Ontology framework for multi-core based cloud computing

Abstract

Cloud Computing is the most recent trend in computing today. Though it is considered being a “new” technology, but the concept was given by John MacCarthy in 1960, that states “computation may organize as Public Utility someday”. Cloud computing facilitates the users what they need, as they need it. Business applications to IT infrastructure including virtual servers and storage are all parts of cloud service model. Cloud systems cannot be just considered as another form of resource provisioning infrastructure but infact it provides multiple opportunities to host and also offers various kind of cloud applications. With the help of cloud computing the development and provisioning time of different services reduces to minimum. Despite numerous benefits, cloud usability is hindered by Interoperability of different cloud environments across the globe. The problem of interoperability and standardization between different cloud providers is responsible for issue like “vendor-locking”. It becomes difficult for users to switch between cloud service providers. Another factor impeding cloud computing is single-core processors. Many of the cloud providers have already altered their services from single-core to multi-core processors. Though multi-core processors in cloud computing comes with its own flaws; yet one of the major flaw is resource scheduling. Resources in cloud computing are limited and hence underutilization of these resources is not acceptable in any business organization. In this thesis, existing resource scheduling methodologies for multi-core processors have been compared. Different cloud interoperability standards have also been discussed and a framework has been designed and implemented which addresses the problems of interoperability and multi-core resource scheduling. To address interoperability, ontology has been used for representing cloud platform. The experimental results gathered through cloudsim. The results clearly demonstrate that the proposed framework has better performance as compared to the FCFS (First Come First Serve) approach.