Containers in research: Reproducibility, portability, and efficiency

Containers offer portability and reproducibility, making them ideal for operational deployments. Despite their focus in these environments, they remain highly valuable for research settings such as HPC for the same attributes of portability and reproducibility. Examples include: • Sharing containers to separate the learning curve between software installation and use • Sharing definition files to ensure research software produces reproducible results • Using isolated environments to safely test potentially breaking code changes• Abstracting data sources to simplify workflow automation • Utilizing software that requires elevated permissions • Enabling access to GPUs and interactive tools • Extending the longevity of research software • Aligning workflows with evolving HPC standards, such as decentralized compute These examples will be discussed in the context of research software engineering projects we have undertaken with New Zealand researchers.ABOUT THE AUTHORChris Zweck obtained a PhD in glaciology from the Antarctic Research Centre at the University of Tasmania. After postdocs at the Geophysical Institute, Fairbanks, Alaska and Alfred Wegener Institute, Bremerhaven, he worked as an associate professor at the Department of Hydrology and Water Resources at the University of Arizona and Faculty of Earth Sciences at the University of Iceland. Since moving to New Zealand he has worked at the MetService and GNS. -----For more information about the eResearch NZ / eRangahau Aotearoa conference, visit: https://eresearchnz.co.nz/