Chameleon Trovi Support for Complex Experiment Appliances

Overview

Chameleon has historically enabled researchers to orchestrate complex appliances—large, multi-node clusters configured via OpenStack Heat—to conduct advanced experiments. Meanwhile, Trovi has emerged as a service for sharing reproducible artifacts. This project aims to bridge the two by adding support in Trovi for importing, discovering, and launching complex appliances. By integrating these capabilities, researchers will be able to one-click deploy complex appliances directly from the Trovi dashboard, archive them for future reference, and reproduce experiments on demand.

Key Outcomes

  • Extended Trovi API: Enable the import and management of complex appliances as artifacts.
  • Streamlined One-Click Launch: Integrate with Chameleon’s existing provisioning workflows so users can launch multi-node clusters directly from Trovi.
  • Enhanced Dashboard Experience: Provide UI assistance for discovering, reviewing, and customizing complex appliance artifacts.
  • Improved Artifact Reproducibility: Automate the process of exporting CC-snapshot images and other resources to ensure everything is preserved across sites (UC, TACC), highlighting any parameters that need user attention for cross-site portability.

Topics: Reproducible Research, Cloud Computing & Orchestration, OpenStack Heat, UI/UX & Web Development Skills: Python, APIs, Cloud (OpenStack), DevOps & Automation, Frontend Difficulty: Hard Size: Large Mentors: Mark Powers Tasks:

  • Extensions to the Trovi API
    • Add support for importing complex appliances as artifacts (including Heat templates, metadata, and associated disk images).
    • Develop methods for tagging, versioning, and categorizing these appliances, making them easier to discover.
  • One-Click Launch of Complex Appliances
    • Integrate with Chameleon’s orchestration engine, enabling single-click cluster deployments from the Trovi UI.
    • Validate correct configuration and resource availability through automated checks.
  • Trovi Dashboard Enhancements
    • Update the front-end to provide intuitive controls for customizing or parameterizing complex appliances before launching.
    • Offer a clear workflow for reviewing dependencies, resource requirements, and usage instructions.
  • Automated Export & Multi-Site Testing
    • Streamline the export of snapshots or images into Trovi as part of the appliance import process.
    • Optionally re-run the imported appliances at multiple sites (UC, TACC), detecting any unparameterized settings or missing dependencies.
Kate Keahey
Kate Keahey
Scientist at the Argonne National Laboratory and Senior Fellow at the Computation Institute at the University of Chicago

Kate is a Scientist at Argonne National Laboratory and a Senior Fellow at the Computation Institute at the University of Chicago.

Mark Powers
Mark Powers
Research Software Engineer at the University of Chicago

Mark Powers is a research software engineer at the University of Chicago, and is the DevOps lead for the Chameleon Cloud testbed.