A Centralized Multi-Criteria Method for Scheduling Tasks in a Cloud Computing Environment

Abstract

Task scheduling determines the order of mapping
tasks to virtual machines to meet objectives. In this paper, a
batch mode heuristic method that is centralized, dynamic, and
multi-objective has been presented for scheduling independent
tasks with a deadline and belonging to several user levels, using
the cloud elasticity in the public cloud environment. In this
method, it has been intended to improve the objectives of
makespan, deadline violation, total execution cost, and load
balancing by considering the tasks’ prioritization based on the
criteria of user level, deadline, task length, and selection of
heterogeneous virtual machines according to processing power,
workload and usage cost. The proposed method was simulated
using the CloudSim tool. Besides, the method’s ability to achieve
the mentioned goals has been evaluated in comparison with
similar methods. The evaluation results, established on standard
test data, show that the proposed method has a good
performance in improving its objectives.