A two-stage scheduling method for deadline-constrained task in cloud computing

In a cloud environment, reducing energy consumption while ensuring diverse quality of service (QoS) guarantees is challenging for task schedulers. Specifically, the energy-efficient scheduling for real-time tasks is more complicated because such tasks have strict time constraints. In this paper, we propose a two-stage scheduling method for deadline-constrained tasks. In the first stage, Enhanced Ant Colony Optimization (EACO) is a global scheduler that allocates incoming cloud tasks to suitable virtual machines (VMs). It can minimize makespan and energy consumption while guaranteeing strict deadline constraints. In the second stage, the Modified Backfilling (MBF) algorithm reorders VM's waiting queue to improve the task completion rate. We conduct two experiment series on synthetic and real trace datasets using the Cloudsim toolkit. Extensive experiments show that compared with other well-known task scheduling methods, our method can effectively reduce makespan by 25.28% and energy consumption by 23% on average. The task completion rate can be increased by 6.27%. The proposed method has a significant improvement compared with other well-known algorithms.

Automated summary

This paper proposes a method for task scheduling in a cloud environment, aiming to reduce energy consumption while ensuring quality of service. The method achieves efficient handling of deadline-constrained tasks through a two-stage scheduling approach, reducing completion time and energy consumption while improving task completion rate. Experimental results show significant improvement in task scheduling performance compared to other methods.