Network Aware Priority Based Task Scheduling Optimization in Cloud Computing

Abstract

Cloud computing offers scalable, flexible computational resources for today’s distributed systems, kind of like it can stretch or shrink as needed, not always but usually. In practice, task scheduling really has to be handled well because it affects resource utilization, it lowers the waiting period, and it can improve overall cloud performance. Most older scheduling algorithms tend to emphasize execution time only, while they ignore communication latency, and also the task’s priority which matters a lot sometimes. This paper proposes a lightweight simulation-based scheduling framework called NetworkAwareScheduler that improves Virtual Machine allocation using network-aware and priority-based optimization techniques. The proposed scheduler evaluates completion time, network latency, and task priority during task allocation. The system was implemented using Java and tested under different scheduling scenarios including low-latency, high-latency, and mixed-latency environments. Experimental results demonstrated significant improvements in throughput, waiting time reduction, makespan optimization, VM workload balancing, and communication-aware scheduling efficiency compared to traditional scheduling approaches.

Introduction

This paper presents a NetworkAwareScheduler, a lightweight cloud task scheduling framework designed to improve resource allocation efficiency in distributed cloud computing environments. Cloud computing platforms process large numbers of user tasks across heterogeneous Virtual Machines (VMs), making effective scheduling essential for maximizing resource utilization, throughput, and system performance. Traditional scheduling algorithms such as FCFS, Round Robin, Min-Min, and Max-Min primarily focus on processor allocation and execution order, often neglecting communication overhead, network latency, workload balancing, and task priority, which are critical factors in modern distributed cloud infrastructures.

The study highlights that inefficient task allocation can lead to overloaded VMs, increased waiting times, communication bottlenecks, reduced throughput, and poor resource utilization. While advanced optimization techniques such as Genetic Algorithms (GA), Particle Swarm Optimization (PSO), and Ant Colony Optimization (ACO) have improved scheduling performance, they often introduce higher computational complexity and implementation overhead.

To address these limitations, the proposed NetworkAwareScheduler incorporates network-aware and priority-based scheduling optimization into a lightweight simulation framework. The scheduler dynamically evaluates available Virtual Machines using a weighted cost function that considers completion time, communication latency, and task priority. By integrating these factors into scheduling decisions, the framework aims to improve execution efficiency while reducing communication delays and workload imbalance.

The simulation environment uses dynamically generated tasks and heterogeneous Virtual Machines instead of real cloud infrastructure. Tasks are characterized by attributes such as task length, priority level, and network latency, while VMs are defined by processing capability (MIPS), current workload, and communication latency. Multiple workload scenarios—including low-latency, high-latency, mixed-latency, and priority-heavy environments—are created to evaluate scheduler performance under different operating conditions.

The proposed scheduling algorithm follows a structured process. Tasks are first sorted according to priority, ensuring that high-priority tasks receive earlier scheduling opportunities. For each task, the scheduler computes execution time, completion time, and total communication latency for every available VM. A weighted scheduling cost is then calculated using:

Cost=0.60(Completion Time)+0.25(Network Latency)−0.15(Priority)\text{Cost} = 0.60(\text{Completion Time}) + 0.25(\text{Network Latency}) - 0.15(\text{Priority})Cost=0.60(Completion Time)+0.25(Network Latency)−0.15(Priority)

where completion time receives the highest weight due to its impact on execution efficiency, network latency accounts for communication overhead, and task priority reduces the overall cost for important tasks. The VM with the lowest calculated cost is selected, and its workload is updated dynamically to maintain balanced resource utilization.

The implementation is developed in Java using a modular architecture consisting of components for task management, VM management, scenario generation, scheduling algorithms, performance evaluation, and CSV-based result export. The framework compares the proposed NetworkAwareScheduler with a baseline HLGO Scheduler and measures key performance metrics such as makespan, throughput, waiting time, and workload distribution.

Conclusion

This research work presented a lightweight simulation-based scheduling framework titled \"Network Aware Priority Based Task Scheduling Optimization in Cloud Computing.\" The proposed framework introduced a custom scheduling algorithm called NetworkAwareScheduler, which improves task allocation efficiency by considering communication latency, task priority, completion time, and Virtual Machine workload during scheduling decisions. Unlike traditional cloud scheduling algorithms that mainly focus on execution order and CPU allocation, the proposed scheduler integrates communication-aware optimization directly into the scheduling process through a weighted cost function. The implementation was carried out using Java programming language and developed as a modular object-oriented simulation framework. The project successfully simulated multiple cloud scheduling scenarios including low-latency, high-latency, and mixed-latency environments. Experimental evaluation was conducted using important scheduling metrics including makespan, throughput, average waiting time, priority weighted waiting time, and VM workload balancing. The experimental results demonstrated that the proposed NetworkAwareScheduler achieved a 26.8% reduction in makespan, 43.0% reduction in average waiting time, 36.7% improvement in throughput, and 47.6% reduction in priority weighted waiting time compared to the HLGO Scheduler. The testing and validation phase confirmed that communication-aware optimization plays an important role in distributed cloud computing environments where network latency significantly affects overall execution efficiency. The modular structure of the framework simplifies future integration with advanced cloud platforms such as CloudSim, Kubernetes, Docker, Amazon Web Services (AWS), and Microsoft Azure. Future enhancements include: (1) cloud deployment using Docker and Kubernetes for scalable scheduling environments; (2) integration with CloudSim, AWS, and Microsoft Azure for real cloud experimentation; (3) machine learning-based scheduling using Reinforcement Learning and Genetic Algorithms; (4) dynamic VM scaling and adaptive workload balancing; (5) energy-aware and fault-tolerant scheduling optimization; (6) integration with IoT and edge computing infrastructures for latency-aware scheduling; (7) real-time monitoring dashboards and automated workload prediction systems.

References

[1] R. Buyya, C. S. Yeo, and S. Venugopal, \"Market-oriented cloud computing: Vision, hype, and reality for delivering IT services as computing utilities,\" in Proc. 10th IEEE Int. Conf. HPCC, 2008, pp. 5–13. [2] R. N. Calheiros, R. Ranjan, A. Beloglazov, C. A. De Rose, and R. Buyya, \"CloudSim: A toolkit for modeling and simulation of cloud computing environments,\" Software: Practice and Experience, vol. 41, no. 1, pp. 23–50, 2011. [3] A. Beloglazov and R. Buyya, \"Energy efficient allocation of virtual machines in cloud data centers,\" in Proc. 10th IEEE/ACM CCGrid, 2010, pp. 577–578. [4] M. Armbrust et al., \"A view of cloud computing,\" Commun. ACM, vol. 53, no. 4, pp. 50–58, 2010. [5] P. Mell and T. Grance, \"The NIST definition of cloud computing,\" NIST Tech. Rep. 800-145, 2011. [6] T. Erl, R. Puttini, and Z. Mahmood, Cloud Computing: Concepts, Technology and Architecture. Prentice Hall, 2013. [7] J. Dean and S. Ghemawat, \"MapReduce: Simplified data processing on large clusters,\" Commun. ACM, vol. 51, no. 1, pp. 107–113, 2008. [8] I. Foster et al., \"Cloud computing and grid computing 360-degree compared,\" in Proc. Grid Computing Environments Workshop, 2008, pp. 1–10. [9] H. Topcuoglu, S. Hariri, and M. Wu, \"Performance-effective and low-complexity task scheduling for heterogeneous computing,\" IEEE Trans. Parallel Distrib. Syst., vol. 13, no. 3, pp. 260–274, 2002. [10] M. Maheswaran et al., \"Dynamic matching and scheduling of independent tasks onto heterogeneous computing systems,\" in Proc. 8th HCW, 1999, pp. 30–44. [11] F. Dong and S. G. Akl, \"Scheduling algorithms for grid computing: State of the art and open problems,\" Queen\'s Univ., Tech. Rep. 2006-504, 2006. [12] S. Pandey et al., \"A particle swarm optimization-based heuristic for scheduling workflow applications in cloud computing,\" in Proc. 24th IEEE AINA, 2010, pp. 400–407. [13] K. Etminani and M. Naghibzadeh, \"A min-min max-min selective algorithm for grid task scheduling,\" in Proc. IEEE Internet Conf., 2007, pp. 1–7. [14] M. Kalra and S. Singh, \"A review of metaheuristic scheduling techniques in cloud computing,\" Egyptian Informatics J., vol. 16, no. 3, pp. 275–295, 2015. [15] X. Li, Y. Mao, and M. Liu, \"Task scheduling optimization in cloud computing based on heuristic algorithms,\" Journal of Networks, vol. 8, no. 3, pp. 547–553, 2013. [16] A. Verma and S. Kaushal, \"Cost-time efficient scheduling plan for executing workflows in cloud environment,\" J. Grid Computing, vol. 13, no. 4, pp. 495–506, 2015. [17] S. Kumar and A. Verma, \"Independent task scheduling in cloud computing by improved genetic algorithm,\" Int. J. Adv. Res. Comput. Sci., vol. 8, no. 5, pp. 1200–1205, 2017. [18] C. Vecchiola, X. Chu, and R. Buyya, \"Aneka: A software platform for .NET-based cloud computing,\" High Speed and Large Scale Scientific Computing, vol. 18, pp. 267–295, 2009. [19] Amazon Web Services, \"AWS Cloud Computing Services,\" 2024. [Online]. Available: https://aws.amazon.com/ [20] Kubernetes Authors, \"Kubernetes Documentation,\" 2024. [Online]. Available: https://kubernetes.io/docs/ [21] Microsoft Corporation, \"Microsoft Azure Documentation,\" 2024. [Online]. Available: https://learn.microsoft.com/azure/ [22] Google Cloud, \"Google Cloud Platform Documentation,\" 2024. [Online]. Available: https://cloud.google.com/ [23] Oracle Corporation, \"Java Platform Standard Edition Documentation,\" 2024. [Online]. Available: https://docs.oracle.com/javase/ [24] JetBrains, \"IntelliJ IDEA Documentation,\" 2024. [Online]. Available: https://www.jetbrains.com/idea/ [25] G. Singh and V. Bawa, \"A review on scheduling algorithms in cloud computing,\" Int. J. Computer Applications, vol. 95, no. 14, pp. 7–11, 2014.

Copyright

Copyright © 2026 Kunal Tiwari, Prashanth K. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.