Workspace Types & Hardware

Workspace Types

Linux Workspaces

Type Description Best For
Amazon Linux 2 General-purpose Linux environment Command-line work, scripting, development
Ubuntu Popular Linux distribution with extensive package support Data science, development, general research
RHEL Enterprise Linux with long-term support Production workloads, enterprise software

Windows Workspaces

Type Description Best For
Windows Server Windows environment with remote desktop access Windows-only software, GUI applications

Specialized Workspaces

Type Description Best For
SageMaker Jupyter notebook environment for ML Machine learning, data science
EMR Big data processing cluster Large-scale data processing

Hardware Configurations

General Purpose

Configuration vCPUs Memory Use Case
Small 2 4 GB Light development, testing
Medium 4 16 GB Standard analysis, development
Large 8 32 GB Moderate computational work
X-Large 16 64 GB Heavy computational work

Compute Optimized

Configuration vCPUs Memory Use Case
Compute Medium 8 16 GB Parallel processing
Compute Large 16 32 GB High-performance computing
Compute X-Large 36 72 GB Intensive simulations

Memory Optimized

Configuration vCPUs Memory Use Case
Memory Medium 4 32 GB Large dataset analysis
Memory Large 8 64 GB In-memory processing
Memory X-Large 16 128 GB Very large datasets

GPU Enabled

Configuration vCPUs Memory GPU Use Case
GPU Small 4 16 GB 1x NVIDIA T4 ML inference, light training
GPU Medium 8 32 GB 1x NVIDIA V100 ML training
GPU Large 32 128 GB 4x NVIDIA V100 Deep learning, large models
Warning

GPU instances have limited availability and higher costs. Request GPU access only when your workload specifically requires GPU acceleration.

Storage Options

Storage Type Default Maximum Notes
Root Volume (SSD) 50 GB 500 GB Operating system and applications
Data Volume Optional 1 TB Additional storage for datasets

Choosing the Right Configuration

Factor Guidance
Workload Type Development → Small/Medium; Data analysis → Medium/Large; Simulations → Compute optimized; ML → GPU
Data Size < 10 GB → Standard root volume; 10-100 GB → Increase root volume; > 100 GB → Add data volume
Duration Short tasks → Larger instance for speed; Long-running → Smaller instance for cost
Budget Larger instances cost more per hour; balance size vs. duration
Tip

Start with a smaller configuration and scale up if needed. You can always terminate and create a new workspace with different specifications.