1. What is the primary aim of task scheduling methods?
The primary aim of task scheduling methods is to increase the acceleration of execution by allocating resources to workloads with different execution times. This ensures efficient resource utilization and balances the workload effectively. Task scheduling methods are classified into dynamic and static methodologies, with the goal of mapping resources to tasks effectively while maintaining a trade-off between resource utilization and user requirements. Cloud task scheduling is a complex optimization problem, and various optimization algorithms have been developed to address it. However, traditional metaheuristics-based algorithms often result in higher computational times and require exploring larger search regions. To overcome these challenges, novel techniques like DA-AGSA have been proposed to improve workflow scheduling in cloud environments.
read more
2. What hybrid algorithm optimizes cost and running time in cloud environments?
Qin [14] introduced a hybrid collaborative multiobjective fruit-fly optimization algorithm (HCMFOA) for optimizing the cost and running time in cloud environments. The HCMFOA utilized a reference point-based clustering technique for dividing a single swarm into multi-sub-swarms. It included two rules for initializing fruit-fly location and three neighborhood operations. The experimental outcomes demonstrated that the HCMFOA achieved better performance compared to prior state-of-the-art models.
read more
3. How are tasks executed in cloud computing environments?
Tasks in cloud computing environments are executed based on their dependencies and levels. Independent tasks can be executed simultaneously on different resources, while dependent tasks rely on the output data of previous tasks. In the provided example, tasks 3 and 4 are independent and executed subsequently on dissimilar resources. The execution of task 2 depends on task 1. The IaaS cloud provider offers resource heterogeneity with various Virtual Machines (VMs) having different configurations. The execution time of a task on a resource is calculated using specific equations, considering factors like processing capacity, generated data, and bandwidth between VMs. The data transferring rate between scheduled tasks on similar VMs is zero. The DA-AGSA technique is used to find the optimal schedule in workflows to minimize total execution time and cost.
read more
4. How does DA-AGSA technique compare with other optimization algorithms?
The DA-AGSA technique is compared with five metaheuristics optimization algorithms, namely FOA, Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO), Genetic Algorithm (GA), and DA-GSA. The effectiveness of the DA-AGSA technique is evaluated based on makespan, cost, and reliability. The makespan is determined as the latest finished time on all the VMs, while the cost is calculated by multiplying the task duration with the allocated VMs price for all tasks. The reliability is defined as the probability of task execution over the allocated processor successfully without errors. The experimental results show that the DA-AGSA technique outperforms the other algorithms in terms of makespan, cost, and reliability, making it a promising optimization technique for workflow simulations.
read more