期刊
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
卷 17, 期 11, 页码 7820-7829出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TII.2020.3011506
关键词
Cyber-physical cloud systems (CPCS); soft-error and lifetime reliability; security; slack; workflow scheduling
类别
资金
- National Natural Science Foundation of China [61802185, 61872185]
- Natural Science Foundation of Jiangsu Province [BK20180470]
- Fundamental Research Funds for the Central Universities [30919011233, 30919011402]
- Open Research Fund of Engineering Research Center of Software/Hardware Co-Design Technology and Application (Ministry of Education) at East China Normal University
Cyber-physical cloud systems (CPCS) integrate cyber-physical systems (CPS) with cloud computing infrastructures to improve performance, while introducing new reliability and security challenges. This article proposes a dependable algorithm for scheduling workflow applications on CPCS, utilizing slack to recover failed tasks and dynamically adjusting task frequencies to enhance soft-error reliability. Extensive experiments show the scheme reduces failure probability by up to 52.1% and improves scheduling feasibility by up to 83.5% compared to other approaches.
Cyber-physical cloud systems (CPCS) are integrations of cyber-physical systems (CPS) and cloud computing infrastructures. Integrating CPS into cloud computing infrastructures could improve the performance in many aspects. However, new reliability and security challenges are also introduced. This fact highlights the need to develop novel methodologies to tackle these challenges in CPCS. To this end, this article is oriented toward enhancing the soft-error reliability of real-time workflows on CPCS while satisfying the lifetime reliability, security, and realtime constraints. In this article, we propose a dependable algorithm for scheduling workflow applications on CPCS. The proposed algorithm uses slack to recover failed tasks and allows all tasks to share the available slack in the system. To improve soft-error reliability, the algorithm first determines the priority of tasks, then assigns the maximum frequency to each task, and finally assigns the recoveries to tasks dynamically. Slack also can be used to utilize security services for satisfying system security requirements. The lifetime reliability constraint is met by dynamically scaling down the operating frequency of low-priority tasks. Extensive experiments on real-world workflow benchmarks demonstrate that the proposed scheme reduces the probability of failure by up to 52.1% and improves the scheduling feasibility by up to 83.5% compared to a number of representative approaches.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据