Choosing research compute
An overview of University-managed or facilitated research compute options.
Research compute options
The University of Auckland provides access to these research compute services:
- Nectar virtual desktop
- Managed virtual machines
- Nectar self-service virtual machines
- Secure Research Environment (SRE)
- High Performance Computing (HPC)
Comparison of research compute services
| Nectar virtual desktop | Managed virtual machines | Nectar self-service virtual machines | Secure Research Environment | REANNZ HPC Platform, previously NeSI | |
|---|---|---|---|---|---|
| Ideal use | Test or demo a workflow, light to moderately intensive computations. | Moderately intensive computations, collaboration, test a workflow, less active management by the researcher. | Moderately intensive computations, collaboration, test a workflow, more active management by the researcher but more control. | Moderately intensive computation, storage and software in access controlled environment. | Highly intensive computations, executed in parallel or serial, via a queued job scheduler. |
| Suitable for sensitive research data | Only if you select a desktop library based in the Auckland zone. | Yes | Requires a consult to determine | Yes | Contact REANNZ |
| Service duration | Expires 14 days after last activity | Approved for up to 1 year. Renewable | Approved for up to 1 year. Renewable | By arrangement | Contact REANNZ |
| Access to GPUs | No | Planned to be available from late 2025 | Yes | By arrangement | Yes |
| Environment | Windows or Linux | Windows or Linux | Windows or Linux | Windows or Linux | Linux or JupyterLab |
| Standard hardware limits | See: Virtual Desktop Service | 16 CPUs 64GB RAM | See: Allocation Management | By arrangement | See: NeSi |
| Located in Aotearoa New Zealand | Yes | Yes | Yes | Yes | Yes |
Considerations for choosing research compute services
Data security classification
Review the Research data classification standard to identify appropriate classification(s) as this will inform which research compute services are appropriate.
Hardware needs
Some analyses require a certain level of compute power to complete tasks within realistic timeframes, while other analyses may benefit from specialised hardware (such as Graphical Processing Units, GPUs). Researchers can develop a workflow with a subset of the data on a smaller compute allocation for testing, and when ready, compute power can be scaled up.
Collaborator access
The number of project members that need access and whether or not they are internal to the University may affect the compute services available to a project.
Legal, ethical, and policy requirements
You may be required to handle or store your research data in particular ways based on legislation (such as the Privacy Act 2020), ethics approvals, or agreements with funders or data providers. See the Research Data Management Policy guidance for more information.
Storage
Research Drives can be connected to Managed virtual machines and the REANNZ High Performance Computing (HPC) platform, whereas they are not compatible with Nectar self-service virtual machines or Nectar virtual desktop. System administration experience: The Centre for eResearch will manage user access permissions, security patching, and most software installation on Managed virtual machines, but you will be required to manage these on Nectar self-service virtual machines.
Environment experience
Most services provide options for Windows or Linux environments, but part of the process of applying for access to REANNZ's High Performance Computing (HPC) is an assessment of prior HPC experience.
Security responsibilities
Service-specific security requirements are outlined on each service page. All users of research compute services are required to abide by the:
- IT Security Policy
- IT Acceptable Use Policy
- Research Data Management Policy, supported by the Data classification standard, Research data retention requirements, and Research Data Management Policy guidance.
Request a consultation
Researchers are encouraged to consult with experts from the Centre for eResearch to determine the most suitable research compute service for their needs.
Contact
Research Virtual Machines Support Services
Email: research-vm@auckland.ac.nz
Sean Matheny
Senior Platforms & Services Engineer
Email: s.matheny@auckland.ac.nz