Skip to content

Interactive Sessions

CANFAR computing environments for astronomical research - Jupyter notebooks, and non-interactive applications.

🎯 Session Types Overview

Choose the right interface for your research:

🚀 Session Fundamentals

What are Interactive Sessions?

Interactive sessions provide on-demand access to pre-configured computing environments running in containers. Each session type offers different interfaces optimized for specific astronomical workflows.

Key Benefits

No Installation Required
Access complex astronomy software through your web browser without local installation or configuration.
Pre-Configured Environments
Containers include popular astronomy packages like AstroPy, CASA, and scientific Python libraries ready to use.
Persistent Data Access
All sessions automatically connect to your ARC storage and can access VOSpace for long-term data management.
Scalable Resources
Choose flexible or fixed resource allocation based on your computational requirements.

📊 Session Type Comparison

Session Type Interface Best For GUI Support
Notebook JupyterLab Data analysis, prototyping, documentation ✅ Web-based
Desktop Full Linux desktop CASA, image viewers, traditional software ✅ Desktop GUI
CARTA CARTA interface Radio astronomy visualisation ✅ Specialised
Firefly Firefly viewer Catalogue analysis, image display ✅ Web-based
Contributed Various Specialised applications ⚠️ Varies
Batch None (headless) Large-scale processing ❌ Headless

🔧 Session Management

Creating Sessions

Via Science Portal:
Launch sessions through the CANFAR Science Portal web interface with point-and-click simplicity.
Via Command Line:
Use the CANFAR CLI for scripted session creation and automation.
Via Python API:
Integrate session management into custom workflows using the CANFAR Python Client.

Session Lifecycle

Creation (30 seconds - 3 minutes)
Container download (first time) and startup with storage mounting
Active Use
Full access to computing resources and storage systems
Idle Management
Sessions automatically suspend after periods of inactivity to conserve resources
Termination
Container deletion with data preserved in persistent storage

Data Persistence

Important: Session containers are temporary. Always save important work to /arc/ storage or VOSpace before ending sessions.

📈 Resource Allocation

Flexible Allocation (Default)

Advantages: - Faster session startup - Can burst to higher resource usage when available - Optimal for interactive work and development

Best For: - Data exploration and analysis - Development and testing - Educational workshops

Fixed Allocation

Advantages: - Guaranteed consistent performance - Predictable resource availability - Better for production workloads

Best For: - Large-scale processing - Performance-critical analysis - Time-sensitive computations

Resource Selection Guide

Workflow Type Recommended Mode CPU/Memory Duration
Interactive Analysis Flexible 2-4 CPU, 4-8GB Hours
Large Dataset Processing Fixed 4-8 CPU, 16-32GB Hours-Days
Development & Testing Flexible 1-2 CPU, 2-4GB Hours
Production Pipelines Fixed Varies by workload Days

🔗 Integration with Platform Services

Storage Integration

All interactive sessions automatically mount:

  • ARC Home (/arc/home/[user]/): Personal configurations and scripts
  • ARC Projects (/arc/projects/[project]/): Shared research data and results
  • Scratch (/scratch/): High-speed temporary storage for processing

Additional storage accessible via API: - VOSpace (vos:): Long-term archives and data sharing

Container Environments

Sessions run in container environments that include:

  • Operating system (typically Ubuntu Linux)
  • Astronomy software packages (AstroPy, CASA, etc.)
  • Scientific computing libraries (NumPy, SciPy, Matplotlib)
  • Development tools and utilities

CVMFS Software Repositories

All CANFAR sessions provide access to read-only CVMFS (CernVM File System) software repositories. This feature provides instant access to the vast collections of pre-built scientific software maintained by the Digital Research Alliance of Canada (Alliance).

See the Software Repositories (CVMFS) guide for more information and examples.

Authentication & Permissions

Sessions inherit your CANFAR permissions:

  • Automatic access to your group projects
  • Secure integration with CADC services
  • API access for automated workflows

🎯 Choosing Your Session Type

For Data Analysis

New to CANFAR? → Start with Jupyter Notebooks
Familiar interface combining code, documentation, and visualisation
Need GUI Applications? → Use Desktop Sessions
Full Linux desktop for CASA, image viewers, and traditional software

For Astronomy Specialisations

Radio AstronomyCARTA Viewer
Optimized for radio interferometry data visualisation and analysis
Survey DataFirefly Viewer
Efficient table and image viewing for large astronomical catalogues
Specialised ToolsContributed Applications
Community-maintained applications for specific research domains

For Production Work

Large-Scale ProcessingBatch Sessions
Automated workflows for processing large datasets without interactive interfaces

API Access to Sessions

If you have a session running and it exposes an API, you can programatically interact with it remotely, either from another session or from your laptop.

The base of the session API will be available at the session URL. For more details, consult the API docs for each of the session types: - CARTA API Docs - Firefly API Docs - JupyterLab API Docs