Skip to content

jupytercloud-project/jupyterhub-libcloudspawner

Repository files navigation

About LibCloudSpawner

Chat on Gitter

LibcloudSpawner enables JupyterHub to spwan single-user notebooks server inside fresh cloud instance.

LibcloudSpawner is based on Apache Libcloud which aims to provide an abstraction API for a lot of cloud providers. Apache Libcloud provides the ability to manage resources in AWS, Google Cloud or Azure. But for the moment, only OpenStack is supported.

Overview

A Jupyterhub Spawner start each single-user notebook server. LibcloudSpawner is a spawner that operate with cloud API to manage singleuser notebook server instance life-cycle. Each Jupyterhub user will have one cloud instance for his usage.

Benefits

  • needs one small instance for JupyterHub server, all other resources are dynamic
  • single-user servers have real OS, feel free to integrate them in your information system (ActiveDirectory, authentication, mountpoint...)
  • simple access to VT-io or pci-passthrough devices (GPU)
  • ability to host third-party services on user instance (VNC, RStudio...)
  • no need for Docker or Kubernetes

Drawbacks

  • user server could be very slow to start (depending cloud provider and image size)
  • no need for Docker or Kubernetes

See the full documentation for informations, use-case and tutorials