Amazon EC2 (Elastic Compute Cloud)
Scalable virtual machine compute capacity in the cloud powered by the high-performance AWS Nitro System.
Scalable virtual machine compute capacity in the cloud powered by the high-performance AWS Nitro System.
Coming from Google Cloud?
Both provide enterprise IaaS virtual machines with high-performance networking, custom machine families, and persistent block disks.
- Full root OS-level access to Linux and Windows Server environments.
- Decoupled network-attached persistent block storage (EBS gp3/io2 vs GCP Persistent Disks).
- Deep discount models for fault-tolerant batch workloads (AWS Spot Instances vs GCP Spot/Preemptible VMs).
- Automated multi-zone fleet scaling (AWS Auto Scaling Groups vs GCP Managed Instance Groups).
- Hypervisor: AWS uses custom Nitro hardware ASICs; GCP uses KVM with Andromeda network virtualization.
- Custom Sizing: GCP allows arbitrary custom vCPU/RAM ratios; AWS offers pre-configured instance sizing tiers.
- Instance Maintenance: GCP automatically performs Live Migration without VM reboot; AWS notifies for planned retirement windows.
- Access Management: GCP uses Cloud IAM + OS Login; AWS uses IAM Instance Profiles + SSM Session Manager (no port 22 needed).
What is Amazon EC2 (Elastic Compute Cloud)?
Amazon Elastic Compute Cloud (Amazon EC2) provides on-demand, resizable compute capacity in the cloud. Built on the modular AWS Nitro System, EC2 offloads hypervisor, storage, and networking tasks onto dedicated hardware cards, delivering near 100% host resources to virtual machines with minimal virtualization jitter.
Decision Guide: When to Choose This Service
Use Amazon EC2 when you require full OS-level control (Linux/Windows), custom kernel modules, bare-metal access, legacy software migrations, or specialized hardware (Graviton4, NVIDIA H100 GPUs).
Do NOT use EC2 for simple containerized web microservices (use AWS Fargate or App Runner) or short-lived event-driven tasks (use AWS Lambda) where managing VMs creates unnecessary operational toil.
Core Architectural Pillars
Dedicated ASIC cards offloading hypervisor, EBS storage, and VPC networking for near-zero host overhead.
Diverse families including General Purpose (M7g), Compute (C7g), Memory (R7g), Storage (I4i), and GPU (P5).
Multi-AZ fleet management with Target Tracking scaling and automated instance healing.
Surplus compute capacity available at up to 90% discount with 2-minute interruption warnings.
Zero-SSH browser and CLI access without opening port 22 or managing keypairs.