现代数据中心网络资源管理技术分析与综述

doi:10.3969/j.issn.1000-436x.2014.02.021

摘要/Abstract

摘要：

现代数据中心网络结构特征和应用模式的深刻变革，给网络资源管理带来了全新的挑战。地址自动配置技术、传输控制技术、流量管理技术以及虚拟化管理技术等是现代数据中心网络资源管理的重要内容，也是近年来学术界研究的重要方向。结合当前的研究现状，对以上几个方面的最新研究成果进行分析综述，并就网络资源管理未来的发展趋势进行展望。

关键词: 数据中心, 网络资源管理, 地址自动配置, 拥塞控制, 流量管理, 虚拟化管理, 云计算

Abstract:

Data center is the basic facility of Internet and cloud computing. Since network resource is critical resource in data center thorough study and analysis of its management, which can help to improve data center's performance, save costs and expenses, and is of great significance. However, as the structure characteristic and application mode are pro-found changing, it brings tremendous challenges to modern data center network resource management. Currently, lots of studies have been made in this field. In order to have an overall perspective of current research, some of the most impor-tant aspects were summarized and analyzed, which are automatic address configuration, congestion control, network traf-fic management and virtualization management. They are also the hotspots in current academe. Based on the comprehen-sive comparison and analysis, some trends of the future data center network resource management were pointed out in the end.

Key words: data center, network resource management, automatic address configuration, congestion control, traffic management, virtualization management, cloud computing

邓罡,龚正虎,王宏,陈琳,刘志宏. 现代数据中心网络资源管理技术分析与综述[J]. 通信学报, 2014, 35(2): 166-181.

ENGGang D,Zheng-hu GONG,Hong WANG,Lin CHEN,Zhi-hong LIU. Analysis and survey of resource management on modern data center networks[J]. Journal on Communications, 2014, 35(2): 166-181.

图/表 6

图1

表1

图2

表2

图3

表3

参考文献 62

[1]	ZHANG W , SONG Y RUAN L et al. Resource management in internet-oriented data centers[J]. Journal of Software, 2012,23(2): 179-199.
[2]	LIN W , QI D . Survey of resource scheduling in cloud computing[J]. Computer Science, 2012,29(10): 1-6.
[3]	QIAN Q , LI C , ZHANG X et al. Survey of virtual resource management in cloud data center[J]. Application Research of Computers, 2012,29(7): 2411-2415.
[4]	Dynamic Host Configuration Protocol[S]. RFC2131, 1997.
[5]	Dynamic Configuration of IPv4 Link-Local Addresses[S]. RFC3927, 2005.
[6]	AL-FARES M , LOUKISSAS A , VAHDAT A . A scalable, commodity data center network architecture[A]. Proc of SIGCOMM '08[C]. Seattle, WA, USA, 2008.63-74.
[7]	MYSORE R N , PAMBORIS A , FARRINGTON N et al. PortLand: a scalable fault-tolerant layer 2 data center network fabric[A]. Proc of SIGCOMM '09[C]. Barcelona, Spain. 2009.
[8]	GUO C , WU H , TAN K et al. Dcell: a scalable and fault-tolerant network structure for data centers[A]. Proc of SIGCOMM'08[C]. Seattle, WA, USA, 2008.75-86.
[9]	GUO C , LU G , LI D et al. Bcube: a high performance, server-centric network architecture for modular data centers[A]. Proc of SIGCOMM'09[C]. Barcelona, Spain, 2009.
[10]	CHEN K , GUO C , WU H et al. Generic and automatic address configuration for data center networks[A]. Proc of SIGCOMM'10[C]. New Delhi, India, 2009.39-50.
[11]	MA X , HU C , CHEN K et al. Error tolerant address configuration for data center networks with malfunctioning devices[A]. Proc of INFOCOM'12[C]. Florida, USA, 2012.
[12]	HU C , YANG K , ZHENG H et al. Automatically configuring the network layer of data centers for cloud computing[A]. IBM Journal of Research and Development, 2011,55(6): 1-10.
[13]	GREENBERG A , HAMILTON J R , JAIN N et al. VL2: a scalable and flexible data center network[A]. Proc of SIGCOMM'09[C]. Barcelona, Spain, 2009. 51-62.
[14]	ALIZADEH M , GREENBERG A , MALTZ D A et al. Data center TCP (DCTCP)[A]. Proc of SIGCOMM'10[C]. New Delhi, India, 2010.
[15]	VAMANAN B , HASAN J , VIJAYKUMAR T N . Deadline-aware datacenter TCP (D2TCP)[A]. Proc SIGCOMM'12[C]. Helsinki, Finland, 2012.
[16]	WU H , FENG Z , GUO C et al. ICTCP: incast congestion control for TCP in data center networks[A]. ACM CoNEXT'10[C]. Philadelphia, USA, 2010.
[17]	WILSON C , BALLANI H , KARAGIANNIS T et al. Better never than late: meeting deadlines in datacenter networks[A]. Proc SIGCOMM'11[C]. Toronto, Ontario, Canada, 2011.
[18]	HONG C , CAESAR M , GODFREY P B . Finishing flows quickly with preemptive scheduling[A]. Proc of SIGCOMM'12[C]. Helsinki, Finland, 2012.
[19]	ZATS D , DAS T , MOHAN P et al. DeTail: reducing the flow completion time tail in datacenter networks[A]. Proc of SIGCOMM'12[C]. Helsinki, Finland, 2012.
[20]	KANDULA S , SENGUPTA S , GREENBERG A et al. The nature of data center traffic: measurements and analysis[A]. Proc of IMC'09[C]. Chicago, Illinois, USA, 2009.202-208.
[21]	BENSON T , AKELLA A , MALTZ D A . Network traffic characteristics of data centers in the wild[A]. Proc of IMC'10[C]. Melbourne, Australia, 2010.267-280.
[22]	SHIEH A , KANDULA S , GREENBERG A et al. Sharing the data center network[A]. Proc of Usenix NSDI'11[C]. Berkeley, CA,USA, 2011.
[23]	BENSON T , ANAND A , AKELLA A et al. Understanding data center traffic characteristics[A]. Proc of SIGCOMM'09[C]. Barcelona, Spain, 2009.92-99.
[24]	WANG M , MENG X , ZHANG L . Consolidating virtual machines with dynamic bandwidth demand in data centers[A]. Proc of INFOCOM'11[C]. Shanghai, China, 2011.
[25]	EPSTEIN A , BREITGAND D . Improving consolidation of virtual machines with risk-aware bandwidth oversubscription in compute clouds[A]. Proc of INFOCOM'12[C]. Florida, USA, 2012.
[26]	AL-FARES M , RADHAKRISHNAN S , RAGHAVAN B et al. Hedera:dynamic flow scheduling for data center networks[A]. Proc of Usenix NSDI'10[C]. California, USA, 2010.
[27]	MCKEOWN N , ANDERSON T , BALAKRISHNAN H et al. Open-Flow: enabling innovation in campus networks[J]. Computer Communication Review, 2008,38(2): 69-74.
[28]	MENG X , PAPPAS V , ZHANG L . Improving the scalability of data center networks with traffic-aware virtual machine placement[A]. Proc of INFOCOM'10[C]. San, Diego, CA, USA, 2010.
[29]	BIRAN O , CORRADI A , FANELLI M et al. A stable network-aware vm placement for cloud systems[A]. Proc of 12th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing[C]. Ottawa, Canada, 2012.
[30]	JIANG J , LAN T , HA S et al. Joint VM placement and routing for data center traffic engineering[A]. Proc of INFOCOM'12[C]. Florida, USA, 2012.
[31]	HOPPS C . Analysis of an Equal-Cost Multi-Path Algorithm[S]. RFC 2992, IETF, 2000.
[32]	RAICIUC , PLUNTKE C , BARRE S et al. Data center networking with multipath TCP[A]. Proc of HOTNETS '10[C]. Monterey, CA, USA, 2010.
[33]	FORD A , RAICIU C , HANDLEY M et al. TCP Extensions for Multipath Operation with Multiple Addresses[A]. Internet-draft, IETF, 2012.
[34]	Open vswitch project[EB/OL]. .
[35]	VMWare vSphere: vnetwork distributed switch[EB/OL]. .
[36]	Cisco nexus 1000V series switches[EB/OL]. .
[37]	LUO Y , MURRAY E , FICARRA T L . Accelerated virtual switching with programmable NICs for scalable data center networking[A]. VISA'10[C]. New York, NY, USA, 2010.
[38]	XU L , ZHANG Y , WU J et al. Network technology research under cloud computing environment[A]. Journal on Communications[C]. New York, NY, USA, 2012,33(Z1): 16-221.
[39]	802.1Qbg-edge Virtual Bridging[EB/OL]. .
[40]	802.1Qbh-bridge port extension[EB/OL]. .
[41]	PCI-SIG. Single root I/O virtualization and sharing specication[EB/OL]. .
[42]	DONG Y , YANG X , LI J et al. High performance network virtualization with SR-IOV[A]. Journal of Parallel and Distributed Computing[C]. 2012,(72): 1471-1482.
[43]	TRIPATHI S , DROUX N , SRINIVASAN T . Crossbow: from hardware virtualized NICs to virtualized networks[A]. Proc of VISA'09[C]. Barcelona, Spain, 2009.
[44]	RAM K K , MUDIGONDA J , COX L A . sNICh: efficient last hop networking in the data center[A]. Proc of ANCS '10[C]. San Diego, CA, USA, 2010.
[45]	NAHIR A , ORDA A , RAZ D . Workload factoring with the cloud: a game-theoretic perspective[A]. Proc of INFOCOM'12[C]. Florida, USA, 2012.
[46]	HAYES B . Cloud computing[A]. Commun, ACM[C]. 2008,51(7): 9-11.
[47]	LUO J , JIN J , SONG A et al. Cloud computing: architecture and key technologies[A]. Journal on Communications[C]. 2011,32(7): 3-21.
[48]	FEHLING C , LEYMANN F , MIETZNER R . A framework for optimized distribution of tenants in cloud applications[A]. Proc of 3rd International Conference on Cloud Computing'10[C]. Miami, FL, USA, 2010.
[49]	Amazon elastic compute cloud[EB/OL]. .
[50]	Amazon simple storage service[EB/OL]. .
[51]	IBM blue cloud project[EB/OL]. .
[52]	802.1Q-virtual LANs[EB/OL]. .
[53]	802.1ad-provider bridges[EB/OL]. .
[54]	802.1ah-provider backbone bridges[EB/OL]. .
[55]	802.1aq-shortest path bridging[EB/OL]. .
[56]	GUO C , LU G , WANGH J et al. SecondNet: a data center network virtualization architecture with bandwidth guarantees[A]. Proc of Philadelphia[C]. New York, USA, ACM, 2010.
[57]	BALLANI H , COSTA P , KARAGIANNIS T et al. Towards predictable datacenter networks[A]. Proc of SIGCOMM'11[C]. New York, ACM, 2011.242-253.
[58]	QIN A , YU H . Research on the multi-tenant isolation mechanism based on indentity mapping[J]. Chinese High Technology Letters, 2011,21(10): 1007-1013.
[59]	MUDIGONDA J , STIEKES B , YALAGANDULA P et al. NetLord: a scalable multi-tenant network architecture for virtualized datacenters[A]. Proc of SIGCOMM'11, Toronto, Canada, 2011.
[60]	POPA L , KRISHNAMURTHY A , RATNASAMY S et al. FairCloud:sharing the network in cloud computing[A]. Proc of HOTNETS'11, Cambridge, MA,USA, 2010.
[61]	KERRAVALA Z . As the value of enterprise networks escalates, so does the need for configuration management[EB/OL]. .
[62]	What's behind network downtime?[EB/OL]. .

名称	配置类型	配置规模	实现方式	适用范围	容错性	动态适应性	配置效率		适用于数据中心
名称	配置类型	配置规模	实现方式	适用范围	容错性	动态适应性	无故障	带故障	适用于数据中心
DHCP	主机	K级	软件实现	同一子网	高	高	高	高	否
Zeroconf	主机	K级	软件实	同一子网	高	高	较高	较高	否
DAC	主机及网络中间设备	M级	软件实现	拓扑相关数据中心	低	低	高	低	是
ETAC	主机及网络中间设备	M级	软件实现	拓扑相关数据中心	较高	低	较高	较高	是
DCZerocon	主机及网络中间设备	M级	额外硬件设备	拓扑无关数据中心	高	高	高	高	是
注：中间设备指的是交换机或路由器等网络中间连接设备，容错性高指的是在网络发生错误时仍能对地址进行配置，较高指的是仅能对部分设备进行配置，而低则指不能完成配置。

名称	软超时敏感	抢先调度	交换机支持	应用程序支持	兼容TCP
DCTCP	否	否	ECN	无	是
ICTCP	否	否	无	无	是
D³	是	否	速率请求应答	提供超时时限	否
D²TCP	是	否	ECN	提供超时时限	是
PDQ	是	是	流分级、抢先调度	提供超时时限	否
DeTail	是	否	无	提供超时时限	否

虚拟化方法	虚拟化方式	虚拟化目标	虚拟机迁移	访问控制	资源利用率	多路径
VLAN	静态	资源隔离	不支持	支持	低	不支持
VL2	动态	资源整合	支持	支持	高	支持
PortLand	动态	资源整合	支持	支持	高	支持
Seawall	动态	公平共享	—	—	高	—
SecondNet	动态	性能隔离	不支持	支持	高	不支持
Oktopus	动态	性能隔离	—	—	高	—
NetLord	动态	资源整合性能隔离	支持	支持	高	支持
OpenFlow	动态	性能隔离	支持	支持	高	支持