来26.1%的最大边介数下降,其增链效率提升13%. 2-Tier层间拓扑级问题的进一步研究有望促进更具弹复性新型物理层通信技术的出现.
Abstract:
To solve the traffic-skew problem of the hyperscale data center networks, a new data center network architecture, called the 2-Tier, was proposed. First, considering the natural contradiction between the structural symmetry of hyperscale data center networks and the user application data skew, a two-layer network architecture of symbiotic traffic-skew mitigated capability and evolutionary deployment was designed, and its network model was given. The upper layer implemented the on-demand mobile topology based on Overlay technology, and the underlying layer adopted the symmetric network topology based on current data center best practices. Then, its mathematical model was designed based on multi-graphs. And then based on the reconfigurable underlying communication technology, the technical implementation scheme was presented, and three evolvable deployment technology routes were designed. From the perspective of algorithm implementation, two kinds of detection methods in traffic hot zone based on 2-Tier architecture were discussed. The theoretical analysis shows that the 2-Tier architecture is effective, reducing the value of the maximum edge betweenness, increasing the transmission capacity and mitigate the traffic-skew. Simulation results show that compared with the existing reconfigurable architectures, the 2-Tier architecture can reduce the maximum edge betweenness by 26.1% and increase the link augment efficiency by 13%. The further research on the topological level of 2-Tier is expected to promote the emergence of some new physical layer communication technologies with more resilience.

参考文献/References:

[1] Choi S, Burkov B, Eckert A, et al. FBOSS: Building switch software at scale[C]//Proceedings of the 2018 Conference of the ACM Special Interest Group on Data Communication. Budapest, Hungary, 2018: 342-356. DOI:10.1145/3230543.3230546.
[2] Singh A, Ong J, Agarwal A, et al. Jupiter rising: A decade of clos topologies and centralized control in Google’s datacenter network[C]// Proceedings of the 2015 ACM SIGCOMM Conference. New York, USA, 2015: 183-197. DOI: 10.1145/2829988.2787508.
[3] HUAWEI. Global connectivity index[EB/OL].(2018-05-29)[2019-06-29]. https://www.huawei.com/minisite/gci/en/.
[4] Grand Research Store. Hyperscale data centers market insights 2019, global and Chinese analysis and forecast to 2024[EB/OL].(2019-05-16)[2019-06-29]. https://www.grandresearchstore.com/services/global-hyperscale-data-centers-2019-2024-542.
[5] Synergy Research Group. Hyperscale data center count approaches 400 mark[EB/OL].(2019-01-22)[ 2019-06-29]. https://dwz.cn/wQOZFAYQ.
[6] Chowdhury M, Zhong Y, Stoica I. Efficient coflow scheduling with varys[C]//Proceedings of the 2014 ACM Conference on SIGCOMM-SIGCOMM ’14. Chicago, Illinois, USA, 2014: 443-454. DOI:10.1145/2619239.2626315.
[7] Al-Fares M, Radhakrishnan S, Raghavan B, et al. Hedera: Dynamic flow scheduling for data center networks[C]// 2010 7th USENIX Symposium on Networked Systems Design and Implementation. San Jose, USA, 2010: 281-296. DOI: 10.1.1.212.9046.
[8] Apache Flink. Stateful computations over data streams[EB/OL].(2019-06-29)[ 2019-06-29]. https://flink.apache.org.
[9] Abadi M, Barham P, Chen J, et al. Tensorflow: A system for large-scale machine learning[C]// 2016 12th USENIX Symposium on Operating Systems Design and Implementation. Savannah, USA, 2016: 265-283.
[10] Vanini E, Pan R, Alizadeh M, et al. Let it flow: Resilient asymmetric load balancing with flowlet switching[C]// 2017 14th USENIX Symposium on Networked Systems Design and Implementation. Boston, USA, 2017: 407-420.
[11] 李丹, 陈贵海, 任丰原, 等. 数据中心网络的研究进展与趋势[J]. 计算机学报, 2014, 37(2): 259-274. DOI:10.3724/SP.J.1016.2014.00259.
Li D, Chen G H, Ren F Y, et al. Data center network research progress and trends[J].Chinese Journal of Computers, 2014, 37(2): 259-274. DOI:10.3724/SP.J.1016.2014.00259. (in Chinese)
[12] 魏祥麟, 陈鸣, 范建华, 等. 数据中心网络的体系结构[J]. 软件学报, 2013, 24(2): 295-316.
  Wei X L, Chen M, Fan J H, et al. Architecture of the data center network[J].Journal of Software, 2013, 24(2): 295-316.(in Chinese)
[13] 王聪, 王翠荣, 王兴伟, 等. 面向云计算的数据中心网络体系结构设计[J]. 计算机研究与发展, 2012, 49(2): 286-293.
  Wang C, Wang C R, Wang X W, et al. Network architecture design for data centers towards cloud computing[J].Journal of Computer Research and Development, 2012, 49(2): 286-293.(in Chinese)
[14] Kandula S, Padhye J, Bahl P. Flyways to de-congest data center networks[C]// Proceedings of the 8th ACM Workshop on Hot Topics in Networks(HotNets’09). New York, USA, 2009: 1-6.
[15] CALIENT Technologies. Photonic optical circuit switching[EB/OL].(2019-06-29)[2019-06-29]. https://www.calient.net/.
[16] Chen K, Singla A, Singh A, et al. OSA:An optical switching architecture for data center networks with unprecedented flexibility[J]. ACM Transactions on Networking, 2014, 22(2): 498-511. DOI:10.1109/tnet.2013.2253120.
[17] Ghobadi M, Mahajan R, Phanishayee A, et al. Projector: Agile reconfigurable data center interconnect[C]// Proceedings of the 2016 ACM SIGCOMM Conference. New York, USA, 2016: 216-299. DOI: 10.1145/2934872.2934911.
[18] Farrington N, Porter G, Radhakrishnan S, et al. Helios: A hybrid electrical/optical switch architecture for modular data centers[J]. ACM SIGCOMM Computer Communication Review, 2010, 41(4): 339-350. DOI:10.1145/1851182.1851223.
[19] Singla A, Hong C Y, Popa L, et al. Jellyfish: Networking data centers randomly[C]// 2012 9th USENIX Symposium on Networked Systems Design and Implementation. San Jose, CA, 2012: 225-238.
[20] Valadarsky A, Shahaf G, Dinitz M, et al. Xpander: Towards optimal-performance datacenters [C]//Proceedings of the 12th International Conference on Emerging Networking Experiments and Technologies-CoNEXT ’16. Irvine, California, USA, 2016: 205-219. DOI:10.1145/2999572.2999580.
[21] Chatzieleftheriou A, Legtchenko S, Williams H, et al. Larry: Practical network reconfigurability in the data center[C]// 2018 15th USENIX Symposium on Networked Systems Design and Implementation. Santa Clara, USA, 2018: 141-156.
[22] Singla A, Singh A, Ramachandran K. Proteus: A topology malleable data center network[C]// Proceedings of the 9th ACM SIGCOMM Workshop on Hot Topics in Networks. New Delhi, India, 2010: 1-6. DOI: 10.1145/1868447.1868455.
[23] Testa F, Pavesi L. Optical switching in next generation data centers[M]. Berlin: Springer International Publishing, 2018: 23-44. DOI:10.1007/978-3-319-61052-8.
[24] Shu R, Cheng P, Chen G, et al. Direct universal access: Making data center resources available to FPGA[C]// 2019 16th USENIX Symposium on Networked Systems Design and Implementation. Boston, USA, 2019: 127-140.
[25] Wang M, Cui Y, Xiao S, et al. Neural network meets DCN: Traffic-driven topology adaptation with deep learning[C]// Proceedings of the ACM on Measurement and Analysis of Computing Systems. New York, USA, 2018: 2601-2625. DOI: 10.1145/3224421.
[26] Halperin D, Kandula S, Padhye J, et al. Augmenting data center networks with multi-gigabit wireless links[C]//Proceedings of the ACM SIGCOMM 2011 Conference on SIGCOMM-SIGCOMM ’11. Toronto, Ontario, Canada, 2011: 38–49. DOI:10.1145/2018436.2018442.
[27] Luo L L, Guo D K, Wu J, et al. VLCcube: A VLC enabled hybrid network structure for data centers[J]. IEEE Transactions on Parallel and Distributed Systems, 2017, 28(7): 2088-2102. DOI:10.1109/tpds.2016.2646366.
[28] Legtchenko S, Chen N, Cletheroe D, et al. XFabric: A reconfigurable in-rack network for rack-scale computers[C]// 2016 13th USENIX Symposium on Networked Systems Design and Implementation. Santa Clara, USA, 2016: 15-29.
[29] HUAWEI. AI fabric intelligent and lossless data center network[EB/OL].(2019-05-01)[2019-06-29]. https://dwz.cn/56LPBJsc.
[30] Fiedler M. Algebraic connectivity of graphs[J]. Czechoslovak Mathematical Journal, 1973, 23(2): 298-305.
[31] Chowdhury M, Zaharia M, Ma J, et al. Managing data transfers in computer clusters with orchestra[C]//Proceedings of the ACM SIGCOMM 2011 Conference on SIGCOMM-SIGCOMM ’11. Toronto, Ontario, Canada, 2011: 98-109. DOI:10.1145/2018436.2018448.
[32] Pothen A, Simon H D, Liou K P. Partitioning sparse matrices with eigenvectors of graphs[J]. SIAM Journal on Matrix Analysis and Applications, 1990, 11(3): 430-452. DOI:10.1137/0611030.
[33] Lee Y T, Sun H. Constructing linear-sized spectral sparsification in almost-linear time[J]. SIAM Journal on Computing, 2018, 47(6): 2315-2336. DOI:10.1137/16m1061850.

备注/Memo

备注/Memo:
收稿日期: 2019-09-26.
作者简介: 张登科(1981—),男,博士生;王兴伟(联系人),男,博士,教授,博士生导师,wangxw@mail.neu.edu.cn.
基金项目: 国家自然科学基金资助项目(61872073)、国家重点研发计划资助项目(2017YFB0801701).
引用本文: 张登科,王兴伟,黄敏.2-Tier:共生抗倾斜可演化部署数据中心网络体系结构[J].东南大学学报(自然科学版),2020,50(2):402-408. DOI:10.3969/j.issn.1001-0505.2020.02.026.
更新日期/Last Update: 2020-03-20