Speaker
Description
At the CS3 conference in 2018, a panel of very wise people gravely informed us that sync and share is dead.
Moving beyond simple sync and share, over this past year the focus of AARNet’s CloudStor team has been towards building a platform that further enables our users to do more with their data without leaving the system. From collaborative document editing to data analytics with Jupyter Notebooks, we’re now investigating connecting CloudStor with the Australian NeCTAR Research Cloud, combining compute and storage to provide researchers the best of both worlds.
This talk covers the development process of an API that builds on top of our existing tenant portal, which is designed to allow institutions access to manage their users and provide self service functionality. Of particular relevance to this use case is the ability to create and manage group drives - shared project spaces for any given group of researchers, with their own quota allocations.
By exposing this functionality via an API, we aim to enable users to provision virtual machines on NeCTAR’s OpenStack cluster with automatically attached storage from CloudStor via the WebDAV interface. This allows researchers to access data in a group drive directly from a file system, instead of having to download, process and upload. In cases where a group drive does not already exist within CloudStor, one can be automatically provisioned for the corresponding NeCTAR project, further facilitating the objective of data life cycle management along the entire arc of research -- from data creation to eventual publication.