An Efficient Job Classification Technique to Enhance Scheduling in Cloud to Accelerate the Performance

Abstract

As cloud computing is becoming more ubiquitous with increasing espousal of advanced technologies, more and more efficient techniques are required to enhance the system performance. The computing systems in the cloud comprise heterogeneity of components or resources. Challenge here is efficient scheduling and resource allocation to jobs requesting the computing devices in order to achieve customer satisfaction. Retention of customer satisfaction is one of the primary factors for an organization to exist. To achieve the aforementioned goal, it is required to implement the principles of software engineering in every task that is accomplished in the organization. Thus, with organizations marching towards cloud environment, the jobs are initially clustered or grouped and subsequently scheduled for the resource arbitration and allocation. This paper focuses on clustering or grouping of jobs. The unsupervised technique or the clustering of the jobs is done based on the logistic regression approach. As this approach is more robust and the parameters considered for classification are more independent, simulation evidence suggests the classification technique cluster the jobs more effectively and provide a consistent utilization of the available resources. This ensures that the non functional requirement of availability of jobs to their customers is achieved thereby enhancing the business performance.