面向数据中心的虚拟机整合优化算法研究
发布时间：2018-09-06 10:34
【摘要】：数据中心的建设与发展将有利于数据中心的商业应用与运作,然而近年来,随着云计算的快速发展,通过互联网技术实现的各种应用层出不穷,为了满足日益增长的用户需求,数据中心的规模在不断的扩大,这给数据中心带来了一系列新的问题,如能耗、性能、安全、管理问题等等。特别是高能耗问题,而虚拟机整合是解决数据中心高能耗问题的手段之一。其思想是使用虚拟机实时迁移,以便一些轻载的物理机可以关闭或者切换到低功耗模式,并通过特定的目标函数寻找一个近似最优解。本文在分析现有虚拟机整合算法不足的基础上,提出两种改进的虚拟机整合算法,主要成果包括:首次提出了一种基于多种群蚁群算法的虚拟机整合算法。其思想是根据当前资源需求来减少活跃的物理机的数量,通过特定的目标函数建立迁移计划,各种群的信息熵来决定蚂蚁群体间的信息交流策略,以此来保证算法收敛性和多样性之间的平衡。通过仿真实验分别和基于蚁群系统的虚拟机整合算法以及基于向量代数的虚拟机整合算法进行比较,验证了该算法在降低能量消耗和减少虚拟机迁移次数方面的有效性。首次提出了一种基于文化-多种群蚁群算法的虚拟机整合算法。在该算法中,将多蚁群算法放入文化算法的种群空间中,并将种群空间中每代蚂蚁的最优值通过函数传入文化算法的信仰空间中。通过信仰空间中的进化算法对该方案进行进化,并指导种群空间进行优化,以期通过用最少的迁移次数来使更多的物理机进入休眠状态,来达到减少云数据中心能耗的目的。仿真实验验证了该算法不仅有效的降低了能量消耗而且减少了虚拟机迁移次数。
[Abstract]:The construction and development of the data center will benefit the commercial application and operation of the data center. However, with the rapid development of cloud computing in recent years, various applications realized through the Internet technology emerge one after another, in order to meet the increasing needs of users. The scale of the data center is expanding, which brings a series of new problems to the data center, such as energy consumption, performance, security, management and so on. Especially the problem of high energy consumption, and virtual machine integration is one of the methods to solve the problem of high energy consumption in data center. The idea is to use virtual machines to migrate in real time so that some light-duty physical machines can turn off or switch to low-power mode and find an approximate optimal solution through specific objective functions. Based on the analysis of the shortcomings of existing virtual machine integration algorithms, two improved virtual machine integration algorithms are proposed in this paper. The main achievements are as follows: for the first time, a virtual machine integration algorithm based on multi-colony ant colony algorithm is proposed. The idea is to reduce the number of active physical machines according to the current resource requirements, to establish migration plans through specific objective functions, and to determine the information exchange strategy among ant populations by the information entropy of various groups. In order to ensure the balance between convergence and diversity of the algorithm. The simulation results are compared with the algorithm based on ant colony system and the algorithm based on vector algebra respectively. The results show that the algorithm is effective in reducing the energy consumption and the number of times of virtual machine migration. A virtual machine integration algorithm based on culture-multi-colony ant colony algorithm is proposed for the first time. In this algorithm, the multi-ant colony algorithm is put into the population space of the cultural algorithm, and the optimal value of each generation ant in the population space is passed into the belief space of the cultural algorithm through the function. The scheme is evolved by evolutionary algorithms in belief space, and the population space is optimized to reduce the energy consumption of cloud data centers by making more physical machines dormant with the least number of migration times. The simulation results show that the proposed algorithm not only reduces the energy consumption but also reduces the number of virtual machine migration.
【学位授予单位】：天津工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP302;TP308
本文编号：2226096
[Abstract]:The construction and development of the data center will benefit the commercial application and operation of the data center. However, with the rapid development of cloud computing in recent years, various applications realized through the Internet technology emerge one after another, in order to meet the increasing needs of users. The scale of the data center is expanding, which brings a series of new problems to the data center, such as energy consumption, performance, security, management and so on. Especially the problem of high energy consumption, and virtual machine integration is one of the methods to solve the problem of high energy consumption in data center. The idea is to use virtual machines to migrate in real time so that some light-duty physical machines can turn off or switch to low-power mode and find an approximate optimal solution through specific objective functions. Based on the analysis of the shortcomings of existing virtual machine integration algorithms, two improved virtual machine integration algorithms are proposed in this paper. The main achievements are as follows: for the first time, a virtual machine integration algorithm based on multi-colony ant colony algorithm is proposed. The idea is to reduce the number of active physical machines according to the current resource requirements, to establish migration plans through specific objective functions, and to determine the information exchange strategy among ant populations by the information entropy of various groups. In order to ensure the balance between convergence and diversity of the algorithm. The simulation results are compared with the algorithm based on ant colony system and the algorithm based on vector algebra respectively. The results show that the algorithm is effective in reducing the energy consumption and the number of times of virtual machine migration. A virtual machine integration algorithm based on culture-multi-colony ant colony algorithm is proposed for the first time. In this algorithm, the multi-ant colony algorithm is put into the population space of the cultural algorithm, and the optimal value of each generation ant in the population space is passed into the belief space of the cultural algorithm through the function. The scheme is evolved by evolutionary algorithms in belief space, and the population space is optimized to reduce the energy consumption of cloud data centers by making more physical machines dormant with the least number of migration times. The simulation results show that the proposed algorithm not only reduces the energy consumption but also reduces the number of virtual machine migration.
【学位授予单位】：天津工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP302;TP308
【参考文献】
相关期刊论文 前4条
1 赵君;马中;刘驰;李海山;王新余;;一种多目标蚁群优化的虚拟机放置算法[J];西安电子科技大学学报;2015年03期
2 叶可江;吴朝晖;姜晓红;何钦铭;;虚拟化云计算平台的能耗管理[J];计算机学报;2012年06期
3 李刚健;;基于虚拟化技术的云计算平台架构研究[J];吉林建筑工程学院学报;2011年01期
4 郭一楠;王辉;;文化算法研究综述[J];计算机工程与应用;2009年09期
，本文编号：2226096
本文链接：https://www.wllwen.com/kejilunwen/jisuanjikexuelunwen/2226096.html
