Skip to content

Runtime Guide

This guide helps you choose the right runtime for your use case. The Digitalhub platform provides several specialized runtimes, each designed for different types of workloads and execution patterns.

Quick Runtime Selection

If you need to... Use this runtime Best for
Execute Python code Python Runtime General-purpose Python execution
Run containerized applications Container Runtime Custom containers, complex dependencies
Orchestrate multiple steps Hera Runtime Workflow orchestration, DAGs
Serve ML models ModelServe Runtime Model inference, API serving
Transform tabular data DBT Runtime Data transformation, SQL workflows
Run federated learning Flower Runtime Federated learning, privacy-preserving ML

Runtime Capabilities Matrix

Runtime Local Execution Remote Execution Multi-step Workflows Model Serving Data Processing Federated Learning
Python ✅ ✅ ❌ ✅ ✅ ❌
Container ❌ ✅ ❌ ❌ ✅ ❌
Hera ❌ ✅ ✅ ❌ ✅ ❌
ModelServe ❌ ✅ ❌ ✅ ❌ ❌
DBT ✅ ✅ ❌ ❌ ✅ ❌
Flower ✅ ✅ ❌ ❌ ❌ ✅

Detailed Selection Guide

Python Runtime

Choose when:

  • You have Python code that needs to run
  • You want simple job execution or model training
  • Your workload fits within Python's ecosystem

Container Runtime

Choose when:

  • Your application requires complex dependencies
  • You want to run existing containerized applications
  • Python runtime isn't sufficient for your environment needs

Hera Runtime

Choose when:

  • You have multiple steps to execute in sequence
  • You want to build DAGs (Directed Acyclic Graphs)
  • You need conditional execution or parallel processing

ModelServe Runtime

Choose when:

  • You need to serve machine learning models
  • You want to expose models as REST APIs
  • You need scalable model inference
  • You work with various ML frameworks (scikit-learn, MLflow, HuggingFace, etc.)

DBT Runtime

Choose when:

  • You work with tabular data
  • You need SQL-based data transformations

Flower Runtime

Choose when:

  • You want to implement federated learning
  • You need privacy-preserving machine learning
  • You work with distributed datasets across multiple parties
  • You want to use the Flower framework for FL

Getting Started with Your Chosen Runtime

Once you've selected a runtime, follow these steps:

  1. Read the overview for your chosen runtime
  2. Check the examples to see common usage patterns
  3. Review the execution guide for detailed parameter information
  4. Explore the entity documentation for complete API reference

Need Help Choosing?

If you're still unsure which runtime to use, consider:

  • Start with Python Runtime if you're new to the platform
  • Use Container Runtime if you have complex dependencies
  • Use Hera Runtime if you have multiple coordinated steps
  • Use ModelServe Runtime if you need to serve ML models
  • Use DBT Runtime if you work primarily with SQL and tabular data
  • Use Flower Runtime if you need federated learning capabilities