Task Scheduling Mechanisms for Fog Computing: A Systematic Surveyopen access
- Authors
- Hosseinzadeh, Mehdi; Azhir, Elham; Lansky, Jan; Mildeova, Stanislava; Ahmed, Omed Hassan; Malik, Mazhar Hussain; Khan, Faheem
- Issue Date
- May-2023
- Publisher
- IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
- Keywords
- Task analysis; Processor scheduling; Surveys; Edge computing; Dynamic scheduling; Systematics; Costs; Quality of service; Fog computing; cloud computing; task scheduling; methods; quality of service
- Citation
- IEEE ACCESS, v.11, pp.50994 - 51017
- Journal Title
- IEEE ACCESS
- Volume
- 11
- Start Page
- 50994
- End Page
- 51017
- URI
- https://scholarworks.bwise.kr/gachon/handle/2020.sw.gachon/88401
- DOI
- 10.1109/ACCESS.2023.3277826
- ISSN
- 2169-3536
- Abstract
- In the Internet of Things (IoT) ecosystem, some processing is done near data production sites at higher speeds without the need for high bandwidth by combining Fog Computing (FC) and cloud computing. Fog computing offers advantages for real-time systems that require high speed internet connectivity. Due to the limited resources of fog nodes, one of the most important challenges of FC is to meet dynamic needs in real-time. Therefore, one of the issues in the fog environment is the optimal assignment of tasks to fog nodes. An efficient scheduling algorithm should reduce various qualitative parameters such as cost and energy consumption, taking into account the heterogeneity of fog nodes and the commitment to perform tasks within their deadlines. This study provides a detailed taxonomy to gain a better understanding of the research issues and distinguishes important challenges in existing work. Therefore, a systematic overview of existing task scheduling techniques for cloud-fog environment, as well as their benefits and drawbacks, is presented in this article. Four main categories are introduced to study these techniques, including machine learning-based, heuristic-based, metaheuristic-based, and deterministic mechanisms. A number of papers are studied in each category. This survey also compares different task scheduling techniques in terms of execution time, resource utilization, delay, network bandwidth, energy consumption, execution deadline, response time, cost, uncertainty, and complexity. The outcomes revealed that 38% of the scheduling algorithms use metaheuristic-based mechanisms, 30% use heuristic-based, 23% use machine learning algorithms, and the other 9% use deterministic methods. The energy consumption is the most significant parameter addressed in most articles with a share of 19%. Finally, a number of important areas for improving the task scheduling methods in the FC in the future are presented.
- Files in This Item
- There are no files associated with this item.
- Appears in
Collections - IT융합대학 > 컴퓨터공학과 > 1. Journal Articles
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.