Diogo Castro Federated collaborative workflows for Jupyter | JupyterCon 2023

Federated collaborative workflows for Jupyter simplifying workflow deployment, improving collaboration and user experience, with a focus on interoperability, scalability and data access.

Key takeaways
  • Federated collaborative workflows for Jupyter are designed to simplify workflow deployment, improve collaboration, and enhance user experience.
  • The project aims to deliver a global collaboration service, integrating various services and applications.
  • Key technology developed includes a storage plugin, Federated Federated Science (FES), and the Interoperability Platform (IOP) for seamless access to data and services.
  • The project faces challenges such as concurrent editing and data loss, which are addressed through locking mechanisms and parallel access to notebooks.
  • Federated Science (FES) enables researchers to work together on distributed file systems, reducing data duplication and improving collaboration.
  • Integrated services include Jupyter, Voila, and EFSS systems, which provide a cohesive and easy-to-use system for users.
  • The project’s success relies on the development of APIs and standardization to enable seamless integration with various services and applications.
  • Applications such as CERNbox and OnCloud provide a convenient way to access files and data, using APIs for seamless integration.
  • The project has achieved successful collaboration with other European research institutions, and plans to expand to cover more geographies and applications.
  • Constraints such as concurrent editing, data loss, and scaling limitations are addressed through innovative solutions like parallel access to notebooks.
  • SWAN (Swiss Workflow) is a Jupyter service built on top of FES, providing a platform for researchers to work together on distributed file systems.
  • The project’s foundation will be used to develop more applications beyond Jupyter, leveraging the Federated Science (FES) architecture.
  • CERNbox provides a sync and share service for researchers, allowing access to data and files from anywhere.
  • Applications such as Jupyter, Voila, and EFSS systems are integrated with CERNbox for Seamless access to data and services.
  • The project’s success relies on user acceptance and adoption of federated collaborative workflows, and future work includes enhancing interoperability and expanding the service mesh.
  • The project aims to provide a unified mechanism for accessing and sharing data across different institutions and countries.