Verisign Labs

12061 Bluemont Way Reston 20190 USA VA ssheth@verisign.com https://www.verisignlabs.com/

Virginia Tech

220 Gilbert Street, RM 4303 Blacksburg 24060 USA VA tijay@vt.edu https://www.vt.edu/

NLnet Labs

Science Park 400 Amsterdam 1098 XH NL benno@nlnetlabs.nl https://nlnetlabs.nl

Internet Network Working Group Post-Quantum Cryptography DNS DNSSEC This document proposes a post-quantum cryptography (PQC) strategy for Domain Name System Security (DNSSEC) that includes two types of algorithms: one or more conservatively designed algorithms that are unlikely ever to need to be replaced, and one or more low-impact drop-in algorithms that are used the same way as a traditional signature algorithm. The conservatively designed algorithms can be used in a mode of operation that mitigates the operational impact of a large signature size. The combination provides both the routine performance of the low-impact algorithm and a resilient fallback to the conservatively designed choice. The draft outlines the strategy, provides recommendations for future testing and deployment, and highlights operational considerations in adopting PQC for DNSSEC.

Conventions Used in This Document The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 when, and only when, they appear in all capitals, as shown here.

Introduction DNSSEC provides data origin authentication for DNS resource records. Current algorithms, such as RSASHA256 (8) and ECDSA (13), are vulnerable to cryptanalytically capable quantum computers. While "harvest now/decrypt later" is not a concern for DNSSEC, as it is for some other protocols such as TLS, "trust now/forge later" is a concern for DNSSEC. Ensuring that signatures are valid and secure from inception until expiration is critical. This combined with the fact that standards bodies like the National Institute of Standards and Technology (NIST) are deprecating support for classical algorithms ensures that migration to post-quantum cryptography (PQC) is necessary. Unfortunately, migration with the large signature sizes introduce operational risks. This draft proposes a strategy deploying:

• One or more conservatively designed PQC algorithm in a mode mitigating large signature sizes.
• One or more low-impact drop-in PQC algorithm analogous to traditional DNSSEC signatures.

This dual-algorithm approach ensures routine performance and resilient fallback during PQC transition. This draft is intended as a contribution to ongoing algorithm updates and the algorithm lifecycle per drafts and

Post-Quantum DNSSEC Challenges

Operational Constraints DNS primarily runs over UDP, with packet sizes limited to a maximum of ~1232 bytes. Traditional signatures (e.g., RSASHA256, ECDSA) fit within this limit. PQC signatures (ML-DSA: 2420-4627 bytes, SLH-DSA: 7856-49856 bytes) exceed it, risking excessive TCP fallback, latency, and resolver performance degradation .

Deployment Cycles DNSSEC upgrades occur over years. Novel PQC algorithms may face uncertain adoption timelines, requiring fallback mechanisms. Some algorithms (e.g., SQIsign) impose verification overhead, slowing response times .

Proposed PQC Algorithm Diversity Strategy DNSSEC should deploy two types of PQC signature algorithms: Currently standardized post-quantum secure algorithms that provide cryptographic confidence and resilient fallback. Examples: SLH-DSA in Merkle Tree Ladder (MTL) mode , Falcon, XMSS, LMS. New algorithms such as the ones that remain under NIST onramp evaluation or under consideration by other standards bodies. These provide routine performance with minimal operational impact. They may leverage newer but less well-established mathematical concepts. Examples: MAYO, SNOVA.

Mode of Operation MTL mode signs a Merkle tree ladder rather than individual DNS responses, amortizing signature size across multiple responses . In DNSSEC, this reduces operational impact while maintaining security.

Alternatives and Considerations

• Conservative candidates: SLH-DSA, ML-DSA (possibly combined with traditional algorithms), Falcon, XMSS, LMS.
• Low-impact candidates: New algorithms such as the ones that remain under NIST onramp evaluation or under consideration by other standards bodies.
• Use of modes of operation (like MTL mode) to mitigate large signature sizes.

Recommended Next Steps

• Conduct hackathons testing multiple algorithms in BIND, NSD, and CoreDNS (see current progress in ).
• Measure latency, fallback rates, and resilience under adversarial conditions, including KeyTrap-style attacks .
• Research countermeasures against denial-of-service risks for MTL mode.

Current Community Efforts Several efforts are underway to implement, test, and discuss PQC algorithms in DNSSEC.

• IETF PQC DNSSEC Side Meeting -
• IETF 123 Hackathon - PQC DNSSEC Implementation
• IETF 122 Hackathon - PQC for DNSSEC - New Kids on the Block
• IETF 122 Hackathon - PQC DNSSEC Metrics with MTL Mode

IANA Considerations This document makes no requests of IANA. Future work may include registration of new DNSSEC algorithm codes for PQC algorithms.

Security Considerations The deployment of PQC algorithms strengthens DNSSEC against quantum attacks but introduces operational risks. Proper testing, fallback mechanisms, and mode-of-operation considerations are essential to avoid new vulnerabilities. Continued community participation in PQC DNSSEC research, in particular around low-impact drop-in algorithms, is essential to standarizing secure PQC DNSSEC solutions. Additional considerations will be described based on continued analysis and feedback.

References Normative References Informative References

Acknowledgements Thanks to Andrew Fregly for early contributions in promoting PQ DNSSEC and uniting the research community around a post-quantum research agenda.

Change Log

• 00: Initial draft of the document.