What to ask a domain portfolio vendor before a premium name carries production traffic

A premium domain is not one asset but a chain: registry, registrar account, authoritative DNS, signing keys and the credentials that move all of them. Any link can fail on its own, and the name stays only as available as its weakest link. The vendor questions worth asking are the ones that show which links the vendor actually controls and what you inherit when one of them breaks.

Who holds custody during the transfer

Ask where the name sits between payment and the moment it lands in your registrar account, and who can move it during that window. The risk is a custody gap: an escrow that releases funds before the registrar change is confirmed, or a seller account that still holds the authorization code after you believe you own the asset. Locks set from a registrar dashboard can be lifted by anyone who phishes that dashboard, so ask whether the registry offers server side locks that need an out of band confirmation to remove. A vendor who cannot name the registrar of record for every name in the portfolio is telling you the inventory is spread across accounts nobody audits.

How many operators answer for the zone

One DNS provider is one outage away from a dark domain, however premium the name. Ask whether the zone can be served by two authoritative operators on independent networks, and how records stay in sync, whether by zone transfer or by a provisioning pipeline that writes to both APIs. Then ask about TTL, because it decides how long a mistake survives: with a TTL of 3600 seconds on the records you plan to cut over, a wrong answer cached by resolvers can persist for an hour after you fix it. Lowering it before a migration shortens that window at the cost of more query load, a trade the vendor should schedule rather than improvise.

The DNSSEC handover is where names go dark

DNS security extensions such as DNSSEC protect the answer but introduce a failure mode that most lab checks miss. If the parent zone still publishes a DS record pointing at the old provider's key while the new provider signs with a different key, or not at all, validating resolvers return SERVFAIL and the name disappears for their users. Networks that do not validate keep resolving normally, so the domain looks healthy from the office and broken from a large share of public and mobile resolvers, an estimate, but one incident reports keep confirming. The decision rule: never accept a plan that changes the DS record and the nameservers in the same window; either the DS is withdrawn and the parent TTL expires before the move, or both operators publish each other's keys for the duration of the handover.

What survives the vendor

Ask how renewals work if the vendor stops trading, because unless you can renew each name directly with its registrar and confirm the expiry date yourself through RDAP, the cheapest part of the chain, a billing relationship, becomes its single point of failure.

How much of this can be verified without lab access?

Most of it, from public data. RDAP exposes status flags, registrar of record and expiry; a DNSSEC chain analyzer shows whether the DS at the parent matches a published key; querying each authoritative server directly shows whether the operators agree on the SOA serial. Public data cannot show who holds the account credentials, so that answer still has to come in writing.

Back to CategoryMaker · Domains

← Back to CategoryMaker
Engineering & Architecture Brief

What to ask a domain portfolio vendor before a premium name carries production traffic

A premium domain is not one asset but a chain: registry, registrar account, authoritative DNS, signing keys and the credentials that move all of them. Any link can fail on its own, and the name stays only as available as its weakest link. The vendor questions worth asking are the ones that show which links the vendor actually controls and what you inherit when one of them breaks.

Who holds custody during the transfer

Ask where the name sits between payment and the moment it lands in your registrar account, and who can move it during that window. The risk is a custody gap: an escrow that releases funds before the registrar change is confirmed, or a seller account that still holds the authorization code after you believe you own the asset. Locks set from a registrar dashboard can be lifted by anyone who phishes that dashboard, so ask whether the registry offers server side locks that need an out of band confirmation to remove. A vendor who cannot name the registrar of record for every name in the portfolio is telling you the inventory is spread across accounts nobody audits.

How many operators answer for the zone

One DNS provider is one outage away from a dark domain, however premium the name. Ask whether the zone can be served by two authoritative operators on independent networks, and how records stay in sync, whether by zone transfer or by a provisioning pipeline that writes to both APIs. Then ask about TTL, because it decides how long a mistake survives: with a TTL of 3600 seconds on the records you plan to cut over, a wrong answer cached by resolvers can persist for an hour after you fix it. Lowering it before a migration shortens that window at the cost of more query load, a trade the vendor should schedule rather than improvise.

The DNSSEC handover is where names go dark

DNS security extensions such as DNSSEC protect the answer but introduce a failure mode that most lab checks miss. If the parent zone still publishes a DS record pointing at the old provider's key while the new provider signs with a different key, or not at all, validating resolvers return SERVFAIL and the name disappears for their users. Networks that do not validate keep resolving normally, so the domain looks healthy from the office and broken from a large share of public and mobile resolvers, an estimate, but one incident reports keep confirming. The decision rule: never accept a plan that changes the DS record and the nameservers in the same window; either the DS is withdrawn and the parent TTL expires before the move, or both operators publish each other's keys for the duration of the handover.

What survives the vendor

Ask how renewals work if the vendor stops trading, because unless you can renew each name directly with its registrar and confirm the expiry date yourself through RDAP, the cheapest part of the chain, a billing relationship, becomes its single point of failure.

Technical FAQ: How much of this can be verified without lab access?

Most of it, from public data. RDAP exposes status flags, registrar of record and expiry; a DNSSEC chain analyzer shows whether the DS at the parent matches a published key; querying each authoritative server directly shows whether the operators agree on the SOA serial. Public data cannot show who holds the account credentials, so that answer still has to come in writing.