]> Verisign

12061 Bluemont Way Reston VA 20190 +1 703 948-3200 dwessels@verisign.com https://verisign.com

Verisign

12061 Bluemont Way Reston VA 20190 +1 703 948-3200 pbarber@verisign.com https://verisign.com

Amazon

matweinb@amazon.com https://amazon.com

Google

1600 Amphitheatre Parkway Mountain View CA 94043 warren@kumari.net

USC/ISI

P.O. Box 382 Davis CA 95617 ietf@hardakers.net

General Internet Engineering Task Force DNS DNSSEC Checksum Hash Zone Transfer This document describes a protocol and new DNS Resource Record that provides a cryptographic message digest over DNS zone data at rest. The &RRNAME; Resource Record conveys the digest data in the zone itself. When used in combination with DNSSEC, &RRNAME; allows recipients to verify the zone contents for data integrity and origin authenticity. This provides assurance that received zone data matches published data, regardless of how the zone data has been transmitted and received. When used without DNSSEC, &RRNAME; functions as a checksum, guarding only against unintentional changes. &RRNAME; does not replace DNSSEC. Whereas DNSSEC protects individual RRSets (DNS data with fine granularity), &RRNAME; protects a zone's data as a whole, whether consumed by authoritative name servers, recursive name servers, or any other applications. As specified herein, &RRNAME; is impractical for large, dynamic zones due to the time and resources required for digest calculation. However, The &RRNAME; record is extensible so that new digest schemes may be added in the future to support large, dynamic zones.

In the DNS, a zone is the collection of authoritative resource records (RRs) sharing a common origin (). Zones are often stored as files in the so-called master file format . Zones are generally distributed among name servers using the AXFR (zone transfer ), and IXFR (incremental zone transfer ) protocols. They can also be distributed outside of the DNS, with any file transfer protocol such as FTP, HTTP, and rsync, or even as email attachments. Currently, there is no standard way to compute a hash or message digest for a stand-alone zone. This document specifies an RR type that provides a cryptographic message digest of the data in a zone. It allows a receiver of the zone to verify the zone's integrity and authenticity when used in combination with DNSSEC. The digest RR is a part of the zone itself, allowing verification of the zone, no matter how it is transmitted. The digest uses the wire format of zone data in a canonical ordering. Thus, it is independent of presentation format, such as whitespace, capitalization, and comments. This specification is OPTIONAL to implement by both publishers and consumers of zone data.

The primary motivation for this protocol enhancement is the desire to verify the data integrity and origin authenticity of a stand-alone zone, regardless of how it is transmitted. A consumer of zone data should be able to verify that it is as-published by the zone operator. Note, however, that integrity and authenticity can only be assured when the zone is signed. DNSSEC provides three strong security guarantees relevant to this protocol: whether or not to expect DNSSEC records in the zone, whether or not to expect a &RRNAME; record in a signed zone, and whether or not the &RRNAME; record has been altered since it was signed. A secondary motivation is to provide the equivalent of a checksum, allowing a zone recipient to check for unintended changes and operational errors, such as accidental truncation.

One approach to preventing data tampering and corruption is to secure the distribution channel. The DNS has a number of features that are already used for channel security. Perhaps the most widely used is DNS transaction signatures (TSIG ). TSIG uses shared secret keys and a message digest to protect individual query and response messages. It is generally used to authenticate and validate UPDATE , AXFR , and IXFR messages. DNS Request and Transaction Signatures (SIG(0) ) is another protocol extension that authenticates individual DNS transactions. Whereas SIG records normally cover specific RR types, SIG(0) is used to sign an entire DNS message. Unlike TSIG, SIG(0) uses public key cryptography rather than shared secrets. The Transport Layer Security protocol suite also provides channel security. The DPRIVE working group is in the process of specifying DNS Zone Transfer-over-TLS . One can also easily imagine the distribution of zones over HTTPS-enabled web servers, as well as DNS-over-HTTPS . Unfortunately, the protections provided by these channel security techniques are (in practice) ephemeral and are not retained after the data transfer is complete. They ensure that the client receives the data from the expected server, and that the data sent by the server is not modified during transmission. However, they do not guarantee that the server transmits the data as originally published, and do not provide any methods to verify data that is read after transmission is complete. For example, a name server loading saved zone data upon restart cannot guarantee that the on-disk data has not been modified. Such modification could be the result of an accidental corruption of the file, or perhaps an incompletely saved file . For these reasons, it is preferable to protect the integrity of the data itself. Why not simply rely on DNSSEC, which provides certain data security guarantees? For zones that are signed, a recipient could validate all of the signed RRSets. Additionally, denial-of-existence records prove that RRSets have not been added or removed. However, delegations (non-apex NS records) are not signed by DNSSEC, and neither are any glue records. &RRNAME; protects the integrity of delegation, glue, and other records that are not otherwise covered by DNSSEC. Furthermore, zones that employ NSEC3 with opt-out are susceptible to the removal or addition of names between the signed nodes. Whereas DNSSEC primarily protects consumers of DNS response messages, this protocol protects consumers of zones. There are existing tools and protocols that provide data security, such as OpenPGP and S/MIME . In fact, the internic.net site publishes PGP signatures alongside the root zone and other files available there. However, this is a detached signature with no strong association to the corresponding zone file other than its timestamp. Non-detached signatures are, of course, possible, but these necessarily change the format of the file being distributed; a zone signed with OpenPGP or S/MIME no longer looks like a DNS zone and could not directly be loaded into a name server. Once loaded the signature data is lost, so it cannot be further propagated. It seems the desire for data security in DNS zones was envisioned as far back as 1997. is an obsoleted specification of the first generation DNSSEC Security Extensions. It describes a zone transfer signature, identified as the AXFR SIG, which is similar to the technique proposed by this document. That is, it proposes ordering all (signed) RRSets in a zone, hashing their contents, and then signing the zone hash. The AXFR SIG is described only for use during zone transfers. It did not postulate the need to validate zone data distributed outside of the DNS. Furthermore, its successor, , omits the AXFR SIG, while at the same time introducing an IXFR SIG.

This document specifies a new Resource Record type to convey a message digest of the content of a zone. The digest is calculated at the time of zone publication. If the zone is signed with DNSSEC, any modifications of the digest can be detected. The procedures for digest calculation and DNSSEC signing are similar. Both require data to be processed in a well-defined order and format. It may be possible to perform DNSSEC signing and digest calculation in parallel. The zone digest is designed to be used on zones that have infrequent updates. As specified herein, the digest is re-calculated over the entire zone content each time the zone is updated. This specification does not provide an efficient mechanism for updating the digest on incremental updates of zone data. It is, however, extensible so that future schemes may be defined to support efficient incremental digest updates. It is expected that verification of a zone digest will be implemented in name server software. That is, a name server can verify the zone data it was given and refuse to serve a zone which fails verification. For signed zones, the name server needs a trust anchor to perform DNSSEC validation. For signed non-root zones, the name server may need to send queries to validate a chain of trust. Digest verification could also be performed externally.

The root zone is one of the most widely distributed DNS zone on the Internet, served by more than 1000 separate instances at the time of this writing. Additionally, many organizations configure their own name servers to serve the root zone locally. Reasons for doing so include privacy and reduced access time. describes one way to do this. As the root zone spreads beyond its traditional deployment boundaries, the verification of the completeness of the zone contents becomes more important.

Since its very early days, the developers of the DNS recognized the importance of secondary name servers and service diversity. However, modern DNS service has complex provisioning which includes multiple third-party providers () and hundreds of anycast instances (). Instead of a simple primary-to-secondary zone distribution system, today it is possible to have multiple levels, multiple parties, and multiple protocols involved in the distribution of zone data. This complexity introduces new places for problems to arise. The zone digest protects the integrity of data that flows through such systems.

A Response Policy Zone (RPZ) is "a mechanism to introduce a customized policy in Domain Name System servers, so that recursive resolvers return possibly modified results" . The policy information is carried inside specially constructed DNS zones. A number of companies provide RPZ feeds, which are consumed by name server and firewall products. While RPZ zones can be signed with DNSSEC, the data is not queried directly, and would not be subject to DNSSEC validation.

ICANN operates the Centralized Zone Data Service , which is a repository of top-level domain zone files. Users that have been granted access are then able to download zone data. Adding a zone digest to these would provide CZDS users with assurances that the data has not been modified between origination and retrieval. Note that &RRNAME; could be added to zone data supplied to CZDS without requiring it to be present in the zone data served by production name servers, since the digest is inherently attached to the specific copy of the zone.

Since the zone digest calculation does not depend on presentation format, it could be used to compare multiple copies of a zone received from different sources, or copies generated by different processes. In this case, it serves as a checksum and can be useful even for unsigned zones.

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. The terms Private Use, Reserved, Unassigned, and Specification Required are to be interpreted as defined in .

This section describes the &RRNAME; Resource Record, including its fields, wire format, and presentation format. The Type value for the &RRNAME; RR is 63. The &RRNAME; RR is class independent. The RDATA of the resource record consists of four fields: Serial, Scheme, Hash Algorithm, and Digest.

This document specifies &RRNAME; RRs located at the zone apex. Non-apex &RRNAME; RRs are not forbidden, but have no meaning in this specification. Non-apex &RRNAME; RRs MUST NOT be used for verification. During digest calculation, non-apex &RRNAME; RRs are treated as ordinary RRs. They are digested as-is and the RR is not replaced by a placeholder RR. Unless explicitly stated otherwise, "&RRNAME;" always refers to apex records throughout this document.

The &RRNAME; RDATA wire format is encoded as follows:

The Serial field is a 32-bit unsigned integer in network byte order. It is the serial number from the zone's SOA record ( section 3.3.13) for which the zone digest was generated. It is included here to clearly bind the &RRNAME; RR to a particular version of the zone's content. Without the serial number, a stand-alone &RRNAME; digest has no obvious association to any particular instance of a zone.

The Scheme field is an 8-bit unsigned integer that identifies the methods by which data is collated and presented as input to the hashing function. Herein, SIMPLE, with Scheme value 1, is the only standardized Scheme defined for &RRNAME; records and it MUST be supported by implementations. The Scheme registry is further described in . Scheme values 240-254 are allocated for Private Use.

The Hash Algorithm field is an 8-bit unsigned integer that identifies the cryptographic hash algorithm used to construct the digest. Herein, SHA384 , with Hash Algorithm value 1, is the only standardized Hash Algorithm defined for &RRNAME; records that MUST be supported by implementations. When SHA384 is used, the size of the Digest field is 48 octets. The result of the SHA384 digest algorithm MUST NOT be truncated, and the entire 48 octet digest is published in the &RRNAME; record. SHA512 , with Hash Algorithm value 2, is also defined for &RRNAME; records, and SHOULD be supported by implementations. When SHA512 is used, the size of the Digest field is 64 octets. The result of the SHA512 digest algorithm MUST NOT be truncated, and the entire 64 octet digest is published in the &RRNAME; record. Hash Algorithm values 240-254 are allocated for Private Use. The Hash Algorithm registry is further described in .

The Digest field is a variable-length sequence of octets containing the output of the hash algorithm. The length of the Digest field is determined by deducting the fixed size of the Serial, Scheme, and Hash Algorithm fields from the RDATA size in the &RRNAME; RR header. The Digest field MUST NOT be shorter than 12 octets. Digests for the SHA384 and SHA512 hash algorithms specified herein are never truncated. Digests for future hash algorithms MAY be truncated, but MUST NOT be truncated to a length that results in less than 96-bits (12 octets) of equivalent strength. describes how to calculate the digest for a zone. describes how to use the digest to verify the contents of a zone.

The presentation format of the RDATA portion is as follows: The Serial field is represented as an unsigned decimal integer. The Scheme field is represented as an unsigned decimal integer. The Hash Algorithm field is represented as an unsigned decimal integer. The Digest is represented as a sequence of case-insensitive hexadecimal digits. Whitespace is allowed within the hexadecimal text.

The following example shows a &RRNAME; RR in presentation format:

example.com. 86400 IN &RRNAME; 2018031500 1 1 ( FEBE3D4CE2EC2FFA4BA99D46CD69D6D29711E55217057BEE 7EB1A7B641A47BA7FED2DD5B97AE499FAFA4F22C6BD647DE )

The zone operator chooses an appropriate hash algorithm and scheme, and includes the calculated zone digest in the apex &RRNAME; RRset. The zone operator MAY choose any of the defined hash algorithms and schemes, including the private use code points. The &RRNAME; RRSet MAY contain multiple records to support algorithm agility . [RFC Editor: change that to BCP 201] When multiple &RRNAME; RRs are present, each MUST specify a unique Scheme and Hash Algorithm tuple. It is RECOMMENDED that a zone include only one &RRNAME; RR, unless the zone operator is in the process of transitioning to a new scheme or hash algorithm.

The algorithm described in this section is designed for the common case of offline DNSSEC signing. Slight deviations may be permitted or necessary in other situations, such as with unsigned zones or online DNSSEC signing. Implementations that deviate from the described algorithm are advised to ensure that it produces &RRNAME; RRs, signatures, and dential-of-existence records that are identical to the ones generated by this procedure.

In preparation for calculating the zone digest(s), any existing &RRNAME; records (and covering RRSIGs) at the zone apex are first deleted. Prior to calculation of the digest, and prior to signing with DNSSEC, one or more placeholder &RRNAME; records are added to the zone apex. This ensures that denial-of-existence (NSEC, NSEC3) records are created correctly if the zone is signed with DNSSEC. If placeholders were not added prior to signing, the later addition of &RRNAME; records would also require updating the Type Bit Maps field of any apex NSEC/NSEC3 RRs, which then invalidates the calculated digest value. When multiple &RRNAME; RRs are published in the zone, e.g., during an algorithm rollover, each MUST specify a unique Scheme and Hash Algorithm tuple. It is RECOMMENDED that the TTL of the &RRNAME; record match the TTL of the SOA. However, the TTL of the &RRNAME; record may be safely ignored during verification in all cases. In the placeholder record, the Serial field is set to the current SOA Serial. The Scheme field is set to the value for the chosen collation scheme. The Hash Algorithm field is set to the value for the chosen hash algorithm. Since apex &RRNAME; records are excluded from digest calculation, the value of the Digest field does not matter at this point in the process.

Following addition of placeholder records, the zone may be signed with DNSSEC. When the digest calculation is complete, and the &RRNAME; record is updated, the signature(s) for the &RRNAME; RRSet MUST be recalculated and updated as well. Therefore, the signer is not required to calculate a signature over the placeholder record at this step in the process, but it is harmless to do so.

Herein, only the SIMPLE collation scheme is defined. Additional schemes may be defined in future updates to this document.

For the SIMPLE scheme, the digest is calculated over the zone as a whole. This means that a change to a single RR in the zone requires iterating over all RRs in the zone to recalculate the digest. SIMPLE is a good choice for zones that are small and/or stable, but probably not good for zones that are large and/or dynamic. Calculation of a zone digest requires RRs to be processed in a consistent format and ordering. This specification uses DNSSEC’s canonical on-the-wire RR format (without name compression) and ordering as specified in Sections 6.1, 6.2, and 6.3 of with the additional provision that RRSets having the same owner name MUST be numerically ordered, in ascending order, by their numeric RR TYPE.

When iterating over records in the zone, the following inclusion/exclusion rules apply: All records in the zone, including glue records, MUST be included, unless excluded by a subsequent rule. Occluded data ( Section 3.5) MUST be included. If there are duplicate RRs with equal owner, class, type, and RDATA, only one instance is included ( Section 6.3), and the duplicates MUST be omitted. The placeholder apex &RRNAME; RR(s) MUST NOT be included. If the zone is signed, DNSSEC RRs MUST be included, except: The RRSIG covering the apex &RRNAME; RRSet MUST NOT be included because the RRSIG will be updated after all digests have been calculated.

A zone digest using the SIMPLE scheme is calculated by concatenating all RRs in the zone, in the format and order described in subject to the inclusion/exclusion rules described in , and then applying the chosen hash algorithm:

digest = hash( RR(1) | RR(2) | RR(3) | ... ) where "|" denotes concatenation.

The calculated zone digest is inserted into the placeholder &RRNAME; RR. Repeat for each digest if multiple digests are to be published. If the zone is signed with DNSSEC, the RRSIG record(s) covering the &RRNAME; RRSet MUST then be added or updated. Because the &RRNAME; placeholder was added prior to signing, the zone will already have the appropriate denial-of-existence (NSEC, NSEC3) records. Some DNSSEC implementations (especially "online signing") might update the SOA serial number whenever a new signature is made. To preserve the calculated digest, generation of a &RRNAME; signature MUST NOT also result in a change to the SOA serial number. The &RRNAME; RR and the matching SOA MUST be published at the same time.

The recipient of a zone that has a &RRNAME; RR verifies the zone by calculating the digest as follows. If multiple &RRNAME; RRs are present in the zone, e.g., during an algorithm rollover, a match using any one of the recipient's supported Schemes and Hash Algorithms is sufficient to verify the zone. The verifier MAY ignore a &RRNAME; RR if its Scheme and Hash Algorithm violates local policy. The verifier MUST first determine whether or not to expect DNSSEC records in the zone. By examining locally configured trust anchors, and, if necessary, querying for and validating DS RRs in the parent zone, the verifier knows whether or not the zone to be verified should include DNSSEC keys and signatures. For zones where signatures are not expected, or if DNSSEC validation is not performed, digest verification continues at step below. For zones where signatures are expected, the existence of the apex &RRNAME; record MUST be validated. If the DNSSEC data proves the &RRNAME; RRSet does not exist, digest verification cannot occur. If the DNSSEC data proves the &RRNAME; does exist, but is not found in the zone, digest verification MUST NOT be considered successful. For zones where signatures are expected, the SOA and &RRNAME; RRSets MUST have valid signatures, chaining up to a trust anchor. If DNSSEC validation of the SOA or &RRNAME; RRSets fails, digest verification MUST NOT be considered successful. When multiple &RRNAME; RRs are present, each MUST specify a unique Scheme and Hash Algorithm tuple. If the &RRNAME; RRSet contains more than one RR with the same Scheme and Hash Algorithm, digest verification for those &RRNAME; RRs MUST NOT be considered successful. Loop over all apex &RRNAME; RRs and perform the following steps: The SOA Serial field MUST exactly match the &RRNAME; Serial field. If the fields do not match, digest verification MUST NOT be considered successful with this &RRNAME; RR. The Scheme field MUST be checked. If the verifier does not support the given scheme, verification MUST NOT be considered successful with this &RRNAME; RR. The Hash Algorithm field MUST be checked. If the verifier does not support the given hash algorithm, verification MUST NOT be considered successful with this &RRNAME; RR. The Digest field size MUST be checked. If the size of the given Digest field is smaller than 12 octets, or if the size is not equal to the size expected for the corresponding Hash Algorithm, verification MUST NOT be considered successful with this &RRNAME; RR. The zone digest is computed over the zone data as described in , using the Scheme and Hash Algorithm for the current &RRNAME; RR. The computed digest is compared to the received digest. If the two digest values match, verification is considered successful. Otherwise, verification MUST NOT be considered successful for this &RRNAME; RR. Each time zone verification is performed, the verifier SHOULD report the status as either successful or unsuccessful. When unsuccessful, the verifier SHOULD report the reason(s) that verification did not succeed.

This document defines a new DNS RR type, &RRNAME;, whose value 63 has been allocated by IANA from the "Resource Record (RR) TYPEs" subregistry of the "Domain Name System (DNS) Parameters" registry: Type: &RRNAME; Value: 63 Meaning: Message Digest Over Zone Data Reference: [this document]

IANA is requested to create a new registry on the "Domain Name System (DNS) Parameters" web page as follows: Registry Name: &RRNAME; Schemes Registration Procedure: Specification Required Reference: [this document] Value Description Mnemonic Reference 0 Reserved 1 Simple &RRNAME; collation SIMPLE [this document] 2-239 Unassigned 240-254 Private Use N/A [this document] 255 Reserved

IANA is requested to create a new registry on the "Domain Name System (DNS) Parameters" web page as follows: Registry Name: &RRNAME; Hash Algorithms Registration Procedure: Specification Required Reference: [this document] Value Description Mnemonic Reference 0 Reserved 1 SHA-384 SHA384 [this document] 2 SHA-512 SHA512 [this document] 3-239 Unassigned 240-254 Private Use N/A [his document] 255 Reserved

Users of &RRNAME; with unsigned zones are advised that it provides no real protection against attacks. While zone digests can be used in the absence of DNSSEC, this only provides protection against accidental zone corruption, such as transmission errors and truncation. When used in this manner, it effectively serves only as a checksum. For zones not signed with DNSSEC, an attacker can make any zone modifications appear to be valid by recomputing Digest field of a &RRNAME; RR.

An attacker, whose goal is to modify zone content before it is used by the victim, may consider a number of different approaches. The attacker might perform a downgrade attack to an unsigned zone. This is why talks about determining whether or not to expect DNSSEC signatures for the zone in step . The attacker might perform a downgrade attack by removing one or more &RRNAME; records. Such a removal is detectable only with DNSSEC validation and is why talks about checking denial-of-existence proofs in step and signature validation in step . The attacker might alter the Scheme, Hash Algorithm, or Digest fields of the &RRNAME; record. Such modifications are detectable only with DNSSEC validation. As stated in , cryptographic algorithms age and become weaker as cryptanalysis techniques and computing resources improve with time. Implementors and publishers of zone digests should anticipate the need for algorithm agility on long timescales.

When a zone publishes multiple &RRNAME; RRs, the overall security is only as good as the weakest hash algorithm in use. For this reason, recommends only publishing multiple &RRNAME; RRs when transitioning to a new scheme or hash algorithm. Once the transition is complete, the old scheme or hash algorithm should be removed from the &RRNAME; RRSet.

As with all DNSSEC signatures, the ability to perform signature validation of a &RRNAME; record is limited in time. If the DS record(s) or trust anchors for the zone to be verified are no longer available, the recipient cannot validate the &RRNAME; RRSet. This could happen even if the &RRNAME; signature is still current (not expired), since the zone's DS record(s) may have been withdrawn following a Key Signing Key (KSK) rollover. For zones where it may be important to validate a &RRNAME; RRSet through its entire signature validity period, the zone operator should ensure that KSK rollover timing takes this into consideration.

Nothing in this specification prevents clients from making, and servers from responding to, &RRNAME; queries. Servers SHOULD NOT calculate zone digests dynamically (for each query) as this can be used as a CPU resource exhaustion attack. &RRNAME; responses could be used in a distributed denial-of-service amplification attack. The &RRNAME; RR is moderately sized, much like the DS RR. A single &RRNAME; RR contributes approximately 65 to 95 octets to a DNS response, for digest types defined herein. Other RR types, such as DNSKEY, can result in larger amplification effects.

&RRNAME; is used to detect incomplete or corrupted zone data prior to its use, thereby increasing resilience by not using corrupt data, but also introduces some denial-of-service fragility by making good data in a zone unavailable if some other data is missing or corrupt. Publishers and consumers of zones containing &RRNAME; records should be aware of these tradeoffs. While the intention is to secure the zone data, misconfigurations or implementation bugs are generally indistinguishable from intentional tampering, and could lead to service failures when verification is performed automatically. Zone publishers may want to deploy &RRNAME; gradually, perhaps by utilizing one of the private use hash algorithm code points listed in . Similarly, recipients may want to initially configure verification failures only as a warning, and later as an error after gaining experience and confidence with the feature.

This section is provided to make zone publishers aware of the performance requirements and implications of including &RRNAME; RRs in a zone.

As mentioned previously, the SIMPLE scheme may be impractical for use in zones that are either large or highly dynamic. Zone publishers should carefully consider the use of &RRNAME; in such zones, since it might cause consumers of zone data (e.g., secondary name servers) to expend resources on digest calculation. For such use cases, it is recommended that &RRNAME; only be used when digest calculation time is significantly less than propagation times and update intervals. The authors' implementation () includes an option to record and report CPU usage of its operation. The software was used to generate digests for more than 800 TLD zones available from . The table below summarizes the results for the SIMPLE scheme and SHA384 hash algorithm grouped by zone size. The Rate column is the mean amount of time per RR to calculate the digest, running on commodity hardware in early 2020. Zone Size (RRs) Rate (msec/RR) 10 - 99 0.00683 100 - 999 0.00551 1000 - 9999 0.00505 10000 - 99999 0.00602 100000 - 999999 0.00845 1000000 - 9999999 0.0108 10000000 - 99999999 0.0148 For example, based on the above table, it takes approximately 0.13 seconds to calculate a SIMPLE SHA384 digest for a zone with 22,000 RRs, and about 2.5 seconds for a zone with 300,000 RRs. These benchmarks attempt to emulate a worst-case scenario and take into account the time required to canonicalize the zone for processing. Each of the 800+ zones were measured three times, and then averaged, with a different random sorting of the input data prior to each measurement.

This specification has no impact on user privacy.

The authors wish to thank David Blacka, Scott Hollenbeck, and Rick Wilhelm for providing feedback on early drafts of this document. Additionally, they thank Joe Abley, Mark Andrews, Ralph Dolmans, Donald Eastlake, Richard Gibson, Olafur Gudmundsson, Bob Harold, Paul Hoffman, Evan Hunt, Shumon Huque, Tatuya Jinmei, Mike St. Johns, Burt Kaliski, Shane Kerr, Matt Larson, Barry Leiba, John Levine, Ed Lewis, Matt Pounsett, Mukund Sivaraman, Petr Spacek, Ondrej Sury, Willem Toorop, Florian Weimer, Tim Wicinski, Wouter Wijngaards, Paul Wouters, and other members of the DNSOP working group for their input.

RFC Editor: Please remove this section before publication. This section lists substantial changes to the document as it is being worked on. From -00 to -01: Removed requirement to sort by RR CLASS. Added Kumari and Hardaker as coauthors. Added Change Log section. Minor clarifications and grammatical edits. From -01 to -02: Emphasize desire for data security over channel security. Expanded motivation into its own subsection. Removed discussion topic whether or not to include serial in &RRNAME;. Clarified that a zone's NS records always sort before the SOA record. Clarified that all records in the zone must are digested, except as specified in the exclusion rules. Added for discussion out-of-zone and occluded records. Clarified that update of &RRNAME; signature must not cause a serial number change. Added persons to acknowledgments. From -02 to -03: Added recommendation to set &RRNAME; TTL to SOA TTL. Clarified that digest input uses uncompressed names. Updated Implementations section. Changed intended status from Standards Track to Experimental and added Scope of Experiment section. Updated Motivation, Introduction, and Design Overview sections in response to working group discussion. Gave &RRNAME; digest types their own status, separate from DS digest types. Request IANA to create a registry. Added Reserved field for future work supporting dynamic updates. Be more rigorous about having just ONE &RRNAME; record in the zone. Expanded use cases. From -03 to -04: Added an appendix with example zones and digests. Clarified that only apex &RRNAME; RRs shall be processed. From -04 to -05: Made SHA384 the only supported &RRNAME; digest type. Disassociated &RRNAME; digest types from DS digest types. Updates to Introduction based on list feedback. Changed "zone file" to "zone" everywhere. Restored text about why &RRNAME; has a Serial field. Clarified ordering of RRSets having same owner to be numerically ascending. Clarified that all duplicate RRs (not just SOA) must be suppressed in digest calculation. Clarified that the Reserved field must be set to zero and checked for zero in verification. Clarified that occluded data must be included. Clarified procedure for verification, using temporary location for received digest. Explained why Reserved field is 8-bits. IANA Considerations section now more specific. Added complex zone to examples. From -05 to -06: RR type code 63 was assigned to &RRNAME; by IANA. From -06 to -07: Fixed mistakes in &RRNAME; examples. Added private use Digest Type values 240-254. Clarified that Digest field must not be empty. From -07 to draft-ietf-dnsop-dns-zone-digest-00: Adopted by dnsop. Clarified further that non-apex &RRNAME; RRs have no meaning. Changed "provably [un]signed" to "provably [in]secure". Allow multiple &RRNAME; RRs to support algorithm agility/rollovers. Describe verification when there are multiple &RRNAME; RRs. From -00 to -01: Simplified requirements around verifying multiple digests. Any one match is sufficient. Updated implementation notes. Both implementations produce expected results on examples given in this document. From -01 to -02: Changed the name of the Reserved field to Parameter. Changed the name of Digest Type 1 from SHA384 to SHA384-STABLE. The meaning of the Parameter field now depends on Digest Type. No longer require Parameter field to be zero in verification. Updated a rule from earlier versions that said multiple &RRNAME; RRs were not allowed. From -02 to -03: Changed the name of Digest Type 1 from SHA384-STABLE to SHA384-SIMPLE. Changed document status from Experimental to Standards Track. Removed Scope of Experimentation section. From -03 to -04: Addressing WGLC feedback. Changed from "Digest Type + Paramter" to "Scheme + Hash Algorithm". This should make it more obvious how &RRNAME; can be expanded in the future with new schemes and hash algorithms, while sacrificing some of the flexibility that the Parameter was intended to provide. Note: old RDATA fields: Serial, Digest Type, Parameter, Digest. Note: new RDATA fields: Serial, Scheme, Hash Algorithm, Digest. Add new IANA requirement for a Scheme registry. Rearranged some sections and separated scheme-specific aspects from general aspects of digest calculation. When discussing multiple &RRNAME; RRs, allow for Scheme, as well as Hash Algorithm, transition. Added Performance Considerations section with some benchmarks. Further clarifications about non-apex &RRNAME; RRs. Clarified inclusion rule for duplicate RRs. Removed or lowercased some inappropriately used RFC 2119 key words. Clarified that all &RRNAME; RRs, even for unsupported hash algorithms, must be zeroized during digest calculation. Added Resilience and Fragility to security considerations. Updated examples since changes in this version result in different hash values. From -04 to -05: Clarifications about non-apex and multiple &RRNAME; RRs. Clarifications about benchmark results. Don't compute &RRNAME; on-the-fly. Specification Required for updates to &RRNAME; protocol registries. Other rewording based on WGLC feedback. Updated RFC numbers for some references. Use documentation IP addresses instead of loopback. Updated examples in the appendix. From -05 to -06: Per WG suggestion, no longer include any apex &RRNAME; record in digest calculation. Updated examples in the appendix. Clarified verification procedure by describing a loop over all &RRNAME; RRs. From -06 to -07: Added NIC Chile Labs implementation. From -07 to -08: Update an author's affiliation. Clarified why placeholder RRs are still important (for NSEC/NSEC3). Moved subsection ("Order of RRSets Having the Same Owner Name") with single sentence paragraph up into parent section. From -08 to -09: Moved format, ordering, inclusion/exclusion into a sub section specific to the SIMPLE scheme. Further clarified rules about multiple &RRNAME; RRs (AD comments). Reworded rules about processing of duplicate zone RRs (AD comments). Removed sentence about optional zeroing of digest prior to calculation (AD comments). Other minor changes (AD comments). From -09 to -10: Add clarification and reference to on-disk modification / corruption of zone files. Added concerns that timing of KSK rollovers could affect validation of &RRNAME; record. Addressed SECDIR review and accepted most proposed edits. From SECDIR review, require minimum digest length of 12 octets. From SECDIR review, add SHA512 has hash algorithm 2. From SECDIR review, say that &RRNAME; RRs MAY be ignored by local policy. Moved Implementation Status to an appendix with the intention to retain it in RFC. In registry tables, changed Status column to Implementation Requirement. From -10 to -11: Fixed people's names in the acknowledgments section (blush) Say "has not been modified between origination and retrieval." Say that &RRNAME; TTL doesn't matter during verification. Further clarification that the SHA-384 and SHA-512 hashes are not truncated. Future algs might be truncated, but never below 96 bits. From -11 to -12: SECDIR review: make "recommended" all caps. SECDIR review: tweak explanation of why &RRNAME; RR has copy of SOA serial. SECDIR review: be even more clear about apex &RRNAME; RRs vs non-apex. SECDIR review: Forgot to delete sentence about IANA policy for adding new hash algorithms. SECDIR review: Spell out Key Signing Key first time. SECDIR review: say "private use hash algorithm code points." SECDIR review: Update estimates of &RRNAME; RR size. From -12 to -13: Added reference to draft-ietf-dprive-xfr-over-tls. Dropped Implementation Requirement from registry tables. Added Use of Multiple ZONEMD Hash Algorithms to Security Considerations. Added Using Zone Digest Without DNSSEC to Security Considerations. Added notes about the need for algorithm agility due to crypto algorithm aging. Further clarified that only with DNSSEC can &RRNAME; guarantee integrity and authenticity. For unsigned zones, &RRNAME; serves only as a checksum. Calculation algorithm is designed for common case of offline signing. Deviations may be allowed as long as the end result is the same. Numerous small edits and clarifications from IESG reviewer comments. From -13 to -14: A few more edits and clarifications from IESG reviewer comments. Moved paragraph about multiple digests to new section titled Including &RRNAME; RRs in a Zone. MUST be implemented -> MUST be supported by implementations. Consolidated SHOULD requirements about error reporting to single place at the conclusion of verification. Rephrased "provably insecure" etc as using DNSSEC validation to know whether or not the zone is expected to have keys and signatures.

&RFC2119; &RFC1034; &RFC1035; &RFC4034; &RFC6234; &RFC8174; &RFC1995; &RFC2065; &RFC2136; &RFC2535; &RFC2845; &RFC2931; &RFC3258; &RFC4880; &RFC5155; &RFC5751; &RFC5936; &RFC7696; &RFC8126; &RFC8484; &RFC8499; &RFC8806; &RFC8901; &I-D.ietf-dprive-xfr-over-tls; ICANN Root Server Operators Verisign NIC Chile Labs Internet Corporation for Assigned Names and Numbers Wikipedia DENIC

This appendix contains example zones with accurate &RRNAME; records. These can be used to verify an implementation of the zone digest protocol.

Here, the EXAMPLE zone contains an SOA record, NS and glue records, and a &RRNAME; record.

Here, the EXAMPLE zone contains duplicate RRs, and an occluded RR, and one out-of-zone RR.

Here, the EXAMPLE zone contains multiple &RRNAME; records. It has both SHA384 and SHA512 digests using the SIMPLE scheme. It also includes &RRNAME; records with Scheme and Hash Algorithm values in the private range (240-254). These additional private-range digests are not verifiable.

The URI.ARPA zone retrieved 2018-10-21. Note this sample zone has (expired) signatures, but no signature for the &RRNAME; RR.

> DiG 9.9.4 <<>> @lax.xfr.dns.icann.org uri.arpa axfr ; (2 servers found) ;; global options: +cmd uri.arpa. 3600 IN SOA sns.dns.icann.org. ( noc.dns.icann.org. 2018100702 10800 3600 1209600 3600 ) uri.arpa. 3600 IN RRSIG NSEC 8 2 3600 ( 20181028142623 20181007205525 47155 uri.arpa. eEC4w/oXLR1Epwgv4MBiDtSBsXhqrJVvJWUpbX8XpetAvD35bxwNCUTi /pAJVUXefegWeiriD2rkTgCBCMmn7YQIm3gdR+HjY/+o3BXNQnz97f+e HAE9EDDzoNVfL1PyV/2fde9tDeUuAGVVwmD399NGq9jWYMRpyri2kysr q/g= ) uri.arpa. 86400 IN RRSIG NS 8 2 86400 ( 20181028172020 20181007175821 47155 uri.arpa. ATyV2A2A8ZoggC+68u4GuP5MOUuR+2rr3eWOkEU55zAHld/7FiBxl4ln 4byJYy7NudUwlMOEXajqFZE7DVl8PpcvrP3HeeGaVzKqaWj+aus0jbKF Bsvs2b1qDZemBfkz/IfAhUTJKnto0vSUicJKfItu0GjyYNJCz2CqEuGD Wxc= ) uri.arpa. 600 IN RRSIG MX 8 2 600 ( 20181028170556 20181007175821 47155 uri.arpa. e7/r3KXDohX1lyVavetFFObp8fB8aXT76HnN9KCQDxSnSghNM83UQV0t lTtD8JVeN1mCvcNFZpagwIgB7XhTtm6Beur/m5ES+4uSnVeS6Q66HBZK A3mR95IpevuVIZvvJ+GcCAQpBo6KRODYvJ/c/ZG6sfYWkZ7qg/Em5/+3 4UI= ) uri.arpa. 3600 IN RRSIG DNSKEY 8 2 3600 ( 20181028152832 20181007175821 15796 uri.arpa. nzpbnh0OqsgBBP8St28pLvPEQ3wZAUdEBuUwil+rtjjWlYYiqjPxZ286 XF4Rq1usfV5x71jZz5IqswOaQgia91ylodFpLuXD6FTGs2nXGhNKkg1V chHgtwj70mXU72GefVgo8TxrFYzxuEFP5ZTP92t97FVWVVyyFd86sbbR 6DZj3uA2wEvqBVLECgJLrMQ9Yy7MueJl3UA4h4E6zO2JY9Yp0W9woq0B dqkkwYTwzogyYffPmGAJG91RJ2h6cHtFjEZe2MnaY2glqniZ0WT9vXXd uFPm0KD9U77Ac+ZtctAF9tsZwSdAoL365E2L1usZbA+K0BnPPqGFJRJk 5R0A1w== ) uri.arpa. 3600 IN RRSIG DNSKEY 8 2 3600 ( 20181028152832 20181007175821 55480 uri.arpa. lWtQV/5szQjkXmbcD47/+rOW8kJPksRFHlzxxmzt906+DBYyfrH6uq5X nHvrUlQO6M12uhqDeL+bDFVgqSpNy+42/OaZvaK3J8EzPZVBHPJykKMV 63T83aAiJrAyHzOaEdmzLCpalqcEE2ImzlLHSafManRfJL8Yuv+JDZFj 2WDWfEcUuwkmIZWX11zxp+DxwzyUlRl7x4+ok5iKZWIg5UnBAf6B8T75 WnXzlhCw3F2pXI0a5LYg71L3Tp/xhjN6Yy9jGlIRf5BjB59X2zra3a2R PkI09SSnuEwHyF1mDaV5BmQrLGRnCjvwXA7ho2m+vv4SP5dUdXf+GTeA 1HeBfw== ) uri.arpa. 3600 IN RRSIG SOA 8 2 3600 ( 20181029114753 20181008222815 47155 uri.arpa. qn8yBNoHDjGdT79U2Wu9IIahoS0YPOgYP8lG+qwPcrZ1BwGiHywuoUa2 Mx6BWZlg+HDyaxj2iOmox+IIqoUHhXUbO7IUkJFlgrOKCgAR2twDHrXu 9BUQHy9SoV16wYm3kBTEPyxW5FFm8vcdnKAF7sxSY8BbaYNpRIEjDx4A JUc= ) uri.arpa. 3600 IN NSEC ftp.uri.arpa. NS SOA ( MX RRSIG NSEC DNSKEY ) uri.arpa. 86400 IN NS a.iana-servers.net. uri.arpa. 86400 IN NS b.iana-servers.net. uri.arpa. 86400 IN NS c.iana-servers.net. uri.arpa. 86400 IN NS ns2.lacnic.net. uri.arpa. 86400 IN NS sec3.apnic.net. uri.arpa. 600 IN MX 10 pechora.icann.org. uri.arpa. 3600 IN DNSKEY 256 3 8 ( AwEAAcBi7tSart2J599zbYWspMNGN70IBWb4ziqyQYH9MTB/VCz6WyUK uXunwiJJbbQ3bcLqTLWEw134B6cTMHrZpjTAb5WAwg4XcWUu8mdcPTiL Bl6qVRlRD0WiFCTzuYUfkwsh1Rbr7rvrxSQhF5rh71zSpwV5jjjp65Wx SdJjlH0B ) uri.arpa. 3600 IN DNSKEY 257 3 8 ( AwEAAbNVv6ulgRdO31MtAehz7j3ALRjwZglWesnzvllQl/+hBRZr9QoY cO2I+DkO4Q1NKxox4DUIxj8SxPO3GwDuOFR9q2/CFi2O0mZjafbdYtWc 3zSdBbi3q0cwCIx7GuG9eqlL+pg7mdk9dgdNZfHwB0LnqTD8ebLPsrO/ Id7kBaiqYOfMlZnh2fp+2h6OOJZHtY0DK1UlssyB5PKsE0tVzo5s6zo9 iXKe5u+8WTMaGDY49vG80JPAKE7ezMiH/NZcUMiE0PRZ8D3foq2dYuS5 ym+vA83Z7v8A+Rwh4UGnjxKB8zmr803V0ASAmHz/gwH5Vb0nH+LObwFt l3wpbp+Wpm8= ) uri.arpa. 3600 IN DNSKEY 257 3 8 ( AwEAAbwnFTakCvaUKsXji4mgmxZUJi1IygbnGahbkmFEa0L16J+TchKR wcgzVfsxUGa2MmeA4hgkAooC3uy+tTmoMsgy8uq/JAj24DjiHzd46LfD FK/qMidVqFpYSHeq2Vv5ojkuIsx4oe4KsafGWYNOczKZgH5loGjN2aJG mrIm++XCphOskgCsQYl65MIzuXffzJyxlAuts+ecAIiVeqRaqQfr8LRU 7wIsLxinXirprtQrbor+EtvlHp9qXE6ARTZDzf4jvsNpKvLFZtmxzFf3 e/UJz5eHjpwDSiZL7xE8aE1o1nGfPtJx9ZnB3bapltaJ5wY+5XOCKgY0 xmJVvNQlwdE= ) ftp.uri.arpa. 3600 IN RRSIG NSEC 8 3 3600 ( 20181028080856 20181007175821 47155 uri.arpa. HClGAqPxzkYkAT7Q/QNtQeB6YrkP6EPOef+9Qo5/2zngwAewXEAQiyF9 jD1USJiroM11QqBS3v3aIdW/LXORs4Ez3hLcKNO1cKHsOuWAqzmE+BPP Arfh8N95jqh/q6vpaB9UtMkQ53tM2fYU1GszOLN0knxbHgDHAh2axMGH lqM= ) ftp.uri.arpa. 604800 IN RRSIG NAPTR 8 3 604800 ( 20181028103644 20181007205525 47155 uri.arpa. WoLi+vZzkxaoLr2IGZnwkRvcDf6KxiWQd1WZP/U+AWnV+7MiqsWPZaf0 9toRErerGoFOiOASNxZjBGJrRgjmavOM9U+LZSconP9zrNFd4dIu6kp5 YxlQJ0uHOvx1ZHFCj6lAt1ACUIw04ZhMydTmi27c8MzEOMepvn7iH7r7 k7k= ) ftp.uri.arpa. 3600 IN NSEC http.uri.arpa. NAPTR ( RRSIG NSEC ) ftp.uri.arpa. 604800 IN NAPTR 0 0 "" "" ( "!^ftp://([^:/?#]*).*$!\\1!i" . ) http.uri.arpa. 3600 IN RRSIG NSEC 8 3 3600 ( 20181029010647 20181007175821 47155 uri.arpa. U03NntQ73LHWpfLmUK8nMsqkwVsOGW2KdsyuHYAjqQSZvKbtmbv7HBmE H1+Ii3Z+wtfdMZBy5aC/6sHdx69BfZJs16xumycMlAy6325DKTQbIMN+ ift9GrKBC7cgCd2msF/uzSrYxxg4MJQzBPvlkwXnY3b7eJSlIXisBIn7 3b8= ) http.uri.arpa. 604800 IN RRSIG NAPTR 8 3 604800 ( 20181029011815 20181007205525 47155 uri.arpa. T7mRrdag+WSmG+n22mtBSQ/0Y3v+rdDnfQV90LN5Fq32N5K2iYFajF7F Tp56oOznytfcL4fHrqOE0wRc9NWOCCUec9C7Wa1gJQcllEvgoAM+L6f0 RsEjWq6+9jvlLKMXQv0xQuMX17338uoD/xiAFQSnDbiQKxwWMqVAimv5 7Zs= ) http.uri.arpa. 3600 IN NSEC mailto.uri.arpa. NAPTR ( RRSIG NSEC ) http.uri.arpa. 604800 IN NAPTR 0 0 "" "" ( "!^http://([^:/?#]*).*$!\\1!i" . ) mailto.uri.arpa. 3600 IN RRSIG NSEC 8 3 3600 ( 20181028110727 20181007175821 47155 uri.arpa. GvxzVL85rEukwGqtuLxek9ipwjBMfTOFIEyJ7afC8HxVMs6mfFa/nEM/ IdFvvFg+lcYoJSQYuSAVYFl3xPbgrxVSLK125QutCFMdC/YjuZEnq5cl fQciMRD7R3+znZfm8d8u/snLV9w4D+lTBZrJJUBe1Efc8vum5vvV7819 ZoY= ) mailto.uri.arpa. 604800 IN RRSIG NAPTR 8 3 604800 ( 20181028141825 20181007205525 47155 uri.arpa. MaADUgc3fc5v++M0YmqjGk3jBdfIA5RuP62hUSlPsFZO4k37erjIGCfF j+g84yc+QgbSde0PQHszl9fE/+SU5ZXiS9YdcbzSZxp2erFpZOTchrpg 916T4vx6i59scodjb0l6bDyZ+mtIPrc1w6b4hUyOUTsDQoAJYxdfEuMg Vy4= ) mailto.uri.arpa. 3600 IN NSEC urn.uri.arpa. NAPTR ( RRSIG NSEC ) mailto.uri.arpa. 604800 IN NAPTR 0 0 "" "" ( "!^mailto:(.*)@(.*)$!\\2!i" . ) urn.uri.arpa. 3600 IN RRSIG NSEC 8 3 3600 ( 20181028123243 20181007175821 47155 uri.arpa. Hgsw4Deops1O8uWyELGe6hpR/OEqCnTHvahlwiQkHhO5CSEQrbhmFAWe UOkmGAdTEYrSz+skLRQuITRMwzyFf4oUkZihGyhZyzHbcxWfuDc/Pd/9 DSl56gdeBwy1evn5wBTms8yWQVkNtphbJH395gRqZuaJs3LD/qTyJ5Dp LvA= ) urn.uri.arpa. 604800 IN RRSIG NAPTR 8 3 604800 ( 20181029071816 20181007205525 47155 uri.arpa. ALIZD0vBqAQQt40GQ0Efaj8OCyE9xSRJRdyvyn/H/wZVXFRFKrQYrLAS D/K7q6CMTOxTRCu2J8yes63WJiaJEdnh+dscXzZkmOg4n5PsgZbkvUSW BiGtxvz5jNncM0xVbkjbtByrvJQAO1cU1mnlDKe1FmVB1uLpVdA9Ib4J hMU= ) urn.uri.arpa. 3600 IN NSEC uri.arpa. NAPTR RRSIG ( NSEC ) urn.uri.arpa. 604800 IN NAPTR 0 0 "" "" ( "/urn:([^:]+)/\\1/i" . ) uri.arpa. 3600 IN SOA sns.dns.icann.org. ( noc.dns.icann.org. 2018100702 10800 3600 1209600 3600 ) ;; Query time: 66 msec ;; SERVER: 192.0.32.132#53(192.0.32.132) ;; WHEN: Sun Oct 21 20:39:28 UTC 2018 ;; XFR size: 34 records (messages 1, bytes 3941) uri.arpa. 3600 IN ZONEMD 2018100702 1 1 ( 1291b78ddf7669b1a39d014d87626b709b55774c5d7d58fa dc556439889a10eaf6f11d615900a4f996bd46279514e473 ) ]]>

The ROOT-SERVERS.NET zone retrieved 2018-10-21.

RFC Editor: Please retain this section upon publication. This section records the status of known implementations of the protocol defined by this specification at the time of publication, and is inspired by the concepts described in RFC7942. Please note that the listing of any individual implementation here does not imply endorsement by the IETF. Furthermore, no effort has been spent to verify the information presented here that was supplied by IETF contributors. This is not intended as, and must not be construed to be, a catalog of available implementations or their features. Readers are advised to note that other implementations may exist.

The authors have an open source implementation in C, using the ldns library . This implementation is able to perform the following functions: Read an input zone and output a zone with the &RRNAME; placeholder. Compute zone digest over signed zone and update the &RRNAME; record. Re-compute DNSSEC signature over the &RRNAME; record. Verify the zone digest from an input zone. This implementation does not: Perform DNSSEC validation of the &RRNAME; record during verification.

Shane Kerr wrote an implementation of this specification during the IETF 102 hackathon . This implementation is in Python and is able to perform the following functions: Read an input zone and output a zone with &RRNAME; record. Verify the zone digest from an input zone. Output the &RRNAME; record in its defined presentation format. This implementation does not: Re-compute DNSSEC signature over the &RRNAME; record. Perform DNSSEC validation of the &RRNAME; record.

NIC Chile Labs wrote an implementation of this specification as part of "dns-tools" suite , which besides digesting, can also sign and verify zones. This implementation is in Go and is able to perform the following functions: Compute zone digest over signed zone and update the &RRNAME; record. Verify the zone digest from an input zone. Perform DNSSEC validation of the &RRNAME; record during verification. Re-compute DNSSEC signature over the &RRNAME; record.