基于在线机制的云资源分配与定价的研究
发布时间：2019-01-26 17:38
【摘要】：云计算技术来自于传统的IT技术,是对传统的网络资源服务模式的变革,建立在网格计算、并行计算、虚拟化等IT技术之上。在千变万化的云服务中,云资源提供商以虚拟机的形式向用户提供资源,极大地提高了服务质量和效率。传统的云资源分配和定价的研究主要基于离线设置,对用户的请求做出假设,不能满足云市场下的动态分配,也不能很好地保障服务质量,云资源提供商的利益也会受到损害。针对云计算环境的动态性,以及资源的异构性等特点,如果能设计出合理高效的云资源分配和定价机制,对云资源提供商的利益以及服务质量的提升无疑都是十分有利的。本论文利用机制设计理论对虚拟机资源的分配和定价问题进行了如下探索研究:1、通过对机制设计理论、整数规划法以及VCG拍卖机制理论的研究,提出了一种最优的离线虚拟机分配和定价机制(VCG-VMAP),算法实现了其支付规则和分配规则,并理论证明得出该机制的个人理性和激励兼容性。但当处理数据增大时该机制不能较好地处理,并且不能很好地适应云计算环境的动态性。2、结合在线机制设计理论,提出了一种在线虚拟机的分配和定价的问题模型,设计出基于在线机制的虚拟机分配和定价机制(O-VMAP),并用算法实现其分配规则和支付规则,从在线机制设计理论的角度证明了该机制的激励兼容性和个人理性。该机制能有效地处理大规模数据,同时能很好地适应动态的云计算环境。3、通过具体实例,呈现了在线机制O-VMAP进行云资源分配和定价的具体过程,重点分析了该机制的分配规则和支付规则,进而完成资源的分配和定价,并对该机制的激励兼容性进行了分析。4、在云计算仿真平台CloudSim上通过设定基本类属性参数和扩展其基本功能进行仿真实验,对本文提出的离线机制VCG-VMAP和在线机制O-VMAP的算法进行编程实现。从五个评价指标:社会收益、执行时间、服务用户数、云资源提供商获得的利益、资源利用率对两种机制的特性进行仿真对比分析,仿真结果表明在线机制在动态的云计算环境中的有效性和适应性,并且能获得很好的社会收益。
[Abstract]:Cloud computing technology comes from the traditional IT technology. It is a transformation of the traditional network resource service mode. It is based on grid computing, parallel computing, virtualization and other IT technologies. In the ever-changing cloud services, cloud resource providers provide resources to users in the form of virtual machines, which greatly improve the quality of service and efficiency. The traditional research of cloud resource allocation and pricing is mainly based on offline setting, which makes assumptions about users' requests, which can not satisfy the dynamic allocation in cloud market, and can not guarantee the quality of service well, and the interests of cloud resource providers will also be damaged. In view of the dynamic nature of cloud computing environment and the heterogeneity of resources, if a reasonable and efficient cloud resource allocation and pricing mechanism can be designed, it is undoubtedly very beneficial to the benefit of cloud resource providers and the improvement of service quality. In this paper, the allocation and pricing of virtual machine resources are studied by mechanism design theory as follows: 1. Through the research of mechanism design theory, integer programming method and VCG auction mechanism theory. An optimal off-line virtual machine allocation and pricing mechanism (VCG-VMAP) is proposed in this paper. The algorithm implements its payment rules and allocation rules, and theoretically proves the mechanism's individual rationality and incentive compatibility. However, when the data is increased, the mechanism can not be handled well and can not adapt to the dynamic nature of cloud computing environment. 2. Combined with the theory of online mechanism design, a problem model of allocation and pricing of online virtual machines is proposed. A virtual machine allocation and pricing mechanism (O-VMAP) based on online mechanism is designed, and its allocation rules and payment rules are implemented by the algorithm. The incentive compatibility and individual rationality of the mechanism are proved from the point of view of online mechanism design theory. This mechanism can deal with large scale data effectively, at the same time, it can adapt to the dynamic cloud computing environment. 3. Through concrete examples, it presents the concrete process of online mechanism O-VMAP for cloud resource allocation and pricing. The allocation rules and payment rules of the mechanism are emphatically analyzed, and the allocation and pricing of resources are accomplished, and the incentive compatibility of the mechanism is analyzed. The algorithms of offline mechanism VCG-VMAP and online mechanism O-VMAP are programmed on the cloud computing simulation platform CloudSim by setting the basic class attribute parameters and extending its basic functions. From five evaluation indicators: social income, execution time, number of service users, the benefits of cloud resource providers, resource utilization, the characteristics of the two mechanisms are compared and analyzed. Simulation results show that the online mechanism is effective and adaptive in the dynamic cloud computing environment, and can achieve good social benefits.
【学位授予单位】：华北水利水电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP393.09
本文编号：2415731
[Abstract]:Cloud computing technology comes from the traditional IT technology. It is a transformation of the traditional network resource service mode. It is based on grid computing, parallel computing, virtualization and other IT technologies. In the ever-changing cloud services, cloud resource providers provide resources to users in the form of virtual machines, which greatly improve the quality of service and efficiency. The traditional research of cloud resource allocation and pricing is mainly based on offline setting, which makes assumptions about users' requests, which can not satisfy the dynamic allocation in cloud market, and can not guarantee the quality of service well, and the interests of cloud resource providers will also be damaged. In view of the dynamic nature of cloud computing environment and the heterogeneity of resources, if a reasonable and efficient cloud resource allocation and pricing mechanism can be designed, it is undoubtedly very beneficial to the benefit of cloud resource providers and the improvement of service quality. In this paper, the allocation and pricing of virtual machine resources are studied by mechanism design theory as follows: 1. Through the research of mechanism design theory, integer programming method and VCG auction mechanism theory. An optimal off-line virtual machine allocation and pricing mechanism (VCG-VMAP) is proposed in this paper. The algorithm implements its payment rules and allocation rules, and theoretically proves the mechanism's individual rationality and incentive compatibility. However, when the data is increased, the mechanism can not be handled well and can not adapt to the dynamic nature of cloud computing environment. 2. Combined with the theory of online mechanism design, a problem model of allocation and pricing of online virtual machines is proposed. A virtual machine allocation and pricing mechanism (O-VMAP) based on online mechanism is designed, and its allocation rules and payment rules are implemented by the algorithm. The incentive compatibility and individual rationality of the mechanism are proved from the point of view of online mechanism design theory. This mechanism can deal with large scale data effectively, at the same time, it can adapt to the dynamic cloud computing environment. 3. Through concrete examples, it presents the concrete process of online mechanism O-VMAP for cloud resource allocation and pricing. The allocation rules and payment rules of the mechanism are emphatically analyzed, and the allocation and pricing of resources are accomplished, and the incentive compatibility of the mechanism is analyzed. The algorithms of offline mechanism VCG-VMAP and online mechanism O-VMAP are programmed on the cloud computing simulation platform CloudSim by setting the basic class attribute parameters and extending its basic functions. From five evaluation indicators: social income, execution time, number of service users, the benefits of cloud resource providers, resource utilization, the characteristics of the two mechanisms are compared and analyzed. Simulation results show that the online mechanism is effective and adaptive in the dynamic cloud computing environment, and can achieve good social benefits.
【学位授予单位】：华北水利水电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP393.09
【参考文献】
相关期刊论文 前10条
1 马小龙;刘兰娟;;基于在线机制设计的私有云资源分配研究[J];计算机应用研究;2015年02期
2 沈张果;楼俊钢;马小龙;麻旺勇;;一种基于机制理论的云市场多类型资源分配定价机制[J];计算机科学;2014年06期
3 王霞俊;;CloudSim云计算仿真工具研究及应用[J];微型电脑应用;2013年08期
4 师雪霖清华大学计算机科学与技术系 清华信息科学与技术国家实验室(筹);徐恪;;云虚拟机资源分配的效用最大化模型[J];计算机学报;2013年02期
5 李小玲;王怀民;丁博;郭长国;李小勇;;虚拟网络映射问题研究及其进展[J];软件学报;2012年11期
6 胡志刚;刘艳;;云环境下基于组合双向拍卖的动态资源定价[J];计算机工程;2012年08期
7 刘正伟;文中领;张海涛;;云计算和云数据管理技术[J];计算机研究与发展;2012年S1期
8 李强;郝沁汾;肖利民;李舟军;;云计算中虚拟机放置的自适应管理与多目标优化[J];计算机学报;2011年12期
9 刘之家;张体荣;谢雄程;;基于云计算的“用户期待”任务调度算法的研究[J];大众科技;2011年04期
10 董晓霞;吕廷杰;;云计算研究综述及未来发展[J];北京邮电大学学报(社会科学版);2010年05期
相关硕士学位论文 前2条
1 陈靖宇;云计算资源分配的在线机制设计[D];杭州电子科技大学;2015年
2 冯伟;多目标优化的虚拟机调度模型与关键算法研究[D];复旦大学;2012年
，本文编号：2415731
本文链接：https://www.wllwen.com/guanlilunwen/ydhl/2415731.html
