General Purpose
Use cases: Web or application servers, enterprise applications, gaming servers, development or test environments
Instance types: M7, Mac, M6, M5, M4, T4, T3, T2
Provide a balance of compute, memory, and networking resources
Selecting Instance Type
Selecting an Amazon EC2 Instance Type
Section titled “Selecting an Amazon EC2 Instance Type”EC2 Instance Type Configuration
Section titled “EC2 Instance Type Configuration”An EC2 instance type defines the configuration of CPU, memory, storage, and network performance:
| Instance type | vCPU | Memory | Storage | Network performance |
|---|---|---|---|---|
| m5d.large | 2 | 4 GiB | 1 x 50 NVMe SSD | Up to 10 Gbps |
| m5d.xlarge | 4 | 8 GiB | 1 x 100 NVMe SSD | Up to 10 Gbps |
| m5d.8xlarge | 32 | 128 GiB | 2 x 600 NVMe SSD | 10 Gbps |
Note All current generation instance types support enhanced networking, except for T2 instances.
EC2 Instance Type Name Components
Section titled “EC2 Instance Type Name Components”Instance type names follow a standard convention with multiple parts:
Example: c7gn.xlarge
- Family:
c(compute optimized) - Generation:
7(seventh generation) - Processor family:
g(AWS Graviton processor) - Additional capabilities:
n(Network and EBS optimized) - Size:
xlarge(extra-large instance)
Suitability of Instance Types for Workloads
Section titled “Suitability of Instance Types for Workloads”Compute Optimized
Use cases: Batch processing, distributed analytics, high performance computing (HPC)
Instance types: C7, C6, C5, C4
Work well for compute-bound applications that benefit from high-performance processors
Storage Optimized
Use cases: High-performance databases, real-time analytics, transactional workloads
Instance types: I4, Im4, Is4, I3, D2, D3, H1
Designed for workloads requiring high, sequential read/write access to large datasets on local storage
Memory Optimized
Use cases: In-memory caches, high-performance databases, big data analytics
Instance types: R7, R6, R5, R4, X2, X1, Z1
Designed to deliver fast performance for workloads that process large datasets in memory
Accelerated Computing
Use cases: Machine learning, artificial intelligence (AI), HPC
Instance types: P5, P4, P3, P2, DL1, Trn1, Inf2, Inf1, G5, G4, G3, F1, VT1
Use hardware accelerators to perform functions more efficiently than software running on CPUs
HPC Optimized
Use cases: Deep learning workloads, compute-intensive HPC workloads
Instance types: Hpc7, Hpc6
Purpose built to offer the best price performance for running HPC workloads at scale on AWS
Choosing an Instance Type
Section titled “Choosing an Instance Type”With over 270 available instance types, selection considerations include:
Performance and Cost Requirements
Section titled “Performance and Cost Requirements”- Analyze workload requirements for optimal performance
- Choose instance type that provides needed performance while minimizing cost
- Better to slightly under-spec initially and resize/scale as needed
Available Tools for Selection
Section titled “Available Tools for Selection”Instance Types page in EC2 console:
- Filter by characteristics you choose
- Search and compare choices
- Displays all available instance types in a Region
Recommendation Choose latest generation instance types for better price-to-performance ratio
AWS Compute Optimizer:
- Analyzes instance runtime behavior
- Makes recommendations for optimization
- Evaluate recommendations and modify accordingly
AWS Compute Optimizer
Section titled “AWS Compute Optimizer”AWS Compute Optimizer:
- Recommends optimal instance type, instance size, and Auto Scaling group configuration
- Analyzes workload patterns using Amazon Machine Learning
- Classifies instance findings as:
- Under-provisioned
- Over-provisioned
- Optimized
- None (insufficient data or unsupported instance type)
Currently generates recommendations for M, C, R, T, and X instance families.
Activity: Instance Type Matching
Section titled “Activity: Instance Type Matching”An EC2 instance type defines a configuration of CPU, memory, storage, and network performance characteristics. As a recommendation, choose new generation instance types in a family because they generally have better price-to-performance ratios. Use the Instance Types page in the Amazon EC2 console and AWS Compute Optimizer to find the right instance type for your workload.