A Novel Attempt towards Effective Scheduling Based on Reliability in Cloud Environment

Abstract

Cloud computing is now becoming a trend of modern IT industry. The huge advantages of cloud computing are progressively attracting individuals and several organizations to outsource their data from local to remote cloud servers. Therefore, the widespread cloud users assuredly expect reliable services and also desire an enhanced performance gain due to their adaptation of cloud. In order to accomplish as many user requests as possible, it is very much necessary to utilize cloud resources effectively and wisely. From the review of the new technologies, some crucial challenges for their development in future have been pointed out, among which the reliability of cloud resources has been emerged as a great concern to provide failure free services to the users. Any hardware and software component of cloud can fail at any point of time. Sufficient remedies can be taken to reduce the number of failures. Everyone wants a reliable service, but a dynamic system with high complexity like cloud; it is very difficult to manage the reliability. This paper introduces an effective reliable scheduling algorithm using an allocation matrix where virtual machines (VMs) are dynamically allocated to the user requests based on their reliability and requests' priority. Here, a dependent and serialized approach has been used for calculating the reliability of the cloud components. With this approach the overall reliability of cloud environment will be increased. Furthermore, the work has been simulated in CloudSim simulation environment. The simulation test of the proposed algorithm and an analysis of the non-reliable and reliable scheduling approach have been carried out.