Canadians land on Jupyter
Abstract
Over the past 18 months, Ian Allison and James Colliander have deployed Jupyter to more than 8,000 users at universities across Canada. Ian and James offer an overview of the Syzygy platform (named after the term for an alignment of three or more celestial points) and explain how they plan to scale and deliver the service nationally and how they intend to make Jupyter integral to the working experience of students, researchers, and faculty members.
Ian and James accomplished this feat with a relatively small allocation on the OpenStack installations of their partners and with a tiny core support team. The goal was to bring this transformative technology into the postsecondary mainstream, right now, where it can effect positive transformation in the way we teach and learn. The goal of the platform is to make the Jupyter resource available to as many people as possible while eliminating the barriers encountered along the way. The vision is institutional but with an international alignment on how to better help students learn and teachers teach.
The deployments leverage automation and portability; this is how they succeed with minimal resources. But succeeding with a minimal support infrastructure also required the development of tools and support mechanisms. This toolset has been built around Terraform, Ansible, and Docker, which allowed Ian and James to go from a cloud provider allocation to a working hub in just a couple of steps, in under an hour. This strategy, together with their determination to make access as simple as possible (e.g., by leveraging existing institutional credentials), has been extraordinarily successful and has been met with a large and growing appetite for Jupyter by everyone in the community. These successes are now driving the development of a K–12 platform using a similar approach and vision.
Ian and James discuss how they curate resources to satisfy a large and diverse user base, outline some of the obstacles encountered and the tools created to overcome them, and share testimony from several Syzygy users and lessons learned working with large user groups, such as providing convenient authentication via shibboleth and partitioning resources for large user groups via Docker. They conclude by detailing plans for using Kubernetes and similar tools to expand the reach of a platform that has already allowed over 8,000 Canadians to land on Jupyter.
