Amazon Elastic Compute Cloud

Virtual servers you can freely use on AWS.

Service for launching virtual servers on AWS. You choose the OS, instance type, storage, and network.

• Building web servers
• Batch processing servers
• Migration target for existing applications

• Billed based on running time and instance type
• Combined with Auto Scaling and ELB
• Under the shared responsibility model, OS and above are the customer's responsibility

• Improve availability with a Multi-AZ configuration
• Designing stateless makes Auto Scaling easier
• Optimize cost by choosing instance purchasing options

Amazon EC2 (Elastic Compute Cloud) is an IaaS service that allows you to launch and manage virtual servers (instances) in the AWS cloud. By selecting an AMI (Amazon Machine Image), you can start an instance with the OS and initial software pre-configured. **Instance types** are selected according to the workload. They are divided into families such as general purpose (M / T series), compute optimized (C series), memory optimized (R / X series), storage optimized (I / D series), and accelerated computing (P / G series / GPU). CLF asks for the basic understanding of "selecting a combination of vCPU count and RAM". **Purchase options** are the key to cost optimization and are frequently asked about in both CLF-C02 and SAA-C03. - **On-Demand**: per-second billing, no commitment, suitable for short-term unpredictable workloads - **Reserved Instances (RI)**: up to 75% discount with a 1-year or 3-year commitment. Suited for steady-state workloads - **Savings Plans**: a newer mechanism that applies discount rates equivalent to RIs more flexibly (regardless of instance type) - **Spot Instances**: utilize AWS spare capacity at up to 90% discount. Subject to interruption risk. Suited for fault-tolerant distributed workloads (batch, ML training, etc.) - **Dedicated Host**: occupy a physical server. Supports BYOL (Bring Your Own License) **The combination of Auto Scaling and ELB** is the standard configuration for web applications. Auto Scaling increases or decreases the number of instances in response to traffic, while ELB (Elastic Load Balancing) distributes requests across each instance. This configuration requires a **stateless design** (no session information stored inside an instance), and the standard pattern is to externalize sessions to ElastiCache (Redis / Memcached) or DynamoDB. From the perspective of the **shared responsibility model**, EC2 places the management of the layers above the hypervisor (OS / middleware / applications) on the user (the "security in the cloud" domain). CLF-C02 asks questions such as "which service provides virtual servers?" and "which is an EC2 billing model?". SAA-C03 asks about how to choose among purchase options, the architecture design of Auto Scaling + ELB, AMI customization and horizontal deployment, and the use of instance metadata and user data.

• Amazon EC2 (AWS Official)
• What is Amazon EC2 (AWS Official Documentation)
• Amazon EC2 Instance Purchase Options (AWS Official Documentation)

Last updated: 2026-06-01