Cloudlet Scheduling Algorithm in Cloud Computing Constrained by Client I/O Port

Cloud computing has rapidly gained widespread use over decades according to the exponential increasing in data use. The cloud becomes an important platform provided self-service, on demand, elastic and scalable resources and services to the clients over a network with unlimited computing capacities. To deal with huge number of users and task processing requests, scheduling in cloud is crucial for the task execution. Current scheduling methods are mostly implemented to improve total completion time and cloud resource utilization. However they do not consider the task or cloudlet scheduling when the collision of sending and receiving tasks possibly occur at the client I/O port. Since the client is physically limited by the its chip set architecture, the situation that all tasks are assigned to the cloud resources simultaneously is insufficient for measuring the scheduling performance. This paper concerns the constraint of client I/O port therefore it proposes cloudlet scheduling method to achieve shortest makespan and less waiting time under the cloud environment. The proposed algorithm and other comparative algorithms are implemented in the CloudSim which is the framework simulating the cloud-based models and services. The performance evaluation shows that the proposed method outperforms other algorithms in terms of minimum makespan and less delay.