Hybrid optimization algorithm for task scheduling and virtual machine allocation in cloud computing
- PDF / 1,420,734 Bytes
- 8 Pages / 595.276 x 790.866 pts Page_size
- 39 Downloads / 310 Views
SPECIAL ISSUE
Hybrid optimization algorithm for task scheduling and virtual machine allocation in cloud computing G. Sreenivasulu1 · Ilango Paramasivam2 Received: 19 August 2020 / Revised: 17 September 2020 / Accepted: 21 October 2020 © Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract In the Internet era, cloud computing is evolved as the efficient distributed platform in the recent years. But the major issue related to the cloud platform is task scheduling. Allocating the suitable VM to the tasks is a challenging task in cloud computing. Many algorithms are proposed to optimize the scheduling process in the cloud environment. The existing algorithms have their own drawbacks. This paper proposed the hybrid model which uses the hierarchical process to prioritize the task before submitting to the scheduler. The Bandwidth-aware divisible task (BAT) scheduling model is modified by adding the Bar system model to develop the hybrid optimization mechanism. The Minimum overload and minimum lease policy is employed for applying the pre-emption in the data center to reduce the overload of the virtual machine. The performance of the proposed hybrid model is evaluated using different parameters. The simulation results prove the efficiency of the hybrid model. Keywords Scheduling · Tasks · Virtual machines · Cloud · Load · Optimization
1 Introduction In the recent years, cloud computing has evolved as one of the efficient computing platforms which provides the flexibility of the information technology to the field of medical, mobile system, business, smart system etc. [1, 2]. These advantages made the cloud computing as the emerging technology. Cloud computing offers different services like storage, platform, software, network, database, security and many more [3, 4]. These services are offered to the users in the form of virtualization resources. These virtualization resources are elastic and dynamic in nature. User applications are directly scheduled over the virtualization resources to execute the tasks. Task scheduling is a wide research area in the cloud computing. Many researchers developed different algorithms to solve the task scheduling problem, but * Ilango Paramasivam [email protected] G. Sreenivasulu [email protected] 1
Department of Computer Science, Bharathier University, Coimbatore, Tamilnadu, India
Department of CSE, PSG Institute of Technology and Applied Research, Coimbatore, Tamilnadu, India
2
still it remains as the NP-hard problem. Cloud environment contains the data centers which are distributed geographically [5]. Each data center is composed of thousands of servers and each server is divided in to number of virtual machines with the components of memory, CPU, storage. Group of VMs are assigned to the cloud users to execute the tasks. Scheduling the appropriate resources to the task is a complex issue. Here, we need to consider the properties of the tasks (for instance size, length, time of execution, dependency) before assigning to the virtual resources. There is
Data Loading...
