A dynamic weight–assignment load balancing approach for workflow scheduling in edge-cloud computing using ameliorated moth flame and rock hyrax optimization algorithms
Mustafa Ibrahim Khaleel
DOI: https://doi.org/10.1016/j.future.2024.02.025
IF: 7.307
2024-02-01
Future Generation Computer Systems
Abstract:As the geographically distributed cloud infrastructure continues to grow in scale and the intricacy of workflow applications increases, there is a growing threat to the operational efficiency of the system. This poses the risk of resource inefficiencies and elevated energy consumption. Load balancing becomes a critical concern for this category of applications, particularly when they encounter overwhelming demand surges and resource outages. Numerous comprehensive methods still overlook certain aspects, such as the inclusion of multiple objectives, addressing local minimum problems, and optimizing resource utilization to its fullest potential. Our research focus is on exploring cloud-deployed three-tier applications to address these challenges. This paper presents a new approach to load balancing, introducing a novel weight assignment algorithm. The algorithm effectively distributes tasks among virtual machines using three carefully optimized algorithms. To begin with, we convert the user tasks into the directed acyclic graph (DAG) structured workflow, organizing them according to their execution sequence. The iterative critical path search algorithm, implemented with a layer-based priority assigning technique, addresses the critical path challenge. This process involves finding a critical path, dividing tasks into distinct layers, classifying them based on their computational complexity and priorities, and distinguishing between critical and non-critical tasks. Following the classification process, the tasks are organized and prioritized using an improved Moth Flame Optimization approach based on multiple criteria, which consider criticality, priority, and makespan. This approach significantly reduces scheduling latency and delays. Finally, task allocation is proposed using the high convergence rock hyrax optimization algorithm with a divide-and-conquer strategy, enabling effective exploration and exploitation of overcrowded resources for improved load balancing. The CloudSim simulation tool is used to carry out the simulation, where it assesses various performance aspects such as latency, bandwidth, completion time, throughput, energy usage, instances of not meeting QoS violations, and overhead.
computer science, theory & methods