【作者】 王晨光；

【导师】 辛阳；

【作者基本信息】 北京邮电大学 ， 网络空间安全， 2022， 硕士

【摘要】 当前云计算与虚拟化的迅速发展以及混合云架构的广泛使用,为政府和企业带来了数据与业务的新载体,并为其信息化建设提供了坚实基础,促进政企数据及业务上云的进程。政企在云环境应用中,具有多样的、差异化的需求,云环境也随之变得愈加复杂。私有云与公有云的搭建与部署也带来了混合云架构的广泛使用,而混合云在满足建设者不同需求时也带来了异构多云面临的问题。多云环境下的云服务,虚拟机作为其服务单元和基本载体,直接决定了用户使用和感知到的云服务质量,因此本文选择虚拟机服务质量作为研究对象,对虚拟机QoS进行评估,以判断用户使用云服务的质量。本文中对影响虚拟机QoS的影响因素进行选取,将其分为易量化的虚拟机配置指标及不易量化的虚拟机自身属性。用户在选择云服务商购买云服务时可能出现因预算过多造成的性能浪费或购买依据经验和主观判断等情况,本文针对此问题,提出了一种面向用户的多云环境下虚拟机QoS的评估模型,通过梯度下降的非线性评估算法,将虚拟机量化指标进行性能评估,同时,对虚拟机不易量化的属性进行评估,并通过实验验证对虚拟机不易量化的属性与量化指标评估的相关性,将二者整合为多云环境下虚拟机QoS评估模型。以解决多云环境的虚拟机QoS评估问题,为云服务用户提供可参考的量化依据。本文设计了对比实验和定性分析,对虚拟机QoS进行评估,并计算出QoS计算值与标注值的差距,然后对虚拟机属性标注与计算而得的QoS值进行相关性分析,以证明对虚拟机属性评估与对虚拟机QoS可量化指标评估的相关性。更多还原

【Abstract】 At present,the rapid development of cloud computing and virtualization and the wide use of hybrid cloud architecture have brought new carriers of data and business to the government and enterprises,and provided a solid foundation for their information construction,and promoted the process of cloud in the data and business of government and enterprises.Government and enterprises have diverse and differentiated needs in cloud environment application,and cloud environment becomes more and more complex.The construction and deployment of private cloud and public cloud also bring the widespread use of hybrid cloud architecture,while hybrid cloud also brings the problems faced by heterogeneous multi-cloud when meeting the different needs of builders.Virtual machine,as the service unit and basic carrier of cloud service in a multi-cloud environment,directly determines the cloud service quality used and perceived by users.Therefore,this paper chooses virtual machine service quality as the research object to evaluate virtual machine QoS,so as to judge the quality of cloud service used by users.In this paper,the influencing factors of VM QoS are selected and divided into VM configuration indicators that are easy to quantify and VM attributes that are not easy to quantify.User when choosing to buy cloud,a cloud service provider may occur due to budget performance causes too much waste or purchase according to the experience and subjective judgment,etc.,aiming at this problem,this paper proposes a user-oriented cloudy environment evaluation model of virtual machine QoS,nonlinear evaluation by gradient descent algorithm,the virtual machine performance assessment of the quantitative indicators,at the same time,The hard to quantify attributes of virtual machines are evaluated,and the correlation between the hard to quantify attributes of virtual machines and the evaluation of quantitative indicators is verified through experiments,and the two are integrated into the virtual machine QoS evaluation model in a multi-cloud environment.In order to solve the problem of VM QoS evaluation in a multi-cloud environment,provide cloud service users with a reference quantitative basis.In this paper,comparative experiments and qualitative analysis are designed to evaluate the virtual machine QoS,and the gap between the calculated value and the marked value of QoS is calculated.Then,the correlation analysis between the virtual machine attribute annotation and the calculated QoS value is conducted to prove the correlation between the virtual machine attribute evaluation and the virtual machine QoS quantifiable index evaluation.更多还原