scieee AI-readable full text Open interactive document viewer

Contributing to the HelioCloud Open Science Cloud Computing Community

Antunes, Alex

Abstract

HelioCloud is an open-source software stack that provides research groups an easy onboarding path into big data/high performance computing cloud capabilities, specifically tuned for heliophysics research and backed by 1.5 Petabytes of science data stored in the AWS cloud. Any institution can spin up a own HelioCloud for ephemeral (summer schools, class) and institutional (research groups, long-term project) use. It exists at heliocloud.org and in Github, it works, and to future-proof it, we are opening the doors for community contributions. We discuss our draft community standards for the Jupyter Notebook-based Daskhub, the Portal for creating VMs, the many tutorial Notebooks, and the ancillary tools such as CloudCatalog. What we hope to gain from DASH is an understanding from participants of what will make it easier for you to use and contribute to this open science infrastructure cloud suite. Plus we hope you will check HelioCloud out and consider it for boosting your own scientific workflow.

Full text

Contributing to the HelioCloud Open Science Cloud Computing Community Alex Antunes [1], Brian Thomas [2], Peter Shumate [1], Jeffery Bradford [2], Lisa Knowles [1], Sarah Rourke [2], Jon Vandegriff [1], Omar Shalaby [2] [1] JHUAPL, [2] NASA/GSFC We discuss our draft community standards for the HELIOCLOUD Jupyter Notebook-based Daskhub, the Portal for creating VMs, the many tutorial Notebooks, and the ancillary tools such as CloudCatalog. What we hope to gain from DASH is an understanding from participants of what will make it easier for you to use and contribute to this open science infrastructure cloud suite. Plus we hope you will check HelioCloud out and consider it for boosting your own scientific workflow. Contact: [email protected] https://github.com/heliocloud-data/ Packages! ● platform (Jupyter) ● portal (VMs) ● runtimes (containers) ● cloudcatalog (indexing tools) ● science-tutorials (examples) ● heliocloud_tools (misc) ● ssw-jnplot (Jupyter SSW hack) We release as FOSS, but want to shift to developing as community FOSS Current state & Problems: 1) more Kubernetes than Python 2) Public releases in Github (yay!) 3) Public dev of CloudCatalog, Tutorials, Community Tools in Github (yay!) 4) Private dev of core Platform/Portal & Runtimes in Gitlab (boo!) 5) CI/CD & test env needs to be ported to Github 6) ‘Gatekeeping’ & steering committee roles not yet clear 7) Unlike a PyHC package, not clear who is ‘contributing community’ Question for DASH (vote here): Is it more likely that HelioCloud Draft in progress on ‘how to contribute!’ User Portal to create VMs in cloud (EC2) Burstable Notebook environment which has GPU acceleration Browser-based Service Components (C) Other? (B) people are more likely to fork for their domain rather than contribute (A) the community will issue and sometimes implement Feature Requests