| Internet-Draft | multialgo | July 2023 |
| Thomassen | Expires 11 January 2024 | [Page] |
This document restates the requirements on DNSSEC signing and validation and makes small adjustments order to allow for more flexible handling of configurations that advertise multiple Secure Entry Points (SEP) with different signing algorithms via their DS record or trust anchor set. The adjusted rules allow both for multi-signer operation and for transfer of signed DNS zones between providers, without requiring that each provider uses the same signing algorithm. In addition, the proposal enables pre-publication of a trust anchor in preparation for an algorithm rollover, such as of the root zone.¶
This document updates RFCs 4035, 6840, and 8624.¶
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶
This Internet-Draft will expire on 11 January 2024.¶
Copyright (c) 2023 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶
DNSSEC [RFC4033][RFC4034][RFC4035][RFC6840][RFC9364] adds origin authentication to the DNS protocol. While it typically works smoothly when using a single signing algorithm, complications can occur when multiple algorithms are in use.¶
In particular, current specifications [RFC4035][RFC6840] require that a zone be signed with each signing algorithm listed in a zone's DS RRset or appearing via its trust anchors. This poses a problem for (at least) the following cases:¶
In multi-signer setups where each DNS provider maintains their own key ([RFC8901] Section 2.1.2), providers may not necessarily choose the same signing algorithm. (For example, one may choose to use algorithm 8 while the other picks algorithm 13, both of which will appear in the domain's DS RRset.) While such setups do allow establishing a chain of trust, DNS responses from either provider will only contain signatures of the one signing algorithm used by that provider, violating the specification.¶
A related issue is the transfer of a signed domain name from one provider to another, which requires a short multi-signer period in order to execute a glitch-free transition without disabing DNSSEC for the domain. If the old and the new provider do not use the same signing algorithms, the same problems appear.¶
When performing an algorithm rollover for a zone with a trust anchor, current specifications mandate that the zone has to be double-signed with both the old and the new algorithm before publishing the new trust anchor. For the root zone, this could lead to a potentially rather long phase of double-signing (on the order of a year). As this comes with both financial and SSR costs, it seems desirable to find a way for publishing the new trust anchor without introducing the new algorithm into the zone just yet.¶
For a more detailed explanation of the implications of the current rules as well as of alternative solution approaches, see Appendix A.¶
However, it turns out that these limitations are not fundamental to the construction of the DNS and DNSSEC protocols, but appear as consequences of the current requirements, which (in this very strict form) are not necessary for origin validation.¶
This document explores how the signing and validation rules can be modified to accommodate additional use cases, without compromising on the security guarantees given by DNSSEC.¶
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 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶
The heart of the issue is that even though one signature, in theory, will suffice for validation, the signer cannot, in the general case, know which particular signing algorithm(s) the validator will support -- and hence, providing a "large enough set" (read: all of them) is the approach that had been taken so far.¶
A more relaxed approach is defined which does not require all algorithms' RRSIGs to be present, while ensuring that the set of signatures provided is still "large enough" for reliable DNSSEC operation, so that glitch-free multi-signer operation and TA pre-publication are made possible. This is enabled by a new mechanism that allows the signer to determine which RRSIGs can be skipped, without risking validation failures.¶
For the case of a multi-signer setup with two generally supported algorithms (such as 8 and 13), the scheme requires only one of the two signatures. Similarly, when pre-publishing a trust anchor, associated signatures don't need to be published immediately, provided that the existing TA's algorithm is generally supported.¶
The notion of UNIVERSAL signing algorithms is introduced, and defined as follows:¶
The information contained in the table of [RFC8624] Section 3.1 is transferred into a to-be-erected IANA registry, and a boolean column is added with the heading "universal validation support". Signing algorithms where this column is TRUE are called "UNIVERSAL".¶
"MUST NOT sign" algorithms can never be UNIVERSAL. "MUST validate" is a prerequisite for UNIVERSAL. Changes that affect whether an algorithm is UNIVERSAL require standards action.¶
Algorithms 8 and 13 are the only algorithms currently declared UNIVERSAL.¶
Also, new terminology is established for algorithms in "MUST NOT sign" status: those are called "INSECURE".¶
As soon as a "MUST validate" algorithm is known or expected to have declining validation support, it should be moved to status "MUST NOT sign" (which removes the UNIVERSAL label if present, and renders the algorithm INSECURE). Accordingly, algorithms 5 and 7 are declared "MUST NOT sign".¶
The following algorithms are thus INSECURE: 1, 3, 5, 6, 7, 12¶
When the DS RRset or trust anchor set for a zone includes an unsupported INSECURE algorithm, validators MUST treat the zone as unsigned, even if signed with another supported algorithm.¶
Otherwise, validators MUST accept any valid path.¶
Implementing these rules requires validating resolvers to keep a record of INSECURE algorithms (e.g. via a static array of INSECURE algorithm numbers), so that the zone's security status can be established upon inspection of a DS record or TA set.¶
It is observed that both signers and validators need to know only one of the concepts "UNIVERSAL" and "INSECURE": to use several signing algorithms, signers only need to know which algorithms are UNIVERSAL, while validators only need to know which are INSECURE. This limits the implementation effort.¶
The new validation requirements enable stable multi-signer setups using UNIVERSAL algorithms as well as glitch-free provider transfers and algorithm upgrades from INSECURE to UNIVERSAL algorithms (such as algorithm 7 to 13), without risking SERVFAIL responses in the event that a resolver no longer supports one of the algorithms (e.g. 7). For a detailed discussion, see Section 4.¶
DNS providers in a multi-signer setup are free to limit their responses to serve signatures for one UNIVERSAL algorithm only. This one signature is sufficient to provide a valid path everywhere.¶
When a UNIVERSAL algorithm is in use, signatures of other algorithms are not required. DNS providers are thus free to introduce additional (non-INSECURE) algorithms without coercing other participating providers to do the same.¶
For zones with trust anchors, when there is a trust anchor with a UNIVERSAL algorithm, it is permissible to introduce a new trust anchor for a different algorithm before introducing the corresponding DNSKEY and RRSIGs into the zone. (Of course, they need to be added before the old trust anchor is removed.)¶
[This section needs to be updated to describe the construction of the new IANA registry for the implementation status and requirements of DNSSEC signing algorithms.]¶
The new validation requirements guarantee that when a zone is in a multi-signer setup with two algorithms, the security level is the same as it would be if the zone was in a single-signer setup using the weakest of them (from the resolver's perspective). This resolves undue SERVFAIL issues that could occur with certain algorithm combinations under the previous rules.¶
For example, a zone using only algorithm 7 is treated as insecure by resolvers that do not support this algorithm. When transferring the domain to another provider via a multi-signer setup with algorithm 13, the zone's security status remains "insecure", as the DS RRset still includes INSECURE algorithm 7. The presence of algorithm 13 is inconsequential at this point. Only once algorithm 7 is removed, the zone turns secure.¶
This rule prevents validation breakage when the resolver encounters an unsupported RRSIG from an outdated algorithm, and instead acknowledges the fact that the signer is using an algorithm that is in "MUST NOT sign" status, which (depending on resolver support) might render the zone insecure. This allows for glitch-free algorithm upgrades, with the security status of the zone changing only once the transition is complete.¶
Resolvers supporting both algorithms retain full validation throughtout the transition. In case of a permanent multi-signer setup, the zone maintainer needs to upgrade the INSECURE algorithm to a UNIVERSAL one in order to restore universal validation.¶
The same situation occurs when an algorithm is removed from the set of UNIVERSAL algorithms. In this case, the algorithm will enter "MUST NOT sign" status and become INSECURE. If the zone continues to use the INSECURE algorithm, it will continue to fully validate with supporting resolvers, while non-supporting resolvers will treat the zone as insecure until the algorithm is replaced.¶
Conversely, when an algorithm is added to the set of UNIVERSAL ones, it is conceivable that a signer may move to this algorithm before all validators are upgraded. This is, in fact, not a problem, as resolvers do not need to know the concept of UNIVERSAL. A problem could only occur if the corresponding RRSIG was not supported by the resolver; however, in that case labeling the algorithm as UNIVERSAL would have been premature. Determining universal support cannot be solved on the protocol level, and it is the community's responsibility to only advance an algorithm to UNIVERSAL if safe enough, i.e. if the number of resolvers lacking support it is deemed negligible.¶
In any case, regardless of "who moves first", resolution is never disrupted, and changes to the set of UNIVERSAL algorithms do not trigger overly conservative SERVFAIL responses.¶
Resolvers dropping support for INSECURE algorithms (e.g. 7) without implementing this specification will produce SERVFAIL responses for multi-signer setups involving the disabled algorithm. Implementation of the new validation rules is thus advised as soon as support for an algorithm is dropped.¶
The author would like to thank Shumon Huque and Viktor Dukhovni for early feedback on this proposal. It was developed after discussions on the problem space with Edward Lewis, Jakob Schlyter, Johan Stenstam, Steve Crocker, whose contributions where both insightful and helpful.¶
[RFC4035] Section 2.2 specifies the RRSIG presence requirements as follows:¶
There MUST be an RRSIG for each RRset using at least one DNSKEY of each algorithm in the zone apex DNSKEY RRset. The apex DNSKEY RRset itself MUST be signed by each algorithm appearing in the DS RRset located at the delegating parent (if any).¶
Further, Section 5.11 of [RFC6840] clarifies:¶
A signed zone MUST include a DNSKEY for each algorithm present in the zone's DS RRset and expected trust anchors for the zone.¶
It may seem tempting to just relax this rule, without any further adjustments. However, doing so is not safe depending on the algorithm combination involved. In particular, when using an algorithm that is not universally supported among the resolver population (such as algorithm 7) together with a supported one (such as algorithm 13), resolvers may return SERVFAIL under certain circumstances.¶
More explicitly, a zone that is using some algorithm as its sole signing algorithm is (correctly) treated as insecure by resolvers that do not support that algorithm. However, when attempting to transfer the domain to another DNS provider through a multi-signer setup with a supported algorithm, affected resolvers presented with the unsupported signature only will not be able to distinguish this situation from a downgrade-to-insecure attack where the second signature has been stripped, and will return SERVFAIL.¶
Zone owners and signers thus would have to take great care to not leave a validating resolver without a valid supported path when transitioning e.g. from algorithm 7 to 13.¶
In general (according to the old requirements), when a validating resolver supporting any of the algorithms listed in a given zone's DS record or TA set responds to a query without the CD flag set, it may not treat that zone as insecure, but must return either validated data (AD=1) or RCODE=2 (SERVFAIL). For this purpose, any valid path suffices; the validator may not apply a "logical AND" approach to all advertised algorithms.¶
Accordingly, [RFC6840] Section 5.11 states:¶
This requirement applies to servers, not validators. Validators SHOULD accept any single valid path. They SHOULD NOT insist that all algorithms signaled in the DS RRset work, and they MUST NOT insist that all algorithms signaled in the DNSKEY RRset work.¶
At first glance, the assertions that (1) the signer provide signatures for all advertised algorithms while (2) the resolver shall be content with just one seems somewhat contradictory. However, the role of the RRSIG rules is to ensure that the resolver will find a valid path (using a "logical OR" strategy), regardless of which particular algorithm(s) it supports, and thus be able to distinguish reliably between "all is in order" (validated data) and a downgrade-to-insecure attack (SERVFAIL).¶
With the new notion of UNIVERSAL algorithms, the same goal can be achieved with less stringent signing and slightly modified validation rules (see above).¶
Initial public draft.¶