Enhanced Particle Swarm Optimization for Task Offloading and Scheduling in Cloud-Fog Environment

Awad, Samar; Gamal, Marwa; Abd El Salam, Khalid; Abdel-Kader, Rehab F.

doi:10.21608/ijt.2024.308144.1060

	Enhanced Particle Swarm Optimization for Task Offloading and Scheduling in Cloud-Fog Environment
International Journal of Telecommunications
Volume 04, Issue 02, July 2024, Page 1-16 PDF (1.55 MB)
Document Type: Original Article
DOI: 10.21608/ijt.2024.308144.1060
View on SCiNiTO
Authors
Samar Awad ¹; Marwa Gamal ²; Khalid Abd El Salam ³; Rehab F. Abdel-Kader ⁴
¹Alfrosia street,Ismailia,Egypt
²Electrical Engineering Department, Computer and Control branch, Suez Canal University, Ismailia, Egypt.
³Electrical Engineering Department, Suez Canal University, Ismailia, Egypt. Department of Information System, College of Information Technology, Misr University for Science and Technology (MUST), 6th of October City 12566, Egypt.
⁴Electrical Engineering Department, Faculty of Engineering, Port Said University, Port Said, Egypt
Abstract
The explosion of devices and their varied uses in the Internet of Things (IoT) have created a massive quantity of data that requires significant processing power. Fog computing, as a prolongation of cloud computing, presents a promising new model by bringing pro-cessing power closer to users through fog servers. Compared to accessing distant cloud servers, this significantly reduces latency, or the required time for data to travel. This setup allows users to offload tasks to nearby servers, ultimately improving the Quality of Service (QoS) they experience. Finding the best match between workflow tasks and available resources is critical to minimizing completion time (makespan), especially in delay-sensitive applications requiring fast data processing. However, achieving this op-timal match remains a challenge. This work proposes an Enhanced Particle Swarm Op-timization (EPSO) algorithm specifically designed to address this challenge. The per-formance of EPSO is compared against PSO, Max-Min, and Round-Robin (RR) scheduling methods. Simulations are conducted using diverse scientific workflow domains. The results demonstrate that EPSO outperforms all other methods in minimizing makespan across all tested workflows. Furthermore, EPSO exhibits competitive performance in other metrics like energy consumption and cost while maintaining greater stability and reliability.
Keywords
Cloud-Fog Computing; Offloading; Scientific Workflow; PSO; Task Scheduling


Statistics Article View: 239 PDF Download: 169