What research data infrastructure actually requires

There’s a tendency to conflate “we have storage” with “we have data infrastructure.”

Storage is necessary but not sufficient. Real research data infrastructure includes:

• Curation standards that make data findable and reusable years later
• Governance models that handle access, privacy, and compliance
• Persistent identifiers (DOIs) that survive URL rot
• Sufficient metadata to reproduce what was done

Most institutions have solved storage. Very few have solved the rest.

At UCLA, building Dataverse as a publishing platform — not just a repository — forced us to address all of these. That’s the difference between infrastructure and plumbing.