基于古诺均衡的云资源定价策略研究
|
摘要
随着对计算能力、资源利用率和资源集中化的迫切需求,云计算应运而生。云计算最终目的是共享资源和服务,因此如何有效的管理以及分配资源成为云计算的重中之重。
云计算资源管理系统主要负责接受客户端任务请求,根据任务请求查找相应的资源来供客户使用,因此合理的调度资源是其核心。由于资源的使用并不是无偿的,资源使用者需要根据资源使用情况付费。因此资源调度的核心依据之一是价格,合理的资源定价策略能够保证资源调度的有效性。在经济社会的垄断市场中,不同的产品提供商之间通过竞争来获得期望的最大利润,产品价格由提供商之间的竞争决定。由于云计算与经济社会具有一定相似性,因而将垄断市场下的资源定价理论引入到云计算资源管理系统中,借由该理论来确定资源价格且实现资源提供者利益最大化需求。
本文从分析现有的资源管理模型入手,引入了资源初始价格策略,解决了多个资源提供者相互竞争情况下的资源定价问题。首先制定出资源描述标准,对资源提供者提供的资源进行量化,得出具体的资源描述参数。随后考虑在多个资源提供者共存的环境下,运用古诺均衡定价策略,得出均衡资源价格。由于该价格并不能从本质上反应定价策略的有效性,因此需要借助资源的调度过程来进行说明。在传统的Min-Min任务调度算法基础上,引入了价格机制,并且在云计算仿真平台上实现了该算法。从对实验结果分析可知,定价策略确实能有效的提高资源管理效率,并且在一定程度上满足用户需求。
Cloud computing appear as to the needs of resource calculate, utilization and centralization. The ultimate aim of the cloud computing is to sharing resources and services, therefore, how to manage and assign resources effectively become the most important part of cloud computing.
The cloud computing resources management system is mainly in charge of accept client requests, and find the corresponding resources for them, so the core of the management is reasonable scheduler. Users fee according to the service condition because of resources are not free, therefore, the core of resource schedule is based on reasonable price. A reasonable pricing policy can ensure the effectiveness of resource deployment. In the captive market, products providers through the competition obtain the maximum profits, and then the price of goods is decided by the competition among them. Because of Cloud computing is similar with the economic social in some aspects, this paper introduce the product pricing theory to cloud computing resource management, and maximum the interests of the resources provider.
This paper first analyzed the existing resources management model, and then introduced a price policy to solve the problem of how to pricing in the premise of resource providers compete among each other. Firstly, this paper worked out a resource describes standards, and then got a balance price according Cournot equilibrium strategy. But because the price cannot response the reasonable pricing, thus we need to simulate resources management process. Introduced the price mechanism based on the traditional min-min task scheduler algorithm, and the platform for the algorithm. Simulate the algorithm in the cloud computing platform. The pricing policy can improve resource management effectively, and meet users needs from the results of the experiments analysis.
云计算资源管理系统主要负责接受客户端任务请求,根据任务请求查找相应的资源来供客户使用,因此合理的调度资源是其核心。由于资源的使用并不是无偿的,资源使用者需要根据资源使用情况付费。因此资源调度的核心依据之一是价格,合理的资源定价策略能够保证资源调度的有效性。在经济社会的垄断市场中,不同的产品提供商之间通过竞争来获得期望的最大利润,产品价格由提供商之间的竞争决定。由于云计算与经济社会具有一定相似性,因而将垄断市场下的资源定价理论引入到云计算资源管理系统中,借由该理论来确定资源价格且实现资源提供者利益最大化需求。
本文从分析现有的资源管理模型入手,引入了资源初始价格策略,解决了多个资源提供者相互竞争情况下的资源定价问题。首先制定出资源描述标准,对资源提供者提供的资源进行量化,得出具体的资源描述参数。随后考虑在多个资源提供者共存的环境下,运用古诺均衡定价策略,得出均衡资源价格。由于该价格并不能从本质上反应定价策略的有效性,因此需要借助资源的调度过程来进行说明。在传统的Min-Min任务调度算法基础上,引入了价格机制,并且在云计算仿真平台上实现了该算法。从对实验结果分析可知,定价策略确实能有效的提高资源管理效率,并且在一定程度上满足用户需求。
Cloud computing appear as to the needs of resource calculate, utilization and centralization. The ultimate aim of the cloud computing is to sharing resources and services, therefore, how to manage and assign resources effectively become the most important part of cloud computing.
The cloud computing resources management system is mainly in charge of accept client requests, and find the corresponding resources for them, so the core of the management is reasonable scheduler. Users fee according to the service condition because of resources are not free, therefore, the core of resource schedule is based on reasonable price. A reasonable pricing policy can ensure the effectiveness of resource deployment. In the captive market, products providers through the competition obtain the maximum profits, and then the price of goods is decided by the competition among them. Because of Cloud computing is similar with the economic social in some aspects, this paper introduce the product pricing theory to cloud computing resource management, and maximum the interests of the resources provider.
This paper first analyzed the existing resources management model, and then introduced a price policy to solve the problem of how to pricing in the premise of resource providers compete among each other. Firstly, this paper worked out a resource describes standards, and then got a balance price according Cournot equilibrium strategy. But because the price cannot response the reasonable pricing, thus we need to simulate resources management process. Introduced the price mechanism based on the traditional min-min task scheduler algorithm, and the platform for the algorithm. Simulate the algorithm in the cloud computing platform. The pricing policy can improve resource management effectively, and meet users needs from the results of the experiments analysis.
引文
[1]. Zdnet,云计算与网格计算基于网络的不同技术[EB/Ol], http://cloud.it168.com/a2010/04 10/871/000000871762.shtml,2010-10-14
[2]. Lenk, Klems, Nimis. et al. What's inside the cloud? An architectural map of the cloud landscape[J], Proceedings of the 2009 ICSE Workshop on Software Engineering Challenges of Cloud Computing (CLOUD'09), May 23.2009, Vancouver, Canada.
[3]. T. O. Group. X/Open DCE:Remote Procedure Call[J], November,1995.
[4]. OMG, History of CORBA[EB/Ol],, http://www.omg.org/gettingstarted/history_of_corba.ht m, seen:2010/10/15
[5].Ian Foster, Carl Kesselman. The Grid:Blueprint for a New Computing Infrastructure [M], Morgan Kaufmann Publishers,1998.
[6]. Ian Foster, Carl Kesselman, Jeffrey, Steven, the Physiology of the Grid:An Open Grid Services Architecture for Distributed Systems Integration[M],2002-6-22
[7].百度百科,SOA[EB/Ol],, http://baike.baidu.com/view/21305.htm,2010-10-20
[8].陈康,郑纬民.云计算：系统实例与研究现状[J].软件学报,2009,20(5)：1337-1348.
[9].刘晓乐,计算机云计算及其实现技术分析[J],电子科技,2009,22(12)
[10].ARMBRUST M, FOX A, GRIFFITH R. Above the clouds:A Berkeley view of cloud computing[J],2009/2/10
[11]. Hao Li,Guo Tang,Joan Lu,Shaowen Yao, The QoS Resource Quantification Based on the Grid Banking Model[J], Proceedings of the First International Conference on 'Networked Digital Technologies' (NDT 2009) Ostrava, The Czech Republic, July 29-31,
[12].袁国骏,浅谈云计算及其发展应用[J],实验室科学,2009年4月第二期
[13].芒果, 谷歌推出新软件商店Google Apps Marketplace[EB/Ol],, http://www.techweb.com.cn/world/2010-03-10/554006.shtml,seen:2010/11/01
[14].王柏,徐六通,云计算[J],中兴通讯技术,2010年6月第16卷第3期
[15].IBM云计算中心 &HiPODS,智慧的地球[EB/Ol],, http://etc.chinabyte.com/09874/143 7/IBM_yjs01.pdf,2010-7-14
[16].百度百科,Windows Azure[EB/Ol],, http://www.microsoft.com/azure/webdev.mspx, 2010-11-1
[17].了解云计算标杆Google的技术构架[EB/Ol],, seen:2010-11-1
[18].”中国云”能否改变世界领空?”[EB/Ol],, http://www.cloud114.com/,seen:2010/11/1
[19]. eNet硅谷动力, 中国移动发布“大云”等研发成果[EB/01],,http://www.enet.com.cn/article/2010/0521/A20100521657286.shtml,seen:2010/11/1
[20]. 解放日报,云计算深度解读[EB/01],,http://www.chinacloud.cn/show.aspx?id=2341&cid=17,2010-11-2
[21]. M. Hibler, R. Ricci, L. Stoller, J. Duerig, S. Guruprasad,T. Stack, K. Webb, and J. Lepreau. Large-scale Virtualization in the Emulab Network Testbed[J]. In Proceedings of the 2008 USENIX Annual Technical Conference,2008.
[22]. Hewlett-Packard. HP Integrated Lights-Out 2 User Guide.Technical report[Z], HP,2009.
[23].百度百科,paas[EB/Ol],,http://baike.baidu.com/view/1413359.htm,2010-11-07
[24].李婷,李晓龙,云计算的资源管理方法研究[J],电脑与电信,2010年第一期
[25].百度百科,硬盘数据传输率[J],http://baike.baidu.com/view/355695.htm,seen:2010/12/01
[26]. SPARC Product Directory, "War of the Disks:Hard Disk Drives vs. Flash Solid Stat e Disks"[Z].Retrieved 2009-05-11.
[27]. ITU-T[EB/Ol],, http://www.itu.int/ITU-T/,2010-12-7
[28].王希,高速网络环境下网络带宽测试算法分析[J],南昌工程学院学报.2005年9月,第24卷第3期
[29]. Manish Jain C Dovrolis Pathload:A Measurement Tool for End-to-End Available Ba ndwidth[J],2002
[30]. C Dovrolis.P Ramanathanm.D Moore What Do Packet Dispersion Techniques Measur e[Z],2001
[31].佚名,云计算的优势和劣势[EB/01],, http://www.cnblogs.com/pccai/archive/2010/11/21/ 1883020.html,2010-12-8
[32].赵亮,胡娴等,基于计算经济的网格资源管理研究,计算机工程与科学[J],2008年第30卷第9期
[33].张艾斌,云计算模式与云安全问题研究[J], SCIENCE & TECHNOLOGY ASSOCIATIO N FORUM,2010年6月
[34]. Buyya, R, Abramson, D., and Giddy, J. An Economy Driven Resource Management Architecture for Global Computational Power Grids[R], Las Vegas:The 2000 International Conference on Parallel and Distributed Processing Techniques and Applications (PDPTA 20 00),2000.
[35].马永征.科学数据网格资源调度技术研究[D].北京：中国科学院计算技术研究所,2005.
[36].林渝.网格计算中的资源管理的研究[D]：南昌：南昌大学,2007.
[37].百科,市场机制[EB/Ol],,http://baike.baidu.com/view/131351.htm,2011-1-06 [38]. Hewlett-Packard. HP Integrated Lights-Out 2 User Guide.Technical report[Z], HP,200 9.
[39]. G.BitranandR.Caldentey,An over view of pricing models for revenue management[J], Manufacturing&ServiceOperationsManagement,2003
[40]. Buyya,Economic-based Distributed Resource Management and Scheduling for Grid Co mputing[D]
[41].经济学上的古诺定律和纳什均衡[EB/Ol],http://blog.sina.com.cn/s/blog_4c92b9050100dm n4.html,2010-12-24
[42].穆晓芳等,基于拍卖机制的网格在线信誉系统模型[J],计算机工程与设计,2008年4期
[43].徐海琛,高性能计算技术与大容量存储技术概述[J],国土资源信息化,2004年4期
[44].佚名,面向虚拟基础设施的云服务：IaaS和Eucalyptus[EB/Ol], http://networking.ctoci o.com.cn/tips/333/9378833.shtml,2011-03-12
[45].佚名,云服务环境：Eucalyptus[EB/Ol], http://www.oschina.net/p/eucalyptus,2011-03-12
[2]. Lenk, Klems, Nimis. et al. What's inside the cloud? An architectural map of the cloud landscape[J], Proceedings of the 2009 ICSE Workshop on Software Engineering Challenges of Cloud Computing (CLOUD'09), May 23.2009, Vancouver, Canada.
[3]. T. O. Group. X/Open DCE:Remote Procedure Call[J], November,1995.
[4]. OMG, History of CORBA[EB/Ol],, http://www.omg.org/gettingstarted/history_of_corba.ht m, seen:2010/10/15
[5].Ian Foster, Carl Kesselman. The Grid:Blueprint for a New Computing Infrastructure [M], Morgan Kaufmann Publishers,1998.
[6]. Ian Foster, Carl Kesselman, Jeffrey, Steven, the Physiology of the Grid:An Open Grid Services Architecture for Distributed Systems Integration[M],2002-6-22
[7].百度百科,SOA[EB/Ol],, http://baike.baidu.com/view/21305.htm,2010-10-20
[8].陈康,郑纬民.云计算：系统实例与研究现状[J].软件学报,2009,20(5)：1337-1348.
[9].刘晓乐,计算机云计算及其实现技术分析[J],电子科技,2009,22(12)
[10].ARMBRUST M, FOX A, GRIFFITH R. Above the clouds:A Berkeley view of cloud computing[J],2009/2/10
[11]. Hao Li,Guo Tang,Joan Lu,Shaowen Yao, The QoS Resource Quantification Based on the Grid Banking Model[J], Proceedings of the First International Conference on 'Networked Digital Technologies' (NDT 2009) Ostrava, The Czech Republic, July 29-31,
[12].袁国骏,浅谈云计算及其发展应用[J],实验室科学,2009年4月第二期
[13].芒果, 谷歌推出新软件商店Google Apps Marketplace[EB/Ol],, http://www.techweb.com.cn/world/2010-03-10/554006.shtml,seen:2010/11/01
[14].王柏,徐六通,云计算[J],中兴通讯技术,2010年6月第16卷第3期
[15].IBM云计算中心 &HiPODS,智慧的地球[EB/Ol],, http://etc.chinabyte.com/09874/143 7/IBM_yjs01.pdf,2010-7-14
[16].百度百科,Windows Azure[EB/Ol],, http://www.microsoft.com/azure/webdev.mspx, 2010-11-1
[17].了解云计算标杆Google的技术构架[EB/Ol],, seen:2010-11-1
[18].”中国云”能否改变世界领空?”[EB/Ol],, http://www.cloud114.com/,seen:2010/11/1
[19]. eNet硅谷动力, 中国移动发布“大云”等研发成果[EB/01],,http://www.enet.com.cn/article/2010/0521/A20100521657286.shtml,seen:2010/11/1
[20]. 解放日报,云计算深度解读[EB/01],,http://www.chinacloud.cn/show.aspx?id=2341&cid=17,2010-11-2
[21]. M. Hibler, R. Ricci, L. Stoller, J. Duerig, S. Guruprasad,T. Stack, K. Webb, and J. Lepreau. Large-scale Virtualization in the Emulab Network Testbed[J]. In Proceedings of the 2008 USENIX Annual Technical Conference,2008.
[22]. Hewlett-Packard. HP Integrated Lights-Out 2 User Guide.Technical report[Z], HP,2009.
[23].百度百科,paas[EB/Ol],,http://baike.baidu.com/view/1413359.htm,2010-11-07
[24].李婷,李晓龙,云计算的资源管理方法研究[J],电脑与电信,2010年第一期
[25].百度百科,硬盘数据传输率[J],http://baike.baidu.com/view/355695.htm,seen:2010/12/01
[26]. SPARC Product Directory, "War of the Disks:Hard Disk Drives vs. Flash Solid Stat e Disks"[Z].Retrieved 2009-05-11.
[27]. ITU-T[EB/Ol],, http://www.itu.int/ITU-T/,2010-12-7
[28].王希,高速网络环境下网络带宽测试算法分析[J],南昌工程学院学报.2005年9月,第24卷第3期
[29]. Manish Jain C Dovrolis Pathload:A Measurement Tool for End-to-End Available Ba ndwidth[J],2002
[30]. C Dovrolis.P Ramanathanm.D Moore What Do Packet Dispersion Techniques Measur e[Z],2001
[31].佚名,云计算的优势和劣势[EB/01],, http://www.cnblogs.com/pccai/archive/2010/11/21/ 1883020.html,2010-12-8
[32].赵亮,胡娴等,基于计算经济的网格资源管理研究,计算机工程与科学[J],2008年第30卷第9期
[33].张艾斌,云计算模式与云安全问题研究[J], SCIENCE & TECHNOLOGY ASSOCIATIO N FORUM,2010年6月
[34]. Buyya, R, Abramson, D., and Giddy, J. An Economy Driven Resource Management Architecture for Global Computational Power Grids[R], Las Vegas:The 2000 International Conference on Parallel and Distributed Processing Techniques and Applications (PDPTA 20 00),2000.
[35].马永征.科学数据网格资源调度技术研究[D].北京：中国科学院计算技术研究所,2005.
[36].林渝.网格计算中的资源管理的研究[D]：南昌：南昌大学,2007.
[37].百科,市场机制[EB/Ol],,http://baike.baidu.com/view/131351.htm,2011-1-06 [38]. Hewlett-Packard. HP Integrated Lights-Out 2 User Guide.Technical report[Z], HP,200 9.
[39]. G.BitranandR.Caldentey,An over view of pricing models for revenue management[J], Manufacturing&ServiceOperationsManagement,2003
[40]. Buyya,Economic-based Distributed Resource Management and Scheduling for Grid Co mputing[D]
[41].经济学上的古诺定律和纳什均衡[EB/Ol],http://blog.sina.com.cn/s/blog_4c92b9050100dm n4.html,2010-12-24
[42].穆晓芳等,基于拍卖机制的网格在线信誉系统模型[J],计算机工程与设计,2008年4期
[43].徐海琛,高性能计算技术与大容量存储技术概述[J],国土资源信息化,2004年4期
[44].佚名,面向虚拟基础设施的云服务：IaaS和Eucalyptus[EB/Ol], http://networking.ctoci o.com.cn/tips/333/9378833.shtml,2011-03-12
[45].佚名,云服务环境：Eucalyptus[EB/Ol], http://www.oschina.net/p/eucalyptus,2011-03-12