UVA Outreach Notes¶

This note captures the first-pass outreach plan for running PrivSyn on University of Virginia infrastructure.

Recommended Contact Order¶

1. UVA Research Computing
2. Start with the official support / consultation path.
3. Ask for a consultation covering public-facing services, Slurm or Kubernetes job submission, storage, and whether the existing container-services offering is the right landing zone.

4. Official support page: Research Computing Support Options

5. UVA ITS Help Desk

6. Email: 4help@virginia.edu
7. Use this when you need routing help for DNS, TLS, reverse proxy, SSO, or the right UVA infrastructure owner.

8. UVA publishes 4help@virginia.edu as a general IT support address in official support materials.

9. If you are in the School of Engineering

10. Email: Engr-IT@virginia.edu

11. This is the direct engineering IT help address and is worth copying when the service will be operated inside SEAS workflows.

12. If cloud-hosted university data is involved

13. UVA Information Security should be consulted early because their compliance team handles cloud-vendor review requirements for university data.
14. Official InfoSec overview: University Information Security

What To Ask¶

• Can UVA host a public HTTPS frontend/API for this application?
• Should the public service live on a VM, Kubernetes ingress, or another managed platform?
• Can a web service submit work to Slurm or a campus Kubernetes jobs service?
• Can the app register for UVA SSO through OIDC or SAML?
• What Postgres and object storage options are available?
• Does uploading user datasets trigger a security or compliance review?

Draft Email¶

Subject: Consultation request for UVA-hosted web service backed by research computing jobs

Hello,

I am planning a UVA-hosted deployment of a research web application for differentially private tabular data synthesis. The service has a browser-based frontend and a Python/FastAPI backend, and longer-running synthesis jobs should execute on UVA-managed research computing infrastructure rather than inside the web tier.

I would like guidance on the recommended UVA architecture for:

• a public or campus-accessible HTTPS endpoint for the web UI and API,
• secure submission of synthesis jobs to UVA Research Computing infrastructure,
• storage for uploaded datasets and generated artifacts,
• authentication through UVA SSO, and
• any required security or compliance review for handling uploaded datasets.

The application is containerized, and I am intentionally keeping the implementation modular so it can target cloud infrastructure or UVA-managed systems with the same codebase. If helpful, I can provide a short architecture summary and expected resource profile for typical synthesis jobs.

Could you point me to the right team or help arrange a short consultation?

Thank you,

[Your name] [Department / lab] [Computing ID]

Official References¶

• Research Computing support and consultation paths: rc.virginia.edu/support
• UVA Information Security overview and team contacts: security.virginia.edu/about-university-information-security

Follow-up After RC Response¶

• If RC leadership responds and asks for a deeper architecture discussion, send the follow-up draft in docs/uva-josh-followup-draft.md.
• Bring docs/uva-rc-architecture.md, docs/delivery-layers.md, and docs/library-cli.md to the meeting so the conversation stays concrete on permissions, operations boundaries, prototype scope, and which delivery layer RC is actually willing to support.