Profit-aware scheduling in task-level for datacenter networks. (July 2017)

• Record Type: Journal Article
• Authors: Tao, Xiaoyi; Qi, Heng; Li, Wenxin; Li, Keqiu; Liu, Yang
• Is Part Of: Computers & electrical engineering. Volume 61(2017)
• Page Start: 327
• Page End: 338
• Publication Date: 2017-07
• Subjects: Data center networks -- Task scheduling -- Application performance -- Maximize profit; Computer engineering -- Periodicals; Electrical engineering -- Periodicals; Electrical engineering -- Data processing -- Periodicals; Ordinateurs -- Conception et construction -- Périodiques; Électrotechnique -- Périodiques; Électrotechnique -- Informatique -- Périodiques; Computer engineering; Electrical engineering; Electrical engineering -- Data processing; Periodicals; Electronic journals; 621.302854
• DOI: 10.1016/j.compeleceng.2016.07.013
• Languages: English
• ISSNs: 0045-7906

Abstract

Highlights: A task-level scheduling method for application performance guarantee in data center is proposed. The method relies on priori information on task and application requirement. Task requirement should be strictly satisfied for integrity and performance. Graphical abstract: Abstract: Applications perform massive and diverse tasks in data centers. Tasks completion condition seriously affects application performance. However, most existing flow-level or task-level scheduling methods treat flows in isolation, meanwhile, few works discuss the efficiency of task-level scheduling from the perspective of the task profit. In this paper, we introduce a profit-aware task-level scheduling scheme named PAT, whose target is to maximize the profit of completing tasks within their reasonable time. To this end, a maximizing profit optimization model is proposed on task-level, and an efficient approximate scheduling algorithm is presented. Furthermore, a situation of absent deadline information is discussed and an ePAT method is presented to solve this situation. Based on the proposed algorithm, we design and implement PAT and ePAT. Some comprehensive experiments are conducted to evaluate the performance of our methods. The experimental results show that our methods bring higher profit than other scheduling methods.