DB-ACO: A Deadline-Budget Constrained Ant Colony Optimization for Workflow Scheduling in Clouds
Siyuan Tao,Yuanqing Xia,Lingjuan Ye,Ce Yan,Runze Gao
DOI: https://doi.org/10.1109/tase.2023.3247973
IF: 6.636
2024-01-01
IEEE Transactions on Automation Science and Engineering
Abstract:With the development of cloud computing, a growing number of workflows are deployed in cloud platform that can dynamically provide cloud resources on demand for users. In clouds, one basic problem is how to schedule workflow under the deadline constraint and minimize the execution cost. As the capability of cloud resources getting higher, the required cost is also rising. Capability of some resources exceeds the need of users, which leads to higher cost, and the budget of users should be considered. In this paper, a novel scheduling algorithm, named DB-ACO, is proposed to minimize the execution cost for the workflow with deadline and budget constraints. DB-ACO is verified on four typical scientific workflows, and the experiments results show it outperforms four state-of-the-art methods, especially for CyberShake.Note to Practitioners-Budget and deadline are important requirements for users in cloud computing, which are used as constraints. Extensive works have been devoted to minimize the cost of workflows execution with different scheduling strategies. However, most of them only consider one single constraint and assume the constraint is simple and loose, which is impractical in actual scenarios due to higher requirement of users. This paper investigates a novel scheduling algorithm DB-ACO to optimize cost under budget and deadline. DB-ACO combines heuristic and meta-heuristic, it uses ant colony optimization to optimize the execution cost under the deadline and budget constraints: each ant sorts tasks on the basis of the combination of the pheromone trail and heuristic information, the deadline and budget are distributed fairly to each task by a novel distribution method, then the service selection rules are introduced to build solution.