Request a JupyterHub

You Focus on the Science — We'll do the Rest

The NSF Unidata Science Gateway offers custom JupyterHub clusters deployed on top of the NSF Jetstream2 Cloud at no cost to researchers and educators in the Earth Systems Science community. In addition to enabling users to create and run their own interactive python notebooks, our JupyterHubs can be equipped to track one or many GitHub repositories to easily share content with your users. Furthermore, as JupyterHub is able to deliver preconfigured compute environments to a large number of users, researchers, instructors, and workshop leaders no longer need to spend valuable time assisting their students and colleagues with replicating their environment.

To get started, complete the webform below. Requesting a JupyterHub may take between 15 and 20 minutes, and you will be asked to provide some technical and non-technical information.  If you are unsure about your responses to some fields, note that we provide suggestions for what is standard and sufficient for most applications.

Caveats

  • Logins to the JupyterHub server will be based on a GitHub mechanism (OAuth). As a result, each user must have a GitHub account. (Register for a free account here.)
  • Cluster lifetimes are limited to 1 semester. If you require an extension, for example to last through a summer or winter intersession, contact us once your initial request term is close to expiration.
  • Resources on Jetstream2 are limited. While we try to accommodate every request, we may have to make modifications to your requirements to ensure everybody gets a piece of the pie.

Acknowledgements

NSF Unidata is able to provide this offering by making use of resources on the Jetstream2 Cloud, hosted by Indiana University and managed by the ACCESS* program. The Unidata Science Gateway team is grateful for the support provided to us by members of their staff.

* ACCESS is an advanced computing and data resource program supported by the U.S. National Science Foundation (NSF) under the Office of Advanced Cyberinfrastructure awards #2138259, #2138286, #2138307, #2137603 and #2138296.

Acknowledging the NSF Unidata Science Gateway

If you ultimately benefit from the Unidata Science Gateway and these JupyterHub resources, please cite this DOI doi:10.5065/688s-2w73 in your scholarly publications.

JupyterHub Request Form

Why GitHub? What permissions do admins have?

We use a GitHub OAuth App as a simple and reliable method for us to authenticate users to JupyterHubs on our science gateway.

As an admin on your JupyterHub, you'll be able to start/stop servers, access user servers, and importantly, add and remove users, as well as grant them admin privileges (TAs, co-PI's, etc.).

JupyterHub Purpose
Teardown Period
Preferred Maintenance Days
Number of Users
Storage
What can the shared disk be used for?
The most common use for a shared disk is to allow your JupyterHub users to access large data sets used for demonstrations or assignments.
Computational Profiles
Why would I need multiple profiles?

Examples of scenarios where you may need multiple profiles:

  • You are an instructor teaching multiple classes with different computational needs
  • You require a "low power" environment with fewer CPUs and less RAM for synchronous classroom demonstrations and a "high power" environment for async assignments
What is in the Unidata Standard environment?
The Unidata Standard environment can be found on GitHub and contains common meteorological packages, many developed by NSF Unidata, such as metpy, siphon, and python-awips, in addition to third party packages such as matplotlib, cartopy, pandas, and xarray.
What is in the Unidata Standard environment?
The Unidata Standard environment can be found on GitHub and contains common meteorological packages, many developed by NSF Unidata, such as metpy, siphon, and python-awips, in addition to third party packages such as matplotlib, cartopy, pandas, and xarray.
What is in the Unidata Standard environment?
The Unidata Standard environment can be found on GitHub and contains common meteorological packages, many developed by NSF Unidata, such as metpy, siphon, and python-awips, in addition to third party packages such as matplotlib, cartopy, pandas, and xarray.
Running "computationally intensive" tasks, especially if multiple users are doing so at once, may result in heavy performance losses or a crash of your server. The NSF Unidata Science Gateway team has methods to get around this, however this requires more resources.
What counts as "computationally intensive"?

By "computationally intensive," we mean processes that are likely to use all CPU power available to them.

We do not consider most earth systems science workflows in python as computationally intensive.

Tools/software that are traditionally executed via a command line, such as WRF or software from the LROSE suite, are more likely to cause performance issues or crashes.

If you are unsure whether or not your workflows may be problematic in a JupyterHub environment, use the Iteration Phase to work with us to determine this.

Dask is a python library that enables you to create clusters for parallel and distributed computing. While we make provisioning a Dask cluster for your instruction or research easy, it is important to note that Dask is useful for specific problems in ESS, but not required for many.
Learn more about Dask

Visit the Dask Home Page

Read the Dask Documentation

Project Pythia's Dask Cookbook, a brief introduction to Dask and how it could potentially be used in Earth Systems Science workflows

As GPUs on Jetstream2 are limited, requesters of GPUs on the NSF Unidata Science Gateway must provide a description of how they intend to use their GPUs.
More Info and Caveats
Jetstream2 comes equipped with NVIDIA A100 GPUs. NSF Unidata Science Gateway Staff has been working to provide this GPU capability to our community.
 
As a new capability you may run into some problems as we learn how to best support your endeavors.
If you require assistance or further discussion of requirements that were not quantified by this form, please describe your situation below. We will contact you at the Email you've provided.
We are always excited to hear how our community members are making use of Science Gateway resources, and to learn how we can make your experience better.
 
May we contact you after your Teardown Phase to discuss these matters and others?
Solve this simple math problem and enter the result. E.g. for 1+3, enter 4.
Sorry to make you do this. We're trying to keep the spambots at bay.