•  
  •  
 

Abstract

The Big data originates from various heterogeneous sources worldwide. Due to the diverse origins of data generation, big data involves a variety of tasks. Some require storage resources, while others need computational resources. Additionally, cloud computing provides centralized storage and computational resources for executing these tasks. However, resource optimization and efficient task scheduling remain challenging. Moreover, accurately categorizing tasks and effectively allocating resources based on their nature pose challenges for cloud computing, especially when assigning tasks to heterogeneous resources to minimize task completion time. To address these issues, this paper proposes a Performance-Aware Task Classifier (PATC) algorithm that classifies tasks and allocates resources accordingly while optimizing QoS parameters. The task classification algorithm divides tasks into two categories: data-intensive and compute-intensive. Threshold coefficients for transmission time and CPU usage are used for task classification. Furthermore, resources are allocated to compute-intensive tasks to minimize their execution time as much as possible. To evaluate system performance, various QoS parameters such as Average Response Time, Total Completion Time, Makespan, and Average Resource Utilization are analyzed. Simulation results show that our approach effectively improves these parameters compared to existing methods such as First Come First Serve (FCFS), Shortest Job First (SJF), Modified SJF, Task Scheduling Strategy (TSS), and IMFO (Improved Moth Flame Optimization).

Keywords

Big data, Cloud computing, Intensive tasks, QoS, Resource optimization, Task categorization

Subject Area

Computer Science

Article Type

Article

First Page

1363

Last Page

1377

Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

Download

Share

 
COinS