Sun C1, Guo S1
1Sydney Informatics Hub, Core Research Facilities, The University of Sydney, Camperdown, Australia
Biography:
Dr Chao Sun is the Data Science Group Lead at the Sydney Informatics Hub, Core Research Facilities, University of Sydney, overseeing two development teams that support national research platform projects. With a background in electronics engineering and machine learning (PhD), Chao has spent over a decade collaborating with domain experts across disciplines. Chao began in the Faculty of Arts and Social Sciences, supporting humanities and social science researchers, then expanded to more faculties after joining the Sydney Informatics Hub in 2019. Since 2022, Chao has led development of the text analytics suite for the Language Data Commons of Australia (LDaCA), an ARDC–dHASS partnership, and since 2024 has contributed to the Australian Plant Phenomics Network (APPN).
Abstract:
Developing a web-based tool often means a contemporary UI and better compatibility across different systems; however, hosting the tool with considerations of security, cost, maintenance and scalability is always a challenge, especially for research tools that are designed as an app rather than a persistent service. A standing virtual machine requires regular patching, backups, security reviews and its own user authentication at significant cost for a project and is often forced to retire when the project ends. What is missing is a good, safe, maintenance-free way to deploy a host on demand that keeps running independently of the project.
The LDaCA team has been using BinderHub to deploy our Jupyter Notebook-based tool for years, and when the hosting need emerged after we developed our new web app, "Wordflow", we tried a different approach and found a hidden alternative. The ARDC BinderHub, hosted on the Nectar Research Cloud, is designed to launch Jupyter notebooks — but under the bonnet it is a specialised container launcher with federated login that spins up an instance for every authorised user. Using a minimal notebook launcher and a prebuilt Docker image, we boot our LDaCA Wordflow web app in each user's web browser. Every user gets a private, resource-capped, isolated instance on request; the launch URL pins an exact software release, so reproducible environments can be accessed from links in a tutorial or paper.
The LDaCA team has put this into production as Wordflow's try-it-now channel and workshop platform; the result serves researchers well and provides an identical experience to the local app for those who are not ready, or allowed, to install the app locally. The benefits for the development team are no server to maintain, no idle cost, and a safe, authenticated environment inherited from national digital infrastructure for research, where Australian and New Zealand researchers sign in with existing institutional credentials. This is not the service's originally intended use, yet it opens up more possibilities to serve researchers using the same existing infrastructure. We are grateful to ARDC for providing and supporting our use case. We present how it works, where it fits and where it does not. This is a generalisable pattern for any interactive, containerisable research application, with the potential to make on-demand hosting an additional viable capability for general researchers with limited technical capabilities.
