Abstract: A cloud desktop is a new type of cloud service that can provide customizable and shared resources. As a searchable cloud service, it can be configured to satisfy individual demands. This paper presents a cloud desktop recommendation method based on ontology, a description of user requirements, and matching from both resource providers and consumers. In one case, the service functions and non-functional features of a resource provider are analyzed, and ontologies are used to match user requirements with a candidate desktop image to find a reliable cloud desktop. In another case, user requirements for resource consumers are analyzed, and we utilize collaborative filtering methods to allow the user to select an appropriate cloud desktop resource, avoiding potential deviations in user resource requests. Experimental results show that the proposed method improves the accuracy of the cloud desktop in the resource discovery and selection process; moreover, the proposed method can recommend more suitable data resources according to the potential needs of users.

Key words: user requirements, cloud desktop, semantic ontology, non-functional property, service recommendation