Researchers and Instructors often require specialized computing environments tailored to their projects and courses. But creating and maintaining these environments can be time-consuming and a barrier for students, especially those with limited access to personal computing resources. The University of Oklahoma Libraries leverages the National Research Platform’s Nautilus project, a hosted Kubernetes framework, to address this challenge.
What is Nautilus?
The Nautilus cluster is a cluster of computers that have CPUs, a variety of GPUs, and file storage which is located primarily in the United States. It allows users to participate using shared resources contributed by universities. While this infrastructure can do many things, the University Libraries. is using the service to host pods (containers) for JupyterLab environments to access common software setups for coding languages and applications.
This framework is being used by the University Libraries for their hosted researcher workshops in R, Python, Data and Software Carpentry. Additionally, courses in Meteorology, Linear Algebra, Astronomy, and Biology have integrated Nautilus for use by learners. Several research groups are using the Nautilus framework on their own to create more detailed workflows and processes to aid in their research.
Faculty and instructors who may be interested in using the portal for their course can fill out a request form. A member of the UL staff will instruct on the best practices for using in a course and are informed as to what support is available. Any questions can be asked by sending an email to Data Scholarship and Data Services.
Computational Environments Currently Available
Scipy: A standard Python JupyterLab environment used that provides common algorithms for analysis and visualization of data.
Linear Algebra: Built on the Scipy environment, includes additional software packages (libraries) specific for use in linear algebra workflows.
Meteorology: Built on the Scipy environment , includes additional software packages (libraries) specific for use in meteorology workflows.
Tensorflow and PyTorch: Built on the Scipy environment, includes software packages (libraries) specific for use with Artificial Intelligence/ Machine Learning workflows. These two pods can also be configured to access GPUs.
Constellate: Built on the Scipy environment, includes additional software packages (libraries) specific for use in Natural Language and Text analysis workflows. Include the Constellate library for access content from JSTOR databases.
R and RStudio: A standard R JupyterLab environment and access the the RStudio application for coding R projects.
OpenRefine: Accesses the OpenRefine tool for working with messy data: cleaning it; transforming it from one format into another