A Study On The Qos-constraint Resource Allocation Technology In The Virtualized Environment

Data center as a provider of physical resources, it needs to ensure that performance while optimizing resource utilization and to reduce their operating costs. The current virtualization technology provides the opportunity that can separated a single physical server into multiple virtual servers independently managed. During using the virtualization technology, it has also brought a challenge: how to share of physical resources on-demand and how to manage virtual server resources to the least cost to meet service quality targets. While the physical server uses virtualization technology to meet applied load of the QoS constraints objectives, it also improve the system resource utilization, reduce the complexity of resource management, reduce the use of the number of physical servers and save resources, which has become an important topic of research .This paper studies the concept of virtualization technology, a typical implementation and future development trends, specially details about QoS control study of the status quo of Web services, has done a summary of the existing virtual environment resource allocation strategies and points out the virtualization technology has brought a new challenges to the allocation of server resources development. In addition, the article based on Xen virtualization platform has designed a virtual cluster resource management system program, analysing of the advantages of the virtualization technology used into the cluster system's resource allocation, in order to forecast the application resource requirements better during adaptive resource allocation constraints control process based QoS, using fuzzy logic control approach. Main contents:The virtual cluster Web services architecture with the core of the QoS binding targets control, combines with the basic building blocks theory of the virtualization technology and the feedback control system, proposes functional requirements of the virtual cluster resource management system based on Xen virtualization environment, completes a virtual cluster management system design which provides management, control and condition monitoring function. In the case the various components for the quantify control of QoS corresponds to which in feedback control system.This paper attempts to introduce the modern theory of intelligent control into the QoS constrained resource allocation controller design. Based on the more sophisticated fuzzy logic control theory in intelligent control, we have proposed the method to use fuzzy control theory to QoS constrained resource controller design.This design of resource allocation controller based on QoS-constrained also designes a rule base automatically created and updated algorithms, to achieve the purpose of self-learning system, so that they can better predict and control the next phase of the local demand for local resources. The algorithm by calculating the importance of the rule dynamic updates the rule base in the rules, addressed the disadvantages as the rules have been established continued, the growing memory demand.In this paper, the control experiments on the Xen virtual machine to complete the allocation of CPU resources on restricting Web server response time targets, to verify that this fuzzy logic control model, theory and algorithms in the resource allocation application is the feasibility. The experiment indicate that control model has a more good control effect, in response to speed, convergence and stability with a good performance. The control model and algorithm can be applied to the control of the allocation of other resources.