Workflow Scheduling in the Cloud-Edge Continuum

Enhancing Workflow Efficiency: A Modified Firefly Algorithm for Hybrid Cloud-Edge Environments

“…However, heuristic algorithms may struggle to handle complex dependencies or dynamic workloads, limiting their scalability and effectiveness in large-scale or heterogeneous computing environments. On the other hand, metaheuristic algorithms offer flexibility and adaptability to diverse optimization objectives and constraints 21 . They excel at handling large-scale and combinatorial optimization problems, providing robustness and scalability for complex workflows and distributed computing environments.…”

“…Moreover, scientific workflows are inherently dynamic, with fluctuating resource demands throughout execution, necessitating adaptive scheduling strategies for optimal performance. Balancing competing objectives, such as minimizing makespan, optimizing resource utilization across both cloud and edge, and managing data transfer expenses, presents a formidable challenge that demands a sophisticated scheduling approach 19,20 .Traditional scheduling algorithms, such as Shortest Job First (SJF) or Heterogeneous Earliest Finish Time (HEFT), may fall short in hybrid environments, lacking the adaptability to address the complexities of resource management and data movement inherent to cloud-edge settings 21,22 . Similarly, conventional metaheuristic algorithms like Genetic Algorithm and Particle Swarm Optimization, while offering some degree of adaptability, may not be expressly tailored to tackle the intricate challenges of resource allocation and data transfer optimization within hybrid cloud-edge environments 23 .…”

“…Traditional scheduling algorithms, such as Shortest Job First (SJF) or Heterogeneous Earliest Finish Time (HEFT), may fall short in hybrid environments, lacking the adaptability to address the complexities of resource management and data movement inherent to cloud-edge settings 21,22 . Similarly, conventional metaheuristic algorithms like Genetic Algorithm and Particle Swarm Optimization, while offering some degree of adaptability, may not be expressly tailored to tackle the intricate challenges of resource allocation and data transfer optimization within hybrid cloud-edge environments 23 .…”