Adaptive scheduling mechanism in cloud

Abstract

Cloud user expects its service requests to be served as early as possible. While cloud service providers (CSP) tries to serve user requests within deadline and obtain more profit. These two factors (deadline and profit) are conflicting - if only deadline is considered profit earned may be lesser; if only profit is considered few low profit jobs may starve for resources. So it is a challenging task for CSP to serve user requests considering these factors. This paper proposes an Adaptive Scheduling Mechanism which prioritizes the jobs based on their profit and deadline, and then orders them in priority queue for execution. On peak loads as new jobs with high profit arrive, some of the existing jobs having lesser profit in a priority queue suffer from starvation. In such context, the proposed mechanism adapts to generate another queue, which holds relatively lesser profit jobs from priority queue, and uses FCFS scheduling approach. Then the scheduler schedules jobs alternatively from multilevel queue - choosing one from priority queue and another from FCFS queue. This scheme returns more profit to CSP and also avoids starvation of jobs. The proposed work is simulated using CloudSim and it is observed that it performs better compared to existing work.