← Back to Learning Hub

DNS Records • Intermediate • 6 min read

CDS and CDNSKEY: Automating DNSSEC DS Updates at the Parent

How CDS and CDNSKEY records let a child zone signal a new DS to its parent automatically, enabling hands-off DNSSEC key rollovers under RFC 7344 and 8078.

The hardest part of a DNSSEC key rollover is not inside your own zone, it is the handoff to your parent. When you rotate your Key Signing Key, the DS record your registry publishes must be updated to match, and that step has traditionally meant logging into a registrar portal or filing a request by hand, often with bad timing that breaks validation. CDS and CDNSKEY records exist to automate exactly that handoff. They let your zone signal the DS it wants the parent to publish, so a supporting registry can pick up the change and update delegation on its own. This guide explains what the two records contain, how the parent consumes them, how they enable hands-off key rollovers, and the special signal that turns DNSSEC off safely.

Why the parent handoff is the weak point

DNSSEC builds a chain of trust from the root down. Your parent zone vouches for your Key Signing Key by publishing a DS record, a digest of that key, in the parent zone. When you roll the KSK, the old DS must give way to a new one at precisely the right moment, not before resolvers have your new key and not after they have dropped the old. Get the timing wrong and validating resolvers return SERVFAIL, which looks exactly like your domain disappearing. Our guide on the DNSSEC key rollover walks the timing in detail; CDS and CDNSKEY are what let a machine, rather than a human, carry out the parent side of that dance.

What CDS and CDNSKEY contain

RFC 7344, published in September 2014, defines both records. They are published in the child zone, signed like everything else, and simply express the child's wishes for its own delegation. CDS mirrors the DS record format, and CDNSKEY mirrors the DNSKEY format, so the parent has two equivalent ways to learn what it should publish.

RecordLives inMirrorsFields
DSParent zonen/a (the target)Key tag, algorithm, digest type, digest
CDSChild zoneDSKey tag, algorithm, digest type, digest
DNSKEYChild zonen/aFlags, protocol, algorithm, public key
CDNSKEYChild zoneDNSKEYFlags, protocol, algorithm, public key

A parent that supports CDS reads it and computes the DS directly. A parent that prefers CDNSKEY reads the published key and computes the digest itself. You can publish both; the parent uses whichever it supports. Because these records sit in your signed zone, the parent can verify they are authentic using the DNSSEC chain that already exists, which is what makes trusting them safe.

How the parent consumes them

The acceptance side is defined in RFC 8078, published in March 2017, which describes how a parent moves its DS to match a child's CDS or CDNSKEY. In practice a registry or registrar runs a scanner that periodically polls each signed child zone for these records. When it sees a CDS or CDNSKEY that differs from the DS currently published, it validates the records against the existing chain of trust, applies acceptance checks, and then updates the DS at the registry. No portal visit, no support ticket, no manual digest copying.

RFC 8078 also addresses the harder bootstrap case, establishing the very first DS for a zone that has no existing chain of trust to validate against. As of November 2026 newer work, including RFC 9615, refines that initial bootstrapping so a registry can accept a first DS with stronger assurances, but the day-to-day value of CDS and CDNSKEY remains the routine rollover once the chain is already in place.

Running a hands-off KSK rollover

With CDS and CDNSKEY in place, a KSK rollover becomes something your signer and your registry coordinate for you. The sequence looks like this.

  1. Confirm your registry or registrar actually scans for CDS or CDNSKEY; support is common but not universal as of November 2026, so verify before you rely on it.
  2. Generate the new KSK and publish it in your DNSKEY set alongside the old one, signing the set with both keys.
  3. Publish a CDS and CDNSKEY describing the new key, so the parent learns the DS you want it to add.
  4. Let the scanner pick up the change and update the DS at the registry; confirm with dig DS example.com +short that the new DS appears.
  5. Wait out the parent's DS TTL so resolvers see the new delegation, then retire the old KSK from your DNSKEY set and remove its CDS and CDNSKEY.

The delete signal: turning DNSSEC off

RFC 8078 also defines a special way to say the opposite, that you want DNSSEC removed. Publishing a CDS or CDNSKEY with the defined delete contents, a record whose fields are set to the reserved zero values rather than a real key digest, tells the parent to withdraw the DS entirely and return your zone to unsigned. This matters because going insecure incorrectly, by yanking keys while a DS still points at them, is itself a way to break a zone. The delete signal gives you a clean, standardized off switch that the parent can act on with the same automation it uses for rollovers.

A note on monitoring

Automation does not remove the need to watch. Confirm each DS change landed before you remove old keys, and alert on unexpected CDS or CDNSKEY content, since those records drive changes to your delegation and should only ever reflect keys you intend to publish.

Frequently asked questions

Do CDS and CDNSKEY update the parent on their own?

Not by themselves. They publish your intent in your own zone; a supporting registry or registrar must run a scanner that reads them and updates the DS. If your provider does not scan for them, the records sit there with no effect and you still update the DS manually.

Should I publish CDS, CDNSKEY, or both?

Publishing both is the safe choice. Some parents consume CDS and compute the DS, while others prefer CDNSKEY and compute the digest themselves. Offering both lets the parent use whichever it supports without you having to know which that is.

Is it safe to let a registry change my DS automatically?

Yes, because the records are part of your signed zone and the parent validates them against the existing chain of trust before acting. That cryptographic check is what lets the parent trust a machine-driven change rather than requiring a human to verify it.

How do I use these records to disable DNSSEC?

Publish a CDS or CDNSKEY with the special delete contents defined in RFC 8078, whose fields use the reserved zero values. A supporting parent reads that as a request to withdraw the DS and return your zone to unsigned in a clean, standardized way.

Why did nothing happen after I published CDS?

Most often the parent does not scan for it, the records failed the parent's validation, or the scanner has not run yet. Confirm your registrar supports CDS or CDNSKEY, make sure the records are correctly signed, and allow time for the next scan before troubleshooting further.

CDS and CDNSKEY turn the riskiest manual step in DNSSEC, the parent-side DS update, into something your signer and registry handle together, which is what finally makes routine key rollovers practical. To confirm your domain is signed and its delegation lines up, run it through the Valla DNS domain health check.

Troubleshooting box

If results look inconsistent, compare authoritative nameservers first, then recursive resolvers by region. Capture snapshots every 10 minutes for deterministic incident timelines.

Try VallaDNS free →

Intermediate • 6 min

How to Find a DNS Record's TTL (Current Value)

Find the TTL currently set on a DNS record: read the TTL column in dig and nslookup, tell the authoritative TTL from a resolver's countdown, and confirm a change.

Intermediate • 7 min

How to Check If Nameservers Have Propagated

Nameserver changes live at the parent TLD and follow the registry NS TTL, not your zone. Check the parent delegation and confirm the new NS set worldwide.