Cost-Performance Evaluation of General Compute Instances: AWS, Azure, GCP, and OCI
Jay Tharwani, Arnab A Purkayastha
TL;DR
This work tackles the challenge of selecting cost-effective general-purpose cloud compute resources across four major providers by standardizing to 4 vCPUs and 16 GiB RAM and benchmarking with Geekbench6. It combines performance data with provider pricing (on-demand and annual commitments) to derive practical cost-per-performance insights across Intel, AMD, and ARM architectures. The study finds ARM-based instances typically yield the best price-performance, AMD offers strong cost efficiency, and OCI often provides the lowest cost-per-performance across architectures, with Intel remaining strong for legacy software and enterprise versatility. The results offer actionable guidance for workload-specific cloud resource choices, enabling organizations to optimize both cost and performance in multi-cloud environments.
Abstract
Cloud computing has become the cornerstone of modern IT infrastructure, offering a wide range of general-purpose instances optimized for diverse workloads. This paper provides a comparative analysis of cost and performance for general-purpose compute instances across four major cloud providers: AWS, Azure, Google Cloud Platform (GCP), and Oracle Cloud Infrastructure (OCI). Using standardized configurations of 4 vCPUs and 16 GiB of RAM, the study evaluates instances based on processor architecture (Intel, AMD, ARM), pricing models, and performance benchmarks. Key findings reveal that ARM-based instances deliver superior price-performance ratios for cost-sensitive workloads, while Intel-based instances excel in enterprise-grade applications requiring versatility and reliability. The results aim to guide organizations in selecting the most cost-effective and performance-efficient cloud resources for their specific needs.