云计算环境下资源监控方法的研究与实现

发布时间：2018-07-28 14:05

【摘要】：云计算(Cloud Computing)是以服务概念为主的新型计算方式,利用现有的强大网络、计算和存储资源提供计算、存储和平台服务,并具有良好的可扩充性和稳定性。云计算平台具有虚拟性、层次性以及动态性等特点,比传统的分布式计算(Distributed Computing)更加复杂。因此,资源监控是云计算平台的重要组成部分,其对提高云计算平台的服务质量发挥重要作用,研究云计算环境下监控方法具有重要的意义。目前云环境下的监控系统存在着如下问题：(1)监控对象有局限性,往往只关注云平台的某一层次的资源或服务,不能为系统较为提供全面的监控。(2)现有的一些监控框架不具有通用性,针对某些特定的云平台效果不错,但是不能适合用于其他云平台。(3)资源监控系统中的数据传输算法虽然采用了混合推拉算法,但是其效率还可以进步的提高,从而进一步减少对系统的影响。本文主要做了如下工作,来解决上述问题：本文对云环境下的资源监控进行了深入的研究,抽象了资源监控系统的模型,明确了建立面向服务的资源监控,详细讨论了监控服务的系统结构。其次,针对资源监控系统采集数据细节进行了研究,对云服务的层次和能耗做了分析,提出了较为详细的数据采集指标和监控代理的设计。然后,进一步深入研究了当前比较流行的混合推拉模式,指出了其中存在的问题,并对问题给出了解决方案。为了验证设计方案的效果,提出了四个对比指标,以此为依据进行算法的对比。通过分析多组实验的结果验证了改进方案的有效性。最后,研究了数据的处理,阐述了资源监控系统对数据处理的过程,在此基础上提出了数据处理组件的结构。
[Abstract]:Cloud computing (Cloud Computing) is a new computing method based on the concept of service. Using the existing powerful network, computing and storage resources provide computing, storage and platform services, and it has good scalability and stability. Cloud computing platform is more complex than traditional distributed computing (Distributed Computing) because of its virtual, hierarchical and dynamic characteristics. Therefore, resource monitoring is an important part of cloud computing platform, which plays an important role in improving the quality of service of cloud computing platform. At present, the following problems exist in the monitoring system in the cloud environment: (1) the monitoring objects have limitations, and they usually only pay attention to the resources or services at a certain level of the cloud platform. Can not provide more comprehensive monitoring for the system. (2) some of the existing monitoring framework is not universal, for some specific cloud platform effect is good, But it is not suitable for other cloud platforms. (3) although the hybrid push-pull algorithm is used in the resource monitoring system, its efficiency can be improved, thus further reducing the impact on the system. This paper mainly does the following work to solve the above problems: this paper has carried on the thorough research to the resources monitoring under the cloud environment, has abstracted the resources monitoring system model, has clearly established the service-oriented resources monitoring, The system structure of monitoring service is discussed in detail. Secondly, the details of data collection in resource monitoring system are studied, the level of cloud service and energy consumption are analyzed, and the detailed data acquisition index and the design of monitoring agent are put forward. Then, the current popular hybrid push-pull model is further studied, the existing problems are pointed out, and the solutions to the problems are given. In order to verify the effect of the design, four contrasting indexes are put forward, based on which the algorithm is compared. The effectiveness of the improved scheme is verified by analyzing the results of multi-group experiments. Finally, the processing of data is studied, and the process of data processing in resource monitoring system is expounded. On the basis of this, the structure of data processing component is put forward.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09

【参考文献】