Exosphere: A researcher-friendly interface for cloud computing
Have your computing needs exceeded the capacity of your local machines? Considering the cloud-computing resources funded by NSF, but not sure where to start? Look no further than Exosphere, an open source cloud-computing interface designed for researchers, not system administrators.
Julian Pistorius (CCT Data Science, UArizona) and collaborator Chris Martin created Exosphere in 2017 to fill an urgent need: existing cloud-computing interfaces were either proprietary or required advanced knowledge of system administration. Exosphere’s interface was designed for scientific research teams and will be the one of the primary user interfaces for the forthcoming, NSF-supported Jetstream 2.
The CCT Data Science team currently uses Exosphere to provision virtual infrastructure (hosted by CyVerse) for research computation. One example is the Predictive Ecosystem Analyzer (PEcAn), whose complex software stack is deployed virtually via Exosphere to enable model development and computation.
In November 2020, Chris and Julian presented Exosphere at SuperCompCloud: 3rd International Workshop on Interoperability of Supercomputing and Cloud Technologies, part of the SuperComputing 2020 conference. This workshop paper also had co-authors David L. LeBauer (The University of Arizona) and Sanjana Sudarshan (Indiana University).
From the abstract:
Exosphere provides researcher-friendly software for managing computing workloads on OpenStack cloud infrastructure. Exosphere is a user-friendly alternative to Horizon, the default OpenStack graphical interface. Exosphere can be used with most research cloud infrastructure, requiring near-zero custom integration work.
Chris and Julian also presented Exosphere to the Scientific SIG at a recent OpenStack Project Teams Gathering (PTG). See below for a short video:
It is easy to deploy a version of Exosphere with your own institutional branding.
Contact chris@c-mart.in or julianp@arizona.edu for more information.
Do you have a research computation problem that exceeds the capability of local resources, but unsure how access free cloud computing? Come meet with us at our Drop in-hours.
Acknowledgements
- Jetstream via Indiana University: For ongoing support
- CyVerse: For hosting resources