ANALYSIS OF METHODS OF DISTRIBUTION OF RESOURCES IN THE VIRTUALIZATION MEDIA

Abstract

Existing cloud computing environments approach to initial allocation and further resources distribution can be divided into 3 types: manual assignment of resources, resource planning by the cloud computing environment manager and resource planning by the virtualization environment. Cloud computing environments have features that create the specifics of resource allocation in such environment. Some elements of this issue are common to virtualization problems, and some have features specific to cloud computing environments. In virtualization environments, the problem of planning and optimizing the resource usage is solved in two ways: resource redistribution by the virtualization manager manually; automated resource planning using Distributed Resource Scheduling (DRS) and Distributed Power Management (DPM). In various virtualization environments, DRS and DPM work differently, but the original logic of work is kept. The subject of this article is to study the resource distribution in virtualization environments and in cloud computing environments. The goal is to evaluate existing resource distribution methods in virtual environments, to identify their compliance with the mandatory characteristics of cloud computing. The task: to analyze existing resource distribution methods in virtualization environments and cloud computing environments, to determine their advantages and disadvantages. According to the results of the comparative analysis, it can be concluded that the methods most often used in cloud computing environments are very primitive, and the problem of resource allocation in a cloud environment can be more effectively solved by the same methods, what are used in virtualization environments. Conclusions. The disadvantage of each of the methods analyzed is that the needs of applications that run inside instances are counted only in the context of the required amount of processor resources, RAM, and free disk space. This does not take into account the specificity of the applications, and how this application will split the resources with already deployed on the same instance applications. Therefore, it does not always take place the selection of optimal host / storage resource for placing an instance, which leads to a significant reduction in application performance and the effectiveness of using the "cloud" resource.