Load Balancing in Cloud Computing

Cloud computing is one of the top trending technologies which primarily focuses on the end user’s use cases. The service provider needs to provide services to many clients. These increasing number of requests from the clients are giving rise to the new inventions in the load scheduling algorithms. There are different scheduling algorithms which are already present in the cloud computing, and some of them includes the Shortest Job First (SJF), First Come First Serve (FCFS), Round Robin (RR) etc. Though there are different parameters to consider when load balancing in cloud computing, makespan (time difference between start time of first task and finish of last task on the same machine) and response time are the most important parameters. This research surveys different load balancing algorithms and aims to improve the SJF load balancing algorithm in cloud computing. In this project, a Modified Shortest Job First (MSJF) and Generalized Priority (GP) load scheduling algorithms are combined to reduce the makespan and optimize the resource utilization. Together, MSJF and GP sends the longest task having high MIPS (million instructions per second) requirements to the machine with a high processing power and the shortest task having low MIPS requirements to the machine with a low processing power. Hence, neither the task with the lowest MIPS requirements nor the task with the highest MIPS requirements needs to wait for a very long time for resource allocation. Every task gets fair priority. Results are shown for SJF, MSJF, and GP in order to compare the different number of tasks using cloud simulator.