The Turing Data Safe Haven: An open, scalable, reproducibly deployable, cloud-based Trusted Research Environment for working safely with sensitive data

Presented on 15 February 2023 at the Research Software Engineering in Data and AI workshop hosted by The Alan Turing Institute and The University of Warwick (15-17 February 2023). The recording of this talk can be streamed on YouTube. Abstract Researchers often need to analyse sensitive data in order to answer important questions for health, government and society. Balancing the need to protect individual privacy and commercial confidentiality with the need to ensure that reliable insights can be made using sensitive data is challenging, and Trusted Research Environments (TREs) that provide secure analysis environments for working with sensitive data are a key part of striking this balance. The Alan Turing Institute has recently open-sourced its Data Safe Haven, a secure, scalable, reproducibly deployable, cloud-based TRE that we have been using for the last 4 years to support our researchers in working safely and productively with sensitive data. By openly publishing the code and documentation required to deploy, configure and manage a Data Safe Haven instance, we make it easier for others to deploy their own Trusted Research Environments, reducing the effort required to support their researchers in working safely with sensitive data. We've also been working closely with the wider community to better align work in this area, initially on the development of the information governance principals and design choices for our Data Safe Haven, which we published in 2019. More recently, we've been working closely with the DARE UK programme to define common requirements for Trusted Research Environments and have co-founded the RSE TRE Community, bringing together those who are developing and deploying TREs to more closely co-ordinate our work. In this talk we will give an overview of the Turing's Data Safe Haven and the information governance and design principals behind it, as well as the work we have been doing with the wider community to develop a common set of TRE requirements and to co-ordinate across organisations on the development, deployment and management of TREs.This work was supported by The UKRI Strategic Priorities Fund under the EPSRC Grant EP/T001569/1, particularly the "Tools, Practices and Systems" theme within that grant, and by The Alan Turing Institute under the EPSRC grants EP/N510129/1, EP/W001381/1, EP/W037211/1 and EP/X03870X/1