RFC 9471
DNS Glue Requirements in Referral Responses
RFC 9471: DNS Glue Requirements in Referral Responses ✨
Okay bestie, let's get into RFC 9471 💖 It's an IETF Standards Track document from September 2023, and it's giving very official, very important energy 🎀. It updates RFC 1034, and the header doesn't list anything it obsoletes. The authors are M. Andrews (ISC), S. Huque (Salesforce), P. Wouters (Aiven), and D. Wessels (Verisign) 👑. The ISSN is 2070-1721 🦋.
Abstract
The DNS uses glue records to let iterative clients find the addresses of name servers that sit inside a delegated zone 🧚. Authoritative servers are expected to hand back every available glue record for in-domain name servers in a referral response. If message size limits mean not all of those glue records can fit, the server must set the TC (Truncated) flag 🚩. That tells the client the response is incomplete and that it should try another transport to get the full thing 💅. This document updates RFC 1034 to clarify what correct server behavior looks like ✨.
Status of This Memo
This is an Internet Standards Track document 🌈.
It's a product of the Internet Engineering Task Force (IETF), so it represents the consensus of the IETF community 🫶. It went through public review and was approved for publication by the Internet Engineering Steering Group (IESG). If you want more on Internet Standards, check out Section 2 of RFC 7841 📚.
For the current status, any errata, and how to send feedback, go to https://www.rfc-editor.org/info/rfc9471 🔗.
Copyright Notice
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) as they stand on the publication date. Please read these carefully, because they spell out your rights and restrictions for this document 📜. Code Components pulled out of this document must include the Revised BSD License text, as described in Section 4.e of the Trust Legal Provisions, and they come with no warranty, as described in the Revised BSD License ⚖️.
Table of Contents
1. Introduction
1.1. Requirements Language
2. Types of Glue in Referral Responses
2.1. Glue for In-Domain Name Servers
2.2. Glue for Sibling Domain Name Servers
2.3. Glue for Cyclic Sibling Domain Name Servers
2.4. Missing Glue
3. Requirements
3.1. Glue for In-Domain Name Servers
3.2. Glue for Sibling Domain Name Servers
3.3. Update to RFC 1034
4. Security Considerations
5. Operational Considerations
6. IANA Considerations
7. References
7.1. Normative References
7.2. Informative References
Acknowledgements
Authors' Addresses1. Introduction 🌸
Okay so the Domain Name System (DNS) [RFC1034] [RFC1035] uses glue records to let iterative clients find the addresses of name servers that live inside a delegated zone 🦋. Glue records get added to the parent zone as part of delegation and are returned in referral responses. Without them, a resolver following the referral has no way of finding those addresses 😭.
Authoritative servers are expected to return all available glue records for in-domain name servers in a referral response. If message size constraints stop all of those glue records from fitting over the chosen transport, the server MUST set the TC (Truncated) flag to tell the client that the response is incomplete. The client SHOULD use another transport to get the full response 💖. This document just clarifies that expectation ✨.
DNS responses sometimes carry optional data in the additional section. In-domain glue records, though, are NOT optional 🙅♀️. Several other protocol extensions are also not optional when they're used. That includes TSIG [RFC8945], OPT [RFC6891], and SIG(0) [RFC2931] 🎀.
As of this writing, addresses (A or AAAA records) for a delegation's authoritative name servers are the only type of glue defined for the DNS 💅.
Note that this document only clarifies requirements for name server software implementations 💻. It doesn't introduce or change any requirements about the data placed in DNS zones or registries. In other words, it only makes requirements about "available glue records" (meaning the ones given in a zone), and it says nothing about whether they must be present in a zone 🙈. If some glue records are missing from a given zone, an authoritative name server may be unable to return a useful referral response for the corresponding domain 😬. The IETF may want to consider a separate update to the requirements for including glue in zone data, beyond what's given in [RFC1034] and [RFC1035] 👑.
This document assumes you're reasonably familiar with DNS operations and protocol terms. Much of the terminology is explained in more detail in "DNS Terminology" [RFC8499] 📚.
1.1. Requirements Language
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 🫶.
2. Types of Glue in Referral Responses 🧚
This section describes the different types of glue you might find in DNS referral responses 🌈. Here's the tea: the type of glue depends on the QNAME. A particular name server (and its matching glue record) can be in-domain for one response and in a sibling domain for another 🔥.
2.1. Glue for In-Domain Name Servers
Here's a simple example of glue records in the delegating zone "test" for the child zone "foo.test". The name servers for foo.test (ns1.foo.test and ns2.foo.test) are both below the delegation point. They're configured as glue records in the "test" zone 💅:
foo.test. 86400 IN NS ns1.foo.test.
foo.test. 86400 IN NS ns2.foo.test.
ns1.foo.test. 86400 IN A 192.0.2.1
ns2.foo.test. 86400 IN AAAA 2001:db8::2:2A referral response from "test" for "foo.test" with glue for in-domain name servers looks like this 🎀:
;; QUESTION SECTION:
;www.foo.test. IN A
;; AUTHORITY SECTION:
foo.test. 86400 IN NS ns1.foo.test.
foo.test. 86400 IN NS ns2.foo.test.
;; ADDITIONAL SECTION:
ns1.foo.test. 86400 IN A 192.0.2.1
ns2.foo.test. 86400 IN AAAA 2001:db8::2:22.2. Glue for Sibling Domain Name Servers
Sibling domain name servers are NS records that aren't contained in the delegated zone itself. Instead, they're contained in another zone delegated from the same parent 👯♀️. In lots of cases, glue for sibling domain name servers isn't strictly required for resolution, because the resolver can send follow-on queries to the sibling zone to resolve the name server addresses (after following the referral to that sibling zone) 🧠. That said, most name server implementations today provide them anyway as an optimization, so iterative resolvers don't need the extra traffic ⭐.
Here, the delegating zone "test" has two delegations, for the child zones "bar.test" and "foo.test" 🍓:
bar.test. 86400 IN NS ns1.bar.test.
bar.test. 86400 IN NS ns2.bar.test.
ns1.bar.test. 86400 IN A 192.0.2.1
ns2.bar.test. 86400 IN AAAA 2001:db8::2:2
foo.test. 86400 IN NS ns1.bar.test.
foo.test. 86400 IN NS ns2.bar.test.A referral response from "test" for "foo.test" with glue for sibling domain name servers looks like this 💜:
;; QUESTION SECTION:
;www.foo.test. IN A
;; AUTHORITY SECTION:
foo.test. 86400 IN NS ns1.bar.test.
foo.test. 86400 IN NS ns2.bar.test.
;; ADDITIONAL SECTION:
ns1.bar.test. 86400 IN A 192.0.2.1
ns2.bar.test. 86400 IN AAAA 2001:db8::2:22.3. Glue for Cyclic Sibling Domain Name Servers
Using sibling domain name servers can create cyclic dependencies 🌀. This happens when one domain specifies name servers from a sibling domain, and the sibling does the same back. A cyclic dependency like this can only be broken when the delegating name server includes glue for the sibling domain in a referral response 🔓. Honestly, it's a bit of a chaotic loop, bestie 😵💫.
Here, the delegating zone "test" has two delegations for the child zones "bar.test" and "foo.test", and each one uses name servers under the other 🦋:
bar.test. 86400 IN NS ns1.foo.test.
bar.test. 86400 IN NS ns2.foo.test.
ns1.bar.test. 86400 IN A 192.0.2.1
ns2.bar.test. 86400 IN AAAA 2001:db8::2:2
foo.test. 86400 IN NS ns1.bar.test.
foo.test. 86400 IN NS ns2.bar.test.
ns1.foo.test. 86400 IN A 192.0.2.3
ns2.foo.test. 86400 IN AAAA 2001:db8::2:4A referral response from "test" for "bar.test" with glue for sibling domain name servers looks like this ✨:
;; QUESTION SECTION:
;www.bar.test. IN A
;; AUTHORITY SECTION:
bar.test. 86400 IN NS ns1.foo.test.
bar.test. 86400 IN NS ns2.foo.test.
;; ADDITIONAL SECTION:
ns1.foo.test. 86400 IN A 192.0.2.3
ns2.foo.test. 86400 IN AAAA 2001:db8::2:4Now for the receipts 🧾. In late 2021, the authors analyzed zone file data from ICANN's Centralized Zone Data Service [CZDS] and found 222 out of approximately 209,000,000 total delegations that had only sibling domain NS Resource Records (RRs) in a cyclic dependency like the one above 🤯.
2.4. Missing Glue 🙈
Okay, this one's a bonus plot twist 😭 It's an example of missing glue, but it can't actually count as a type of glue. It's not common, but real responses that lack required glue and come back with TC=0 have been seen, and they've caused resolution failures 💀
The example below comes from the dig command [DIG] and is based on a response observed in June 2020. The names were changed so they live under documentation domains 🎀. It shows a case where none of the glue records in the zone fit into the space available in the UDP response, and the TC flag was not set. This example shows a referral with DNSSEC records [RFC4033] [RFC4034] [RFC4035], but this behavior has also been seen with plain DNS responses. Some records are truncated for display purposes. At the time of writing, the servers originally responsible for this example have been updated and now correctly set the TC flag ✨
% dig +norec +dnssec +bufsize=512 +ignore @ns.example.net \
rh202ns2.355.foo.example
; <<>> DiG 9.15.4 <<>> +norec +dnssec +bufsize +ignore \
@ns.example.net rh202ns2.355.foo.example
; (2 servers found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 8798
;; flags: qr; QUERY: 1, ANSWER: 0, AUTHORITY: 9, ADDITIONAL: 1
;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags: do; udp: 4096
;; QUESTION SECTION:
;rh202ns2.355.foo.example. IN A
;; AUTHORITY SECTION:
foo.example. 86400 IN NS rh120ns2.368.foo.example.
foo.example. 86400 IN NS rh202ns2.355.foo.example.
foo.example. 86400 IN NS rh120ns1.368.foo.example.
foo.example. 86400 IN NS rh202ns1.355.foo.example.
foo.example. 3600 IN DS 51937 8 1 ...
foo.example. 3600 IN DS 635 8 2 ...
foo.example. 3600 IN DS 51937 8 2 ...
foo.example. 3600 IN DS 635 8 1 ...
foo.example. 3600 IN RRSIG DS 8 2 3600 ...3. Requirements 📋
This section is where the updated glue rules live 💅
3.1. Glue for In-Domain Name Servers
This document clarifies that when a name server generates a referral response, it MUST include all available glue records for in-domain name servers in the additional section or MUST set TC=1 if constrained by message size. Translation: no slacking on in-domain glue, and if it doesn't fit, flag it with TC=1 🚩
At the time of writing, most iterative clients send initial queries over UDP and retry over TCP when they get a response with the TC flag set 🔁 UDP responses are generally limited to between 1232 and 4096 bytes, because of the values commonly used for the EDNS0 UDP Message Size field [RFC6891] [FLAGDAY2020]. TCP responses are limited to 65,535 bytes 🌈
3.2. Glue for Sibling Domain Name Servers
This document clarifies that when a name server generates a referral response, it SHOULD include all available glue records in the additional section. If, after adding glue for all in-domain name servers, the glue for all sibling domain name servers still doesn't fit because of message size constraints, the name server MAY set TC=1, but it's not obligated to do so 🦋
Note that users may experience resolution failures for domains with cyclically dependent sibling name servers when the delegating name server chooses to omit the corresponding glue in a referral response. As described in Section 2.3, such domains are rare 🙈
3.3. Update to RFC 1034
Here's the glow-up for RFC 1034 ✨ The OLD text first:
| Copy the NS RRs for the subzone into the authority section of the
| reply. Put whatever addresses are available into the additional
| section, using glue RRs if the addresses are not available from
| authoritative data or the cache. Go to step 4.And the NEW text:
| Copy the NS RRs for the subzone into the authority section of the
| reply. Put whatever NS addresses are available into the
| additional section, using glue RRs if the addresses are not
| available from authoritative data or the cache. If all glue RRs
| for in-domain name servers do not fit, set TC=1 in the header. Go
| to step 4.The big difference is the new "If all glue RRs for in-domain name servers do not fit, set TC=1 in the header" line 🎀
4. Security Considerations 🚨
This document clarifies correct DNS server behavior and does not introduce any changes or new security considerations. No new drama here, bestie 💜 That said, the glue problems in Section 2.4 are the kind of thing that breaks resolution, so keep an eye on your servers 👀
5. Operational Considerations 🛠️
At the time of writing, most DNS server implementations only set the TC flag if none of the available glue records fit in a UDP response. The updated requirements in this document might increase the fraction of UDP responses with the TC flag set, and that could mean more queries arriving over TCP transport 📈 Plan for it, babe 💖
6. IANA Considerations 🌸
This document has no IANA actions. Nothing to do on the registry front, the end 💅
7. References 📚
7.1. Normative References
[RFC1034] Mockapetris, P., "Domain names - concepts and facilities",
STD 13, RFC 1034, DOI 10.17487/RFC1034, November 1987,
<https://www.rfc-editor.org/info/rfc1034>.
[RFC1035] Mockapetris, P., "Domain names - implementation and
specification", STD 13, RFC 1035, DOI 10.17487/RFC1035,
November 1987, <https://www.rfc-editor.org/info/rfc1035>.
[RFC2119] Bradner, S., "Key words for use in RFCs to Indicate
Requirement Levels", BCP 14, RFC 2119,
DOI 10.17487/RFC2119, March 1997,
<https://www.rfc-editor.org/info/rfc2119>.
[RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC
2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174,
May 2017, <https://www.rfc-editor.org/info/rfc8174>.7.2. Informative References
[CZDS] ICANN, "Centralized Zone Data Service",
<https://czds.icann.org/>.
[DIG] Wikipedia, "dig (command)", September 2023,
<https://en.wikipedia.org/wiki/Dig_(command)>.
[FLAGDAY2020]
Various DNS software and service providers, "DNS Flag Day
2020", October 2020, <https://dnsflagday.net/2020/>.
[RFC2931] Eastlake 3rd, D., "DNS Request and Transaction Signatures
( SIG(0)s )", RFC 2931, DOI 10.17487/RFC2931, September
2000, <https://www.rfc-editor.org/info/rfc2931>.
[RFC4033] Arends, R., Austein, R., Larson, M., Massey, D., and S.
Rose, "DNS Security Introduction and Requirements",
RFC 4033, DOI 10.17487/RFC4033, March 2005,
<https://www.rfc-editor.org/info/rfc4033>.
[RFC4034] Arends, R., Austein, R., Larson, M., Massey, D., and S.
Rose, "Resource Records for the DNS Security Extensions",
RFC 4034, DOI 10.17487/RFC4034, March 2005,
<https://www.rfc-editor.org/info/rfc4034>.
[RFC4035] Arends, R., Austein, R., Larson, M., Massey, D., and S.
Rose, "Protocol Modifications for the DNS Security
Extensions", RFC 4035, DOI 10.17487/RFC4035, March 2005,
<https://www.rfc-editor.org/info/rfc4035>.
[RFC6891] Damas, J., Graff, M., and P. Vixie, "Extension Mechanisms
for DNS (EDNS(0))", STD 75, RFC 6891,
DOI 10.17487/RFC6891, April 2013,
<https://www.rfc-editor.org/info/rfc6891>.
[RFC8499] Hoffman, P., Sullivan, A., and K. Fujiwara, "DNS
Terminology", BCP 219, RFC 8499, DOI 10.17487/RFC8499,
January 2019, <https://www.rfc-editor.org/info/rfc8499>.
[RFC8945] Dupont, F., Morris, S., Vixie, P., Eastlake 3rd, D.,
Gudmundsson, O., and B. Wellington, "Secret Key
Transaction Authentication for DNS (TSIG)", STD 93,
RFC 8945, DOI 10.17487/RFC8945, November 2020,
<https://www.rfc-editor.org/info/rfc8945>.Acknowledgements 💌
The authors want to thank Joe Abley, David Blacka, Brian Dickson, Kazunori Fujiwara, Paul Hoffman, Geoff Huston, John R. Levine, Jared Mauch, George Michaelson, Yasuhiro Orange Morishita, Benno Overeinder, Hugo Salgado, Shinta Sato, Puneet Sood, Petr Spacek, Ralf Weber, Tim Wicinski, Suzanne Woolf, and other members of the DNSOP Working Group for their input. Thank you all, you're literally the reason this came together 🫶
Authors' Addresses 🦋
M. Andrews
ISC
Email: marka@isc.org
Shumon Huque
Salesforce
Email: shuque@gmail.com
Paul Wouters
Aiven
Email: paul.wouters@aiven.io
Duane Wessels
Verisign
Email: dwessels@verisign.com📦 where this page comes from
- ietf: rfc9471 2023-09
- 💖 girlified by claude-haiku-5-5 on : a language model's rewrite of the original, under its terms