NIST rewrote its DNS deployment guidance in March 2026, and teams that built their DNS controls on the 2013 edition now need to show how they meet a much broader standard that treats DNS as a security control, not just infrastructure. SP 800-81 Rev. 3, the Secure Domain Name System (DNS) Deployment Guide, was published as final on 19 March 2026 and supersedes SP 800-81-2. It adds protective DNS, encrypted DNS and DNS logging to the earlier focus on DNSSEC and server hardening, and frames all of it within zero trust and defence in depth. This guide maps the publication section by section into actionable settings, implementation steps and assessment evidence for network, DNS and security operations teams and the auditors who review them. For the wider DNS threat picture see the DNS Security Guide.
Key takeaways
- Five headline recommendations. Use protective DNS wherever feasible, encrypt internal and external DNS traffic wherever feasible, deploy DNSSEC, use dedicated DNS servers, and follow the technical deployment guidance.
- DNS is a policy enforcement point. The guide places DNS inside zero trust: it should not be implicitly trusted and must not become a path around other controls.
- Control the resolvers clients use. Block outbound port 53 and unauthorised DoT, DoH and DoQ, and use device management so endpoints only use approved encrypted resolvers.
- Concrete DNSSEC parameters. Prefer ECDSA or Ed448 over RSA, keep key lifetimes to one to three years, keep DNSKEY signature validity short, prefer NSEC, and automate trust anchor rollover.
- Logging is now a requirement area. Keep current and historical DNS logs, always log queries classified as malicious, and correlate with DHCP and asset data.
What it is and who it applies to
SP 800-81 Rev. 3 is NIST guidance written by Scott Rose (NIST) with Cricket Liu and Ross Gibson (Infoblox). The draft appeared on 10 April 2025 and the final on 19 March 2026; NIST has also published a spreadsheet of potential errata (planning note of 10 July 2026), so check the CSRC page for updates. It is a guide, not a law: US federal agencies use it alongside SP 800-53 controls SC-20, SC-21 and SC-22, and any organisation can adopt it as a DNS baseline.
| Item | SP 800-81-2 (2013) | SP 800-81 Rev. 3 (2026) |
|---|---|---|
| Framing | Securing the DNS protocol and servers | DNS as part of security strategy: zero trust, defence in depth, cyber resilience |
| Protective DNS | Not a focus | Section 2.1: resolver-based blocking, threat intelligence, response policy zones |
| Encrypted DNS | Not covered | Section 4.2.1: DoT, DoH and DoQ, plus control of public resolvers |
| DNSSEC | Core topic | Updated: algorithm preference, lifetimes, NSEC versus NSEC3, compact denial, post-quantum note |
| Logging | Limited | Explicit DFIR and logging guidance |
| Audience scope | Federal IT | Also OT, IoT and critical infrastructure use cases (section 1.2.1) |
Who should use it: anyone operating authoritative servers for public or internal zones, recursive resolvers or forwarders, protective DNS services, or managing endpoint DNS settings. Small organisations relying on a managed DNS provider and a protective DNS service still own the resolver policy, logging and domain hygiene sections.
Structure and key requirements
| Section | Topic | Key recommendations |
|---|---|---|
| 2.1 | Protective DNS | Use DNS as a blocking and enforcement layer; hybrid cloud and on-premises where feasible so protection survives a cloud outage; analyse all record types (MX, SOA, TXT), not only A and AAAA; choose RPZ placement by visibility and cost; update feeds automatically; return NXDOMAIN or similar for blocked names |
| 2.2 | Protecting the protocol | Harden authoritative and recursive services; combine DNSSEC with encryption and TSIG; continuously monitor public domains and impersonation |
| 2.3 | Service and infrastructure | Dedicated DNS hosts; remove DNS server software from hosts that are not name servers; at least two authoritative servers on different network segments and sites; hidden primary; SIEM and SOAR integration |
| 3.1 to 3.4 | Authoritative operations | Restrict zone transfers with ACLs and TSIG or SIG(0); consider ZONEMD; TLS or network-layer confidentiality for transfers; dynamic updates only to the primary from authorised senders; NOTIFY only from the primary or authorised secondaries; check for lame delegations |
| 3.5 and 3.6 | Information leakage and domain integrity | Exclude HINFO, RP and LOC from internet-facing zones; review TXT records; remove dangling CNAMEs and stale delegations; monitor look-alike registrations; park retired delegations |
| 3.7 | TTLs | 30 minutes to one day in most cases; never 0; avoid 30 seconds or less for signed data; TTLs shorter than RRSIG validity |
| 3.8 | DNSSEC signing | See the DNSSEC table below |
| 4.2 | Recursive services | Encrypted DNS; restrict public providers; QNAME minimisation (RFC 7816) as an option; detect DNS exfiltration; DNSSEC validation with the root trust anchor; RFC 5011 automated rollover |
| 4.3 | Resolver operations | Block inbound TCP and UDP 53 to recursive servers; recursion only for internal clients; primary and backup time sources using NTS (RFC 8915) |
| 5 | Stub resolvers and mobile hosts | At least two recursive servers per stub; consider anycast; block outbound DNS except to known resolvers; EMM to enforce resolver settings; mobile hosts use a tunnel or an approved secure resolver |
DNSSEC parameters in Rev. 3
| Setting | Rev. 3 recommendation |
|---|---|
| Algorithm | ECDSA or Ed448 in preference to RSA where possible, for smaller responses |
| Key strength | SP 800-57 strengths: 112 bits now, 128 bits from 2030 |
| Key lifetime | Maximum of one to three years |
| Key protection | HSM for important zones, usually for the key signing key |
| Rollovers | Automate zone signing key rollovers; coordinate key signing key rollovers with the parent |
| DNSKEY RRSIG validity | Short, for example five to seven days |
| Denial of existence | Prefer NSEC; if NSEC3 is needed, tune parameters to limit denial-of-service risk and plan to move away from SHA-1; compact denial of existence is a work in progress |
| Algorithm migration | Follow RFC 6781 ordering: new signatures before new DNSKEY and DS; remove DNSKEYs before their signatures |
| Internal zones | Do not sign unless required |
| Post-quantum | Algorithms not yet specified for DNSSEC; plan to migrate when they are |
For deployment detail see the DNSSEC Deployment Guide, the Protective DNS Guide and the Encrypted DNS Enterprise Guide.
Encrypted DNS and control of public resolvers
- Protocols and ports: DoT on TCP 853, DoH on TCP and UDP 443, DoQ on UDP 853. Route IoT devices through a forwarder, and keep local resolvers in the path so applications cannot bypass enterprise policy.
- Restricting public providers (4.2.2): block outbound UDP and TCP 53 except from authorised forwarders; block unauthorised DoT on TCP 853 and DoQ on UDP 853; block known public DoH services with RPZ entries for their names plus firewall rules for dedicated DoH addresses; use MDM to stop users configuring unapproved encrypted DNS.
- Costs: plan for CPU overhead and harder troubleshooting, and configure a server certificate on each encrypted DNS server.
Identity and NHI requirements
Several Rev. 3 controls depend on machine credentials and privileged accounts that are easy to overlook in a DNS review.
| Credential or identity | Where Rev. 3 relies on it | What to do |
|---|---|---|
| TSIG keys | Zone transfers (3.1), dynamic updates (3.3), authentication alongside encryption (2.2.2) | Treat as shared secrets: one key per server pair or updater, stored in a secrets manager, rotated on a schedule and on staff or provider change |
| Dynamic update clients | Updates only from authorised senders (3.3) | Restrict to named DHCP servers or service accounts; use dedicated hidden primaries for dynamic zones. See the Windows Server DNS and DHCP Security Guide |
| DNSSEC signing keys | HSM for key signing keys (3.8.1) | Restrict HSM roles, log every signing operation, and document key ceremonies |
| Encrypted DNS server certificates | Certificate on each DoT, DoH or DoQ server (4.2.1.3) | Inventory and automate renewal like any other TLS certificate |
| DNS provider and registrar accounts, API tokens | Domain integrity and hygiene (2.2.3, 3.6) | Phishing-resistant MFA for people, scoped and rotated tokens for automation such as infrastructure as code and ACME DNS-01 challenges. See the Authoritative DNS and Registrar Account Security Guide |
| DNS administrators | Dedicated, hardened DNS hosts (2.3.1) | Small, named admin group with privileged access controls and logged changes |
A DNS API token with write access to a public zone can redirect mail, web and certificate validation for a whole domain, so it should be governed as a privileged non-human identity. See the Secrets Management Guide.
Implementation steps
- Inventory DNS roles: list every authoritative server, hidden primary, secondary, recursive resolver, forwarder, protective DNS service and DNS provider account, with owner and location. Remove DNS server software from hosts that are not meant to be name servers.
- Gap assessment: score each role against the sections in the table above and record evidence. Prioritise resolver control, logging and zone transfer authentication, which are often the largest gaps.
- Protective DNS and logging: put all client resolution through protective DNS with automatic feed updates, start logging (DNSTAP where supported) and send logs to the SIEM. Pilot RPZ blocking in log-only mode for one to two weeks to find false positives; rolling back is switching the policy back to log-only.
- Restrict egress: block outbound 53 and unauthorised DoT, DoH and DoQ in stages, first alerting on violations, then blocking for a pilot group, then everywhere. Keep a documented exception path for devices that must use another resolver.
- Encrypt: enable encrypted DNS from managed endpoints to enterprise resolvers using MDM or EMM profiles, and between forwarders and upstream services.
- Authoritative hardening: hidden primary, two or more secondaries on separate segments and sites, TSIG on transfers and updates, NOTIFY restrictions, removal of HINFO, RP and LOC records, TTL review, and a dangling record clean-up. See the Subdomain Takeover and Dangling DNS Guide.
- DNSSEC: sign public zones with ECDSA or Ed448, automate zone signing key rollovers, protect key signing keys in an HSM, enable validation on resolvers with RFC 5011 trust anchor management, and configure NTS time sources. Test rollovers in a staging zone first; the rollback for a failed signing change is restoring the previous DNSKEY set and DS record within the TTL window, which is why short, planned TTLs matter.
- Operate: review logs, feeds and zone content continuously, and re-assess against the errata as NIST updates the publication.
Evidence and assessment
| Requirement area | Evidence an assessor should see |
|---|---|
| Protective DNS | Service configuration, coverage report showing all clients use it, feed update logs, sample blocked events in the SIEM |
| Resolver control | Firewall rules for 53, 853 and known DoH addresses; RPZ entries for DoH names; MDM resolver profiles; reports of blocked bypass attempts |
| Dedicated servers and resilience | Architecture diagram with hidden primary and secondary locations; host build standard |
| Zone transfers and updates | Server configuration showing ACLs and TSIG; key inventory with rotation dates |
| Zone content and hygiene | Last dangling record review, look-alike domain monitoring output, TTL policy |
| DNSSEC | Algorithm and key lifetimes, HSM configuration, rollover records, validation enabled on resolvers, DS records at parent |
| Logging | Retention settings, sample DNSTAP or query logs, DHCP correlation, alert rules for malicious-classified queries |
| Time | NTS configuration with primary and backup sources on validating and encrypted DNS servers |
Detection and measurement: track the share of clients resolving through protective DNS (target 100% of managed devices), blocked attempts to reach unauthorised resolvers per week (should fall after rollout), zone transfers without TSIG (zero), DNSSEC validation failures and signing errors (investigated within one business day), days since the last dangling record review (under 90), and the proportion of DNS logs available for the agreed retention period. Detection content should cover DNS tunnelling and exfiltration as section 4.2.4 describes: abnormal query volume, changing patterns, high-entropy query names and queries for known malicious domains (ATT&CK T1071.004 DNS, T1572 Protocol Tunneling, T1048 Exfiltration Over Alternative Protocol, T1568.002 Domain Generation Algorithms). See the DNS Logging and Detection Guide.
Mapping to other frameworks
| SP 800-81r3 area | NIST SP 800-53 Rev. 5 | NIST CSF 2.0 | CIS Controls v8.1 | ISO/IEC 27001:2022 | MITRE D3FEND |
|---|---|---|---|---|---|
| Authoritative integrity and DNSSEC signing | SC-20 | PR.DS-02 | 12 | A.8.20 | n/a |
| Recursive validation and resolver hardening | SC-21 | PR.IR-01 | 9.2 | A.8.20 | n/a |
| Dedicated, redundant DNS architecture | SC-22 | PR.IR-01 | 12 | A.8.20 | n/a |
| Protective DNS filtering | SC-7, SI-4 | DE.CM-01 | 9.2 | A.8.23 | D3-DNSDL DNS Denylisting |
| DNS logging and exfiltration detection | AU-2, SI-4 | PR.PS-04, DE.CM-01 | 8.6 | A.8.15, A.8.16 | D3-DNSTA DNS Traffic Analysis |
| Secure time sources | SC-45 | PR.PS-01 | 8.4 | A.8.17 | n/a |
Common pitfalls
- Protective DNS that clients bypass: filtering is configured, but browsers and apps use public DoH and nothing blocks port 853 or known DoH endpoints.
- A records only: protective DNS inspects A and AAAA answers while MX, TXT and other records used for command and control go unexamined.
- Unauthenticated zone transfers: ACLs by IP only, or transfers open to any secondary, leaking full internal zone contents.
- Manual DNSSEC rollovers: key rollovers done by hand, years apart, by someone who has left; signatures then expire and the zone fails validation.
- TTL extremes: TTLs of zero or a few seconds on signed data, or TTLs longer than signature validity, causing load and validation failures.
- Logs without context: query logs kept but not correlated with DHCP and asset data, so an alert names an IP address nobody can map to a device.
Practitioner checklist
- Inventory every DNS role and provider account, and remove DNS server software from hosts that are not name servers.
- Route all managed clients through protective DNS that inspects all record types and updates feeds automatically.
- Block outbound port 53 except from forwarders, and block unauthorised DoT, DoH and DoQ.
- Enforce approved encrypted resolvers on endpoints and mobile devices through MDM or EMM.
- Run a hidden primary with at least two secondaries on separate segments and sites.
- Authenticate zone transfers and dynamic updates with TSIG, and manage TSIG keys as secrets.
- Sign public zones with ECDSA or Ed448, automate rollovers and keep key signing keys in an HSM.
- Enable DNSSEC validation with RFC 5011 trust anchor management and NTS time sources.
- Log DNS with DFIR retention, correlate with DHCP and assets, and alert on tunnelling indicators.
- Review zones for dangling records, unneeded record types and look-alike domains at least quarterly.
Standards and references
- NIST: SP 800-81 Rev. 3 Secure Domain Name System (DNS) Deployment Guide
- NIST: SP 800-53 Rev. 5 (SC-20, SC-21, SC-22)
- NIST: Cybersecurity Framework 2.0
- IETF: RFC 6781 DNSSEC Operational Practices, Version 2
- IETF: RFC 5011 Automated Updates of DNS Security (DNSSEC) Trust Anchors
- MITRE D3FEND: DNS Traffic Analysis
Related NHI Mgmt Group resources: DNS Security Guide · Protective DNS Guide · DNSSEC Deployment Guide · DNS Hijack Playbook · Encrypted DNS Enterprise Guide