UCPH self-hosting

Living migration table from EU cloud to on-premises

The AIPLA prototype runs on two layers: the AI Protocol Platform template (application) and Multivac (infrastructure — model access, Ollama hosting, managed Postgres, Cloud Storage). The template is cloud-agnostic and portable as-is. The migration target is to swap the Multivac infrastructure layer for UCPH-managed equivalents while keeping the template unchanged.

This page captures the component-by-component swap and what UCPH IT would need to provide. Half of the final handover package.

Status: stub. To be populated as the architecture decisions on the Architecture page solidify and as UCPH IT responds with hosting capabilities.

Migration table

Cloud component Role in prototype UCPH equivalent Migration effort Notes
Vertex AI (Gemini) LLM + multimodal Ollama on UCPH GPU server Medium — depends on local-readiness fraction per task class Capability gap depends on which local models are at the floor; see Evaluation
Cloud Run Backend hosting UCPH Kubernetes / VM Low Containerised backend portable as-is
Cloud SQL (Postgres + pgvector) Application DB + RAG vectors UCPH-managed Postgres Low pg_dump → restore; pgvector extension widely available
Cloud Storage Uploaded resources (PDFs, images, CSVs) UCPH NFS or S3-compatible Low Path abstraction in app
Firestore (if used) TBD TBD TBD Prefer not to use — see ADR-005
Cloud Logging Operational logs UCPH ELK or similar Low Standard log shipping
Vertex AI Vector Search (if used) RAG vectors pgvector or self-hosted Chroma Low if not used; see ADR-005 Lean is to avoid in favour of pgvector
GCP IAM Service auth UCPH SSO / institutional IDs TBD Depends on student-auth decision in ADR-010

Outstanding questions for UCPH IT

(JB has already pinged them; awaiting response. Populate as answers come in.)

  • Do they have GPU-capable hosting for inference? Spec?
  • What’s the institutional Kubernetes / container hosting story?
  • Do they offer managed Postgres (with pgvector)?
  • What’s the policy on cloud-to-on-prem migration timelines for live research projects?
  • Is there an existing pattern for student-facing services that already meets data-protection review? (We may be able to piggyback rather than green-field this.)

When to trigger migration

The migration becomes worth doing when:

  1. UCPH IT confirms hosting availability with timeline
  2. Local-readiness fraction (per Evaluation) reaches a threshold per task class — likely 70–80% as a working bar
  3. Steady-state usage justifies fixed infrastructure cost over per-API spend

Until then, GCP EU is the right place to be.