Establishing Local DNS Authority in Split-Horizon Environments

To resolve a DNS query, there are three essential behaviors that an implementation can apply: (1) answer from a local database, (2) query the relevant authorities and their parents, or (3) ask a server to query those authorities and return the final answer. Implementations that use these behaviors are called "authoritative nameservers", "full resolvers", and "forwarders" (or "stub resolvers"). However, an implementation can also implement a mixture of these behaviors, depending on a local policy, for each query. We term such an implementation a "hybrid resolver". Most DNS resolvers are hybrids of some kind. For example, stub resolvers frequently support a local "hosts file" that preempts query forwarding, and most DNS forwarders and full resolvers can also serve responses from a local zone file. Other standardized hybrid resolution behaviors include Local Root , mDNS , and NXDOMAIN synthesis for .onion . In many network environments, the network offers clients a DNS server (e.g. DHCP OFFER, IPv6 Router Advertisement). Although this server is formally specified as a recursive resolver (e.g. Section 5.1 of ), some networks provide a hybrid resolver instead. If this resolver acts as an authoritative server for some names, we say that the network has "split-horizon DNS", because those names resolve in this way only from inside the network. Network clients that use pure stub resolution, sending all queries to the network-provided resolver, will always receive the split-horizon results. Conversely, clients that send all queries to a different resolver or implement pure full resolution locally will never receive them. Clients with either pure resolution behavior are out of scope for this specification. Instead, this specification enables hybrid clients to access split-horizon results from a network-provided hybrid resolver, while using a different resolution method for some or all other names. There are several existing mechanisms for a network to provide clients with "local domain hints", listing domain names that have special treatment in this network (). However, none of the local domain hint mechanisms enable clients to determine whether this special treatment is authorized by the domain owner. Instead, these specifications require clients to make their own determinations about whether to trust and rely on these hints. This specification describes a protocol between domains, networks, and clients that allows the network to establish its authority over a domain to a client (). Clients can use this protocol to confirm that a local domain hint was authorized by the domain (), which might influence its processing of that hint. This specification relies on securely identified local DNS servers and globally valid NS records. Use of this specification is therefore limited to servers that support authenticated encryption and split-horizon DNS names that are properly rooted in the global DNS.

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. This document makes use of the terms defined in . The term "Global DNS" is defined in . 'Encrypted DNS' refers to a DNS protocol that provides an encrypted channel between a DNS client and server (e.g., DoT, DoH, or DoQ). The term 'Validated Split-Horizon' is also defined.

A split horizon configuration for some name is considered "validated" if the network client has confirmed that a parent of that name has authorized the local network to serve its own responses for that name. Such authorization generally extends to the entire subtree of names below the authorization point.

The protocol in this document allows the domain owner to create a split-horizon DNS. Other entities which do not own the domain are detected by the client. Thus, DNS filtering is not enabled by this protocol.

There are various mechanisms by which a network client might learn "local domain hints", which indicate a special treatment for particular domain names upon joining a network. This section provides a review of some common and standardized mechanisms for receiving domain hints.

There are several DHCP options that convey local domain hints of different kinds. The most directly relevant is "RDNSS Selection" , which provides "a list of domains ... about which the RDNSS has special knowledge", along with a "High", "Medium", or "Low" preference for each name. The specification notes the difficulty of relying on these hints without validation: Trustworthiness of an interface and configuration information received over the interface is implementation and/or node deployment dependent, and the details of determining that trust are beyond the scope of this specification. Other local domain hints in DHCP include the "Domain Name" , "Access Network Domain Name" , "Client FQDN" , and "Name Service Search" options. This specification may help clients to interpret these hints. For example, a rogue DHCP server could use the "Client FQDN" option to assign a client the name "www.example.com" in order to prevent the client from reaching the true "www.example.com". A client could use this specification to check the network's authority over this name, and adjust its behavior to avoid this attack if authority is not established. The Domain Search option , which offers clients a way to expand short names into Fully Qualified Domain Names, is not a "local domain hint" by this definition, because it does not modify the processing of any specific domain. (The specification notes that this option can be a "fruitful avenue of attack for a rogue DHCP server", and provides a number of cautions against accepting it unconditionally.)

A host can be configured with DNS information when it joins a network, including when it brings up VPN (which is also considered joining a(n additional) network, detailed in ). Existing implementations determine the host has joined a certain network via SSID, IP subnet assigned, DNS server IP address or name, and other similar mechanisms. For example, one existing implementation determines the host has joined an internal network because the DHCP-assigned IP address belongs to the company's IP address (as assigned by the regional IP addressing authority) and the DHCP-advertised DNS IP address is one used by IT at that network. Other mechanisms exist in other products but are not interesting to this specification; rather what is interesting is this step to determine "we have joined the internal corporate network" occurred and the DNS server is configured as authoritative for certain DNS zones (e.g., *.example.com). Because a rogue network can simulate all or most of the above characteristics this specification details how to validate these claims in .

Provisioning Domains (PvDs) are defined in as sets of network configuration information that clients can use to access networks, including rules for DNS resolution and proxy configuration. The PvD Key "dnsZones" is defined in as a list of "DNS zones searchable and accessible" in this provisioning domain. Attempting to resolve these names via another resolver might fail or return results that are not correct for this network.

In IKEv2 VPNs, the INTERNAL_DNS_DOMAIN configuration attribute can be used to indicate that a domain is "internal" to the VPN . To prevent abuse, the specification notes various possible restrictions on the use of this attribute: "If a client is configured by local policy to only accept a limited set of INTERNAL_DNS_DOMAIN values, the client MUST ignore any other INTERNAL_DNS_DOMAIN values." "IKE clients MAY want to require whitelisted domains for Top-Level Domains (TLDs) and Second-Level Domains (SLDs) to further prevent malicious DNS redirections for well-known domains." Within these guidelines, a client could adopt a local policy of accepting INTERNAL_DNS_DOMAIN values only when it can validate the local DNS server's authority over those names as described in this specification.

To establish its authority over some DNS zone, a participating network MUST offer one or more encrypted resolvers via DNR or an equivalent mechanism (see ). At least one of these resolvers' Authentication Domain Names (ADNs) MUST appear in an NS record for that zone. This arrangement establishes this resolver's authority over the zone.

To validate the network's authority over a domain name, participating clients MUST resolve the NS record for that name. If the resolution result is NODATA, the client MUST remove the last label and repeat the query until a response other than NODATA is received. Once the NS record has been resolved, the client MUST check if each local encrypted resolver's Authentication Domain Name appears in the NS record. The client SHALL regard each such resolver as authoritative for the zone of this NS record. Each validation of authority applies only to the specific resolvers whose names appear in the NS RRSet. If a network offers multiple encrypted resolvers, each DNS entry may be authorized for a distinct subset of the network-provided resolvers. A zone is termed a "Validated Split-Horizon zone" after successful validation using a "tamperproof" NS resolution method, i.e. a method that is not subject to interference by the local network operator. Two possible tamperproof resolution methods are presented below.

The client sends the NS query to a pre-configured resolver that is external to the network, over a secure transport. Clients SHOULD apply whatever acceptance rules they would otherwise apply when using this resolver (e.g. checking the AD bit, validating RRSIGs).

The client resolves the NS record using any resolution method of its choice (e.g. querying one of the network-provided resolvers, performing iterative resolution locally), and performs full DNSSEC validation locally . The result is processed based on its DNSSEC validation state (Section 4.3 of ): the response is used for validation. the response is rejected and validation is considered to have failed. the client SHOULD retry the validation process using a different method, such as the one in , to ensure compatibility with unsigned names.

Two examples are shown below. The first example showing an company with an internal-only DNS server resolving the entire zone for that company (e.g., *.example.com) the second example resolving only a subdomain of the company's zone (e.g., *.internal.example.com).

Consider an organization that operates "example.com", and runs a different version of its global domain on its internal network. Today, on the Internet it publishes two NS records, "ns1.example.com" and "ns2.example.com". The host and network first need mutual support one of the mechanisms described in learning. Shown in is learning using DNR and PvD. Validation is then perfomed using either Public DNS or DNSSEC. The client determines the network's DNS server (ns1.example.com) and Provisioning Domain (pvd.example.com) using DNR and PvD, using one of DNR Router Solicitation, DHCPv4, or DHCPv6. The client connects to the DNR-learned DNS server (ns1.example.com), validates its certificate, and queries for pvd.example.com. The client connects to the PvD server, validates its certificate, and retrieves the provisioning domain JSON information indicated by the associated PvD. The PvD contains:

The JSON keys "identifier", "expires", and "prefixes" are defined in .

| | | | | | Response with DNR hostnames & | | | | PvD FQDN (2) | | | |<-----------------------------------------------------------| | ----------------------------\ | | | |-| now knows DNR hostnames & | | | | | | PvD FQDN | | | | | |---------------------------/ | | | | | | | | TLS connection to ns1.example.com (3) | | |------------------------------------>| | | | ---------------------------\ | | | |-| validate TLS certificate | | | | | |--------------------------| | | | | | | | | resolve pvd.example.com (4) | | | |------------------------------------>| | | | | | | | A or AAAA records (5) | | | |<------------------------------------| | | | | | | | https://pvd.example.com/.well-known/pvd (6) | | |---------------------------------------------->| | | | | | | 200 OK (JSON Additional Information) (7) | | |<----------------------------------------------| | | -----------------------\ | | | |-| dnsZones=example.com | | | | | |----------------------| | | | ]]>

The figure below shows the steps performed to verify the local claims of DNS authority using a public resolver. The client uses an encrypted DNS connection to a public resolver (e.g., 1.1.1.1) to issue NS queries for the domains in dnsZones. The NS lookup for "example.com" will return "ns1.example.com" and "ns2.example.com". As the network-provided nameservers are the same as the names retrieved from the public resolver and the network-designated resolver's certificate includes at least one of the names retrieved from the public resolver, the client has finished validation that the nameservers signaled in and are owned and managed by the same entity that published the NS records on the Internet. The endpoint will then use that information from and to resolve names within dnsZones.

| | ---------------------------\ | | |-| validate TLS certificate | | | | |--------------------------| | | | | | | NS? example.com (1) | | |--------------------------------------------------->| | | | | NS=ns1.example.com, ns2.example.com (2) | | |<---------------------------------------------------| | -------------------------------\ | | |-| both DNR ADNs are authorized | | | | ----------------------\--------| | | |-| finished validation | | | | |---------------------| | | | | | | use network-designated resolver | | | for example.com (3) | | |----------------------------------------->| | | | | ]]>

The figure below shows the steps performed to verify the local claims of DNS authority using DNSSEC. The DNSSEC-validating client queries the network encrypted resolver to issue NS queries for the domains in dnsZones. The NS lookup for "example.com" will return a signed response containing "ns1.example.com" and "ns2.example.com". The client then performs full DNSSEC validation locally. As the DNSSEC validation is successful and the network-provided nameservers are the same as the names in the DNSSEC response, and the network-designated resolver's certificate includes at least one of the names returned in the DNSSEC response, the client has finished validation that the nameservers signaled in and are owned and managed by the same entity that published the NS records on the Internet. The endpoint will then use that information from and to resolve names within dnsZones.

| | | | NS=ns1.example.com,ns2.example.com, Signed Answer (RRSIG) (2) | |<---------------------------------------------------------------| | -----------------------------------\ | |-| DNSKEY+NS matches RRSIG, use NS | | | |----------------------------------| | | -------------------------------\ | |-| both DNR ADNs are authorized | | | |------------------------------| | | ----------------------\ | |-| finished validation | | | |---------------------| | | | | use encrypted network-designated resolver for example.com (3) | |--------------------------------------------------------------->| | | ]]>

A subdomain can also be used for all internal DNS names (e.g., the zone internal.example.com exists only on the internal DNS server). For successful validation described in this document the the internal DNS server will need a certificate signed by a CA trusted by the client. For such a name internal.example.com the message flow is similar to the difference is that queries for hosts not within the subdomain (www.example.com) are sent to the public resolver rather than resolver for internal.example.com.

When the VPN tunnel is IPsec, the encrypted DNS resolver hosted by the VPN service provider can be securely discovered by the endpoint using the ENCDNS_IP*_* IKEv2 Configuration Payload Attribute Types defined in . Other VPN tunnel types have similar configuration capabilities, not detailed here.

This specification does not alter DNSSEC validation behaviour. To ensure compatibility with validating clients, network operators MUST ensure that names under the split-horizon are correctly signed or place them in an unsigned zone. If an internal zone name (e.g., internal.example.com) is used with in conjunction with this specification and a public certificate is obtained for validation, that internal zone name will exist in Certificate Transparency logs. It should be noted, however, that this specification does not leak individual host names (e.g., www.internal.example.com) into the Certificate Transparancy logs or to public DNS resolvers.

This document has no IANA actions.

Thanks to Mohamed Boucadair, Jim Reid, Tommy Pauly, Paul Vixie, Paul Wouters and Vinny Parla for the discussion and comments.