Dear colleagues,

Some of you will remember me from my years on the NCC's Database team. This time I am writing from the other side of the table, as a resource holder, with an idea I think is worth the WG's consideration.

The ask in one paragraph: let a resource holder optionally tell the RIPE NCC's RDAP service to refer queries for the more-specifics of their space to an RDAP server the holder runs, and let clients optionally follow that referral. Nothing is mandatory on either side. A holder that offers nothing behaves exactly as today, and a client that does not follow simply sees the covering object, as today. It is the RDAP analogue of the reverse DNS delegation we already do with domain: and nserver:.

Why this matters: an RDAP walk that starts from the IANA bootstrap (RFC 9224) and reaches https://rdap.db.ripe.net/ stops at the covering allocation. If the holder maintains a large, dynamic set of more-specifics and runs a conformant RDAP server for them, there is no way for the RIPE response to hand the client down to it.

The last hop is undiscoverable unless you already know the holder's server exists. RDAP is meant to be walked, not memorised, so that hop ought to be discoverable too. And I should say this is not theoretical for me: we publish per-assignment registration data for our own space from our own RDAP server, and running into exactly this gap is what prompted this mail.

The real prize, I think, is registration data hygiene. A holder with a lot of assignment activity has two poor options today. Register nothing below the allocation, which leaves coarse, unhelpful data for abuse handling, research and routing. Or bulk-load large numbers of assignment objects that start drifting out of date the day they land, which means stale data and central database bloat. An optional referral to the place where the data is actually maintained gives the community accurate, current, granular registration data served from the source, keeps the central database lean, and preserves an unbroken, machine-walkable chain from the bootstrap to the exact object. It rewards holders for publishing good, live data rather than penalising them for it.

The principle is not new for us, only the object of it is. Registrants already maintain their own sub-allocation and assignment data directly. Reverse DNS for the same space is already delegated to the holder's nameservers. And the WG has handled referral concepts before: we retired referral-by: at RIPE 68, though that was an authentication-time mechanism that had stopped earning its keep, which is a different animal from a data-locality referral, and domain referrals go back to the late-90s threads on this list. What has changed is that the standards now express this cleanly, so there is little to invent:

* RFC 7480 section 5.2 already lets a server that is not authoritative answer with a 30x and a Location header. The redirect primitive exists.

* RFC 9910 (RDAP RIR Search, Standards Track, January 2026) defines the rdap-up and rdap-down relations for navigating the number hierarchy, and rdap.db.ripe.net already emits these today. An rdap-down or related link to the holder's base URL is a natural carrier for the referral, with the holder's server publishing rdap-up back to the RIPE object so the hierarchy stays consistent and loop-safe.

* regext is actively working on explicit RDAP referrals right now (draft-ietf-regext-rdap-referrals, "Explicit RDAP Redirects", -04, July 2026): a client can ask to be redirected straight to a related RDAP record, and the draft's own examples already cover IP space. It is early, but it is the natural vehicle for the down-referral, and IP-space input from this WG would land at exactly the right time.

So the genuinely new part is a member-facing, opt-in way to register that base URL (scoped strictly to the holder's own space, TLS only), plus the NCC emitting an already-standardised referral.

Where I would love the WG's steer: is there appetite for an optional, member-configurable RDAP referral for more-specifics, built on the referrals work above rather than anything bespoke? If there is, I am glad to take it forward as a concrete RIPE NCC feature request and/or a slot at the next db-wg session, and to keep it aligned with the NRO and regext so it stays consistent across RIRs. I can also prototype the down-referral against a live, conformant server that already publishes the up link back to the RIPE object.

All the best,
Kaveh