Benchmarking Virtualized Compute Instances in Amazon's Cloud for High-Performance Applications
- Authors
-
-
Harika Naidu Beesabathuni
Author
-
- Keywords:
- Cloud Computing, Virtualization, Amazon EC2, Performance Evaluation, High-Performance Computing, Scientific Computing, Benchmarking
- Abstract
-
This paper presents a performance evaluation of virtualized compute instances within Amazon's Elastic Compute Cloud (EC2) infrastructure, with particular emphasis on their suitability for high-performance computing (HPC) applications. Cloud computing has emerged as a flexible and cost-effective alternative to traditional computing infrastructures, enabling on-demand access to virtualized computational resources. However, the performance implications of virtualization, especially for scientific and communication-intensive workloads, remain an important concern for researchers and practitioners. To address this issue, a comprehensive benchmarking methodology is employed to evaluate the computational, memory, storage, and parallel processing capabilities of Amazon EC2 instances. A range of benchmarking tools, including microbenchmarks and workload-oriented evaluations, are used to measure critical performance metrics such as resource provisioning time, computational throughput, memory hierarchy behavior, disk I/O performance, communication overhead, and application execution characteristics. The experimental results are compared with conventional high-performance computing systems to assess performance differences across workload categories and system configurations. Furthermore, the study investigates the cost-performance trade-offs associated with cloud-based execution and identifies limitations introduced by virtualization and shared infrastructure. The findings provide practical insights into the applicability of commodity cloud environments for scientific computing workloads and highlight scenarios in which cloud resources may complement or substitute traditional HPC systems.
- References
- Downloads
- Published
- 2026-09-25
- Issue
- Vol. 1 No. 5 (2026)
- Section
- Articles
- License
-
Copyright (c) 2026 International Journal of Intelligent Systems and Data Science

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