Guidance on Tech and Data 

The Azure Trusted Research Environment (TRE) provides a secure platform for researchers to deploy virtual machines (VMs) for data analysis. Within each workspace, users can deploy their own isolated VMs; however, they cannot access or view each other’s VMs, ensuring user privacy and data security. Users in the same workspace have access to the shared airlock system for controlled data import and export.
Through the airlock, users can also import files or software installations, subject to approval.

Standard Offering

Researchers can deploy either Windows or Linux VMs using custom-built images provided by OUH. These images come pre-installed with essential software, including: Python, Anaconda, Visual Studio Code, R, RStudio, Azure Storage Explorer, PostgreSQL, Git, LibreOffice

Shared storage location within the workspace, between researchers. This has a maximum of 100Tb.

This list may be updated over time as the TRE evolves to meet new research needs.

VM Size Options 

GPU VMs – availability is currently constrained across Azure nationally

H100 series have very limited availability and deployment delays are likely.

A10 series — please be aware that usage may be subject to constraints.

AI research

This variety of VM sizes allows researchers to select the configuration that best fits their project requirements, balancing performance and cost.

Workspace Services

Workspace services are modular resources that can be deployed within a specific workspace to support various research activities.

Examples of workspace services include:

• Apache Guacamole Virtual Desktop: Provides remote desktop access.
• Azure SQL Database: Offers a managed database for data storage and querying.
• Azure Machine Learning : Facilitate machine learning model training, tracking, and deployment.
• Azure Databricks: Supports collaborative data engineering and analytics.
• Gitea: Self-hosted Git service for code collaboration.

Shared Services

Shared services are modular resources deployed across the TRE, available for use by all workspaces. Examples of shared services include:

• Nexus Package Mirror: Provides a local repository mirror for APT, PIP, Docker, and Conda packages, allowing installation of required packages on fully internet-restricted VMs to maintain secure, compliant environments.
• Gitea: Enables a central, self-hosted Git repository for shared code