Research And Implementation Of User Requirements Oriented Fault Tolerant Method For Cloud Desktop

As users' demand for data security and privacy have been continuously improved,more and more enterprise are beginning to build their own private cloud platforms,with Cloud Desktop being the most typical application.So in order to enhance the user experience of the cloud desktop and provide accurate desktop applications,the cloud desktop as a searchable cloud service can be selected according to their individual demands.This thesis divides cloud desktop resources into two parts:infrastructure(IF)and software facilities(SF).On the one hand,from the perspective of the service provider,there are a large number of desktop images in the data center.When the user operates the desktop,abnormal system or application software may appear and then remain in the image before the system resets,which will directly affect the user's experience.On the other hand,from the perspective of the service consumer,users have a misunderstanding of the cloud desktop resources.If the amount of IF requests is small,it will affect the performance of SF and even lead to system anomalies;If the amount of IF requests is too large,the private cloud resources of the enterprise will be wasted,causing the cloud platform failing to respond to more users' SF application requirements.This thesis proposes a fault-tolerant cloud desktop approach based on user demand,the main tasks is:First of all,in terms of functional attribute,how to select the cloud desktop closest to the user's demand from the resource pool including a lot of similar desktop images.So in this thesis,based on ontology technology to match user requests,the retrieval result is more accurate than the traditional keyword query.Based on the similarity between the concepts of ontology,this thesis presents a fault-tolerant method that meets the requirements:(1)At the service provider side,in order to ensure the quality of desktop resource services,this thesis provides reliable desktop applications by representing and evaluat ing non-functional attributes related to the cloud desktop;(2)At the service consumer end,users have uncertainties about their own applicat ion requirements.Sothis thesis recommends appropriate resources to them by analyzin g the usage evaluation of similar user requirements,to avoid the deviation of user reso urce requests.Finally,a cloud desktop retrieval prototype system based on user demand is designed and implemented.Experimental results show that the ontology-based user demand description and matching method can improve the fault tolerance of the cloud desktop when it is used and an exception occurs.