Platform Comparison

Choosing the Right Platform

RACE offers four platforms to meet different needs. This guide helps you understand the differences and choose the best option for your use case.

Platform Comparison Table

Platform Details

Compute Service

Platform: AWS
Best for: Researchers who need secure, repeatable research environments

Compute Service is a cloud solution that enables IT teams to provide secure, repeatable, and federated control of access to data, tooling, and compute power that researchers need.

Key Features:

• Secure environments - Isolated research environments with controlled data access
• Federated access control - Multi-institutional collaboration support
• Automated setup - Create baseline research environments in minutes, not months
• Simplified data access - Managed data access controls
• Price transparency - Built-in cost tracking and visibility
• Multiple connection methods - SSH, RDP, and web-based access

Ideal use cases:

• Collaborative research projects with external institutions
• Projects requiring standardized, reproducible environments
• Research teams needing secure data access controls
• Studies requiring rapid environment provisioning

Limitations:

• Limited to VM-based workloads
• Cannot access all AWS services directly
• Environments configured by IT team; less flexibility than direct AWS access

Loome

Platform: Azure
Best for: Users who need low-friction cloud access with simple cost management

Loome offers a low-friction entry point to the cloud, with a single portal for managing projects, computing costs, and elastic demand. With autonomy to scale storage and processing on-demand, you can manage workstations, clusters, and docker images with ease.

Key Features:

• Single console - Manage compute, storage, and costs in one place
• Simple provisioning - Easy interface for HPC clusters, storage, and workstations
• Automatic cost tracking - Costs tracked against budgets automatically
• Curated templates - Resource templates for specific analysis needs
• On-demand scaling - Turn resources on/off as needed
• Variety of options - Multiple compute and storage configurations

Ideal use cases:

• High-performance computing clusters
• Research workstations with significant compute power
• Projects requiring easy resource management and cost visibility
• Teams needing simplified cloud access without deep Azure expertise

Limitations:

• Limited to Azure infrastructure
• Resource options curated by platform
• Some advanced Azure features may require direct access

AWS Direct Access

Platform: AWS
Best for: Experienced users who need full AWS capabilities

Direct AWS access provides full access to AWS services through institutional AWS accounts. Researchers have flexibility to configure and manage their own AWS resources.

Key Features:

• Full AWS console access - Use any AWS service
• Complete flexibility - Design your own architecture
• Advanced services - Access to all AWS services including:
  • Compute: EC2, Lambda, Batch
  • AI/ML: Bedrock (generative AI), SageMaker
  • Quantum: Braket
  • Storage: S3, EFS
  • Database: RDS, DynamoDB
• Infrastructure as Code - Use CloudFormation, Terraform, etc.

Ideal use cases:

• Complex multi-service architectures
• Custom networking requirements
• Advanced AWS services (SageMaker, Braket, Bedrock)
• Projects requiring specific AWS configurations

Requirements:

• AWS knowledge and experience
• Understanding of cloud security
• Ability to manage costs independently

Azure Direct Access

Platform: Azure
Best for: Experienced users who need full Azure capabilities

Direct Azure access provides full access to Azure services through institutional Azure subscriptions. Researchers can provision and manage Azure resources independently.

Key Features:

• Full Azure portal access - Use any Azure service
• Complete flexibility - Design your own architecture
• Advanced services - Access to all Azure services including:
  • Compute: Virtual Machines, Azure Batch
  • AI/ML: Azure OpenAI Service, Azure Machine Learning
  • Storage: Blob Storage, Azure Files
• Microsoft integration - Native integration with Microsoft tools

Ideal use cases:

• Complex Azure architectures
• Microsoft-centric workloads
• .NET and Windows-based applications
• Integration with Microsoft 365
• Projects requiring Azure OpenAI Service

Requirements:

• Azure knowledge and experience
• Understanding of cloud security
• Ability to manage costs independently

Decision Guide

Use this flowchart to help choose your platform:

┌─────────────────────────────────────────────────────────────┐
│              What are your cloud skills?                     │
└─────────────────────────────────────────────────────────────┘
                           │
           ┌───────────────┴───────────────┐
           ▼                               ▼
    ┌─────────────┐                 ┌─────────────┐
    │  Beginner   │                 │  Advanced   │
    └─────────────┘                 └─────────────┘
           │                               │
           ▼                               ▼
┌─────────────────────┐         ┌─────────────────────┐
│ Do you need Azure   │         │ Which cloud do you  │
│ specifically?       │         │ prefer?             │
└─────────────────────┘         └─────────────────────┘
     │           │                   │           │
     ▼           ▼                   ▼           ▼
   Yes          No                 AWS        Azure
     │           │                   │           │
     ▼           ▼                   ▼           ▼
  Loome    Compute Service   AWS Direct  Azure Direct

Managed vs Self-Managed: Trade-offs

Managed Services (Compute Service, Loome)

Advantages:

• Quick setup - get started in minutes
• No cloud expertise required
• Built-in security and compliance
• Automatic cost controls and tracking
• Simplified support from RACE team
• Focus on research, not infrastructure

Trade-offs:

• Limited to available configurations
• Cannot access all cloud services directly
• Less flexibility for custom requirements

Self-Managed (AWS Direct, Azure Direct)

Advantages:

• Full access to all cloud services
• Complete architectural flexibility
• Can implement any solution
• Direct control over all settings
• Access to cutting-edge services (Braket, Bedrock, Azure OpenAI)

Trade-offs:

• Requires cloud expertise
• You manage security and compliance
• Must monitor and control costs manually
• Steeper learning curve

When to Use Each Platform

Still Not Sure?

If you’re unsure which platform is right for you:

1. Start with Compute Service - It’s the easiest way to get secure computing resources
2. Contact the RACE team - They can help assess your needs and recommend the best option
3. Try multiple platforms - You can request access to more than one

Next Steps

Once you’ve decided on a platform:

1. Learn about funding schemes and request access
2. Follow the platform guide for detailed instructions:
   • Compute Service Guide
   • Loome Guide
   • AWS Access Guide