Compare compute services across AWS, Azure, and GCP including virtual machines, serverless functions, pricing, performance, and use cases.
Recommendation: AWS is recommended for enterprises requiring the broadest compute selection and deepest service integration. Azure excels for Microsoft-centric organizations leveraging Hybrid Benefit. GCP is ideal for Kubernetes-native workloads and scenarios requiring custom machine types or automatic sustained use discounts.
| Category | AWS | Azure | GCP |
|---|---|---|---|
| VM Instance Types Number and variety of virtual machine configurations | 600+ instance types across 8 families | 500+ VM sizes across 6 families | 40+ predefined types plus custom machine types |
| Serverless Runtime Languages Programming languages supported by serverless functions | Node.js, Python, Java, Go, .NET, Ruby | C#, JavaScript, Python, Java, PowerShell | Node.js, Python, Go, Java, .NET, Ruby, PHP |
| Max Serverless Execution Time Maximum duration for a single serverless function invocation | 15 minutes | 10 minutes (Consumption), unlimited (Premium) | 60 minutes (2nd gen) |
| Custom Machine Types Ability to define custom CPU and memory combinations | ✗ No | ✗ No | Full custom CPU/memory combinations |
| ARM-Based Processors Support for ARM architecture for cost-effective compute | Graviton2/Graviton3 (up to 40% better price-performance) | Ampere Altra (Dpsv5/Epsv5 series) | Tau T2A (Ampere Altra) |
| Serverless Containers Managed serverless container execution | AWS Fargate | Azure Container Instances | Cloud Run |
| Auto Scaling Automatic scaling of compute resources based on demand | EC2 Auto Scaling, Application Auto Scaling | VM Scale Sets, Azure Autoscale | Managed Instance Groups, Cloud Run auto scaling |
| Spot/Preemptible Instances Discounted instances for fault-tolerant workloads | Spot Instances (up to 90% discount) | Spot VMs (up to 90% discount) | Spot VMs (up to 91% discount) |
| Live VM Migration Ability to migrate running VMs without downtime during maintenance | ✗ No | Limited (maintenance events) | Yes (transparent, automatic) |
| Managed Kubernetes Fully managed Kubernetes container orchestration | Amazon EKS | Azure Kubernetes Service (AKS, free control plane) | Google Kubernetes Engine (GKE, most mature) |
Pricing: On-demand, Reserved Instances (1yr/3yr), Savings Plans, Spot Instances. EC2 starts at $0.0058/hr (t4g.nano). Lambda at $0.20 per 1M requests plus compute time.
Performance: Broadest instance selection for workload-specific optimization. Graviton instances deliver leading price-performance for general workloads. Lambda cold starts typically 100-500ms.
Pricing: Pay-as-you-go, Reserved Instances (1yr/3yr), Savings Plans, Spot VMs, Azure Hybrid Benefit. VMs start at $0.004/hr (B1ls). Functions free tier: 1M executions/month.
Performance: Strong performance for Microsoft workloads. Hybrid Benefit provides significant savings for Windows/SQL Server. Functions Premium plan eliminates cold starts.
Pricing: On-demand, Committed Use Discounts (1yr/3yr), Sustained Use Discounts (automatic), Spot VMs. Compute Engine starts at $0.006/hr (e2-micro). Cloud Functions free tier: 2M invocations/month.
Performance: Custom machine types eliminate over-provisioning. Sustained use discounts apply automatically. Cloud Functions 2nd gen built on Cloud Run delivers faster cold starts and longer execution.
AWS offers the broadest range of compute options with EC2 for virtual machines and Lambda for serverless functions.
Amazon EC2 provides resizable compute capacity with 600+ instance types across general purpose, compute optimized, memory optimized, storage optimized, and accelerated computing families. Supports Graviton ARM-based processors for cost-effective performance.
AWS Lambda runs code without provisioning servers, supporting Node.js, Python, Java, Go, .NET, and Ruby. Scales automatically from zero to thousands of concurrent executions with sub-second billing granularity.
ECS and EKS provide managed container orchestration. Fargate offers serverless containers eliminating the need to manage underlying infrastructure.