Apple Inc.

One Apple Park Way Cupertino California 95014 USA mellon@fugue.com

Apple Inc.

One Apple Park Way Cupertino California 95014 USA +1 408 974 3207 cheshire@apple.com

Internet Internet Engineering Task Force Multicast DNS DNS-Based Service Discovery DNS Update SIG(0) RFC Request for Comments I-D Internet-Draft The Service Registration Protocol for DNS-Based Service Discovery uses the standard DNS Update mechanism to enable DNS-Based Service Discovery using only unicast packets. This makes it possible to deploy DNS Service Discovery without multicast, which greatly improves scalability and improves performance on networks where multicast service is not an optimal choice, particularly 802.11 (Wi&nbhy;Fi) and 802.15.4 (IoT) networks. DNS&nbhy;SD Service registration uses public keys and SIG(0) to allow services to defend their registrations against attack.

Introduction DNS-Based Service Discovery is a component of Zero Configuration Networking . This document describes an enhancement to DNS-Based Service Discovery that allows services to register their services using the DNS protocol rather than using Multicast DNS (mDNS). There is already a large installed base of DNS&nbhy;SD clients that can discover services using the DNS protocol. This document is intended for three audiences: implementors of software that provides services that should be advertised using DNS&nbhy;SD, implementors of DNS servers that will be used in contexts where DNS&nbhy;SD registration is needed, and administrators of networks where DNS&nbhy;SD service is required. The document is intended to provide sufficient information to allow interoperable implementation of the registration protocol. DNS-Based Service Discovery (DNS&nbhy;SD) allows services to advertise the fact that they provide service, and to provide the information required to access that service. DNS&nbhy;SD clients can then discover the set of services of a particular type that are available. They can then select a service from among those that are available and obtain the information required to use it. Although DNS-SD using the DNS protocol (as opposed to mDNS) can be more efficient and versatile, it is not common in practice, because of the difficulties associated with updating authoritative DNS services with service information. Existing practice for updating DNS zones is to either manually enter new data, or else use DNS Update . Unfortunately DNS Update requires either that the authoritative DNS server automatically trust updates, or else that the DNS Update client have some kind of shared secret or public key that is known to the DNS server and can be used to authenticate the update. Furthermore, DNS Update can be a fairly chatty process, requiring multiple round trips with different conditional predicates to complete the update process. The SRP protocol adds a set of default heuristics for processing DNS updates that eliminates the need for DNS update conditional predicates: instead, the SRP server has a set of default predicates that are applied to the update, and the update either succeeds entirely, or fails in a way that allows the registering service to know what went wrong and construct a new update. SRP also adds a feature called First-Come, First-Served Naming, which allows the registering service to claim a name that is not yet in use, and, using SIG(0) , to authenticate both the initial claim and subsequent updates. This prevents name conflicts, since a second SRP service attempting to claim the same name will not possess the SIG(0) key used by the first service to claim it, and so its claim will be rejected and the second service will have to choose a new name. Finally, SRP adds the concept of a 'lease,' similar to leases in Dynamic Host Configuration Protocol . The SRP registration itself has a lease which may be on the order of an hour; if the registering service does not renew the lease before it has elapsed, the registration is removed. The claim on the name can have a longer lease, so that another service cannot claim the name, even though the registration has expired. The Service Registration Protocol for DNS&nbhy;SD (SRP), described in this document, provides a reasonably secure mechanism for publishing this information. Once published, these services can be readily discovered by DNS&nbhy;SD clients using standard DNS lookups. The DNS&nbhy;SD specification (“Populating the DNS with Information”), briefly discusses ways that services can publish their information in the DNS namespace. In the case of mDNS, it allows services to publish their information on the local link, using names in the ".local" namespace, which makes their services directly discoverable by peers attached to that same local link. RFC6763 also allows clients to discover services using the DNS protocol. This can be done by having a system administrator manually configure service information in the DNS, but manually populating DNS authoritative server databases is costly and potentially error-prone, and requires a knowledgable network administrator. Consequently, although all DNS&nbhy;SD client implementations of which we are aware support DNS&nbhy;SD using DNS queries, in practice it is used much less frequently than mDNS. The Discovery Proxy provides one way to automatically populate the DNS namespace, but is only appropriate on networks where services are easily advertised using mDNS. This document describes a solution more suitable for networks where multicast is inefficient, or where sleepy devices are common, by supporting both offering of services, and discovery of services, using unicast.

Service Registration Protocol Services that implement SRP use DNS Update to publish service information in the DNS. Two variants exist, one for full-featured hosts, and one for devices designed for "Constrained-Node Networks" . An SRP server is most likely an authoritative DNS server, or else is updating an authoritative DNS server. There is no requirement that the server that is receiving SRP requests be the same server that is answering queries that return records that have been registered.

Protocol Variants

Full-featured Hosts Full-featured hosts are either configured manually with a registration domain, or use the "dr._dns&nbhy;sd._udp.<domain>" query () to learn the default registration domain from the network. RFC6763 says to discover the registration domain using either ".local" or a network-supplied domain name for <domain>. Services using SRP MUST use the domain name received through the DHCPv4 Domain Name option (), if available, or the Neighbor Discovery DNS Search List option . If the DNS Search List option contains more than one domain name, it MUST NOT be used. If neither option is available, the Service Registration protocol is not available on the local network. Manual configuration of the registration domain can be done either by querying the list of available registration zones ("r._dns&nbhy;sd._udp") and allowing the user to select one from the UI, or by any other means appropriate to the particular use case being addressed. Full-featured devices construct the names of the SRV, TXT, and PTR records describing their service(s) as subdomains of the chosen service registration domain. For these names they then discover the zone apex of the closest enclosing DNS zone using SOA queries . Having discovered the enclosing DNS zone, they query for the "_dnssd&nbhy;srp._tcp.<zone>" SRV record to discover the server to which they should send DNS updates. Hosts that support SRP Updates using TLS use the "_dnssd&nbhy;srp&nbhy;tls._tcp.<zone>" SRV record instead.

Constrained Hosts For devices designed for Constrained-Node Networks some simplifications are available. Instead of being configured with (or discovering) the service registration domain, the (proposed) special-use domain name (see ) "default.service.arpa" is used. The details of how SRP server(s) are discovered will be specific to the constrained network, and therefore we do not suggest a specific mechanism here. SRP clients on constrained networks are expected to receive from the network a list of SRP servers with which to register. It is the responsibility of a Constrained-Node Network supporting SRP to provide one or more SRP server addresses. It is the responsibility of the SRP server supporting a Constrained-Node Network to handle the updates appropriately. In some network environments, updates may be accepted directly into a local "default.service.arpa" zone, which has only local visibility. In other network environments, updates for names ending in "default.service.arpa" may be rewritten internally to names with broader visibility.

Why two variants? The reason for these different assumptions is that low-power devices that typically use Constrained-Node Networks may have very limited battery power. The series of DNS lookups required to discover an SRP server and then communicate with it will increase the power required to advertise a service; for low-power devices, the additional flexibility this provides does not justify the additional use of power. It is also fairly typical of such networks that some network service information is obtained as part of the process of joining the network, and so this can be relied upon to provide nodes with the information they need. Networks that are not constrained networks can have more complicated topologies at the Internet layer. Nodes connected to such networks can be assumed to be able to do DNSSD service registration domain discovery. Such networks are generally able to provide registration domain discovery and routing. By requiring the use of TCP, the possibility of off-network spoofing is eliminated.

Protocol Details We will discuss several parts to this process: how to know what to publish, how to know where to publish it (under what name), how to publish it, how to secure its publication, and how to maintain the information once published.

What to publish We refer to the DNS Update message sent by services using SRP as an SRP Update. Three types of updates appear in an SRP update: Service Discovery records, Service Description records, and Host Description records.

• Service Discovery records are one or more PTR RRs, mapping from the generic service type (or subtype) to the specific Service Instance Name.
• Service Description records are exactly one SRV RR, exactly one KEY RR, and one or more TXT RRs, all with the same name, the Service Instance Name (). In principle Service Description records can include other record types, with the same Service Instance Name, though in practice they rarely do. The Service Instance Name MUST be referenced by one or more Service Discovery PTR records, unless it is a placeholder service registration for an intentionally non-discoverable service name.
• The Host Description records for a service are a KEY RR, used to claim exclusive ownership of the service registration, and one or more RRs of type A or AAAA, giving the IPv4 or IPv6 address(es) of the host where the service resides.

describes the details of what each of these types of updates contains, with the exception of the KEY RR, which is defined in . These RFCs should be considered the definitive source for information about what to publish; the reason for summarizing this here is to provide the reader with enough information about what will be published that the service registration process can be understood at a high level without first learning the full details of DNS&nbhy;SD. Also, the "Service Instance Name" is an important aspect of first-come, first-serve naming, which we describe later on in this document.

Where to publish it Multicast DNS uses a single namespace, ".local", which is valid on the local link. This convenience is not available for DNS&nbhy;SD using the DNS protocol: services must exist in some specific unicast namespace. As described above, full-featured devices are responsible for knowing in what domain they should register their services. Devices made for Constrained-Node Networks register in the (proposed) special use domain name "default.service.arpa", and let the SRP server handle rewriting that to a different domain if necessary.

How to publish it It is possible to issue a DNS Update that does several things at once; this means that it's possible to do all the work of adding a PTR resource record to the PTR RRset on the Service Name, and creating or updating the Service Instance Name and Host Description, in a single transaction. An SRP Update takes advantage of this: it is implemented as a single DNS Update message that contains a service's Service Discovery records, Service Description records, and Host Description records. Updates done according to this specification are somewhat different than regular DNS Updates as defined in RFC2136. The RFC2136 update process can involve many update attempts: you might first attempt to add a name if it doesn't exist; if that fails, then in a second message you might update the name if it does exist but matches certain preconditions. Because the registration protocol uses a single transaction, some of this adaptability is lost. In order to allow updates to happen in a single transaction, SRP Updates do not include update prerequisites. The requirements specified in are implicit in the processing of SRP Updates, and so there is no need for the service sending the SRP Update to put in any explicit prerequisites.

How DNS&nbhy;SD Service Registration differs from standard RFC2136 DNS Update DNS&nbhy;SD Service Registration is based on standard RFC2136 DNS Update, with some differences:

• It implements first-come first-served name allocation, protected using SIG(0) .
• It enforces policy about what updates are allowed.
• It optionally performs rewriting of "default.service.arpa" to some other domain.
• It optionally performs automatic population of the address-to-name reverse mapping domains.
• An SRP server is not required to implement general DNS Update prerequisite processing.
• Constrained-Node SRP clients are allowed to send updates to the generic domain "default.service.arpa"

How to secure it Traditional DNS update is secured using the TSIG protocol, which uses a secret key shared between the DNS Update client (which issues the update) and the server (which authenticates it). This model does not work for automatic service registration. The goal of securing the DNS&nbhy;SD Registration Protocol is to provide the best possible security given the constraint that service registration has to be automatic. It is possible to layer more operational security on top of what we describe here, but what we describe here is an improvement over the security of mDNS. The goal is not to provide the level of security of a network managed by a skilled operator.

First-Come First-Served Naming First-Come First-Serve naming provides a limited degree of security: a service that registers its service using DNS&nbhy;SD Registration protocol is given ownership of a name for an extended period of time based on the key used to authenticate the DNS Update. As long as the registration service remembers the name and the key used to register that name, no other service can add or update the information associated with that. FCFS naming is used to protect both the Service Description and the Host Description.

Service Behavior

Public/Private key pair generation and storage The service generates a public/private key pair. This key pair MUST be stored in stable storage; if there is no writable stable storage on the SRP client, the SRP client MUST be pre-configured with a public/private key pair in read-only storage that can be used. This key pair MUST be unique to the device. A device with rewritable storage should retain this key indefinitely. When the device changes ownership, it may be appropriate to erase the old key and install a new one. Therefore, the SRP client on the device SHOULD provide a mechanism to overwrite the key, for example as the result of a "factory reset." When sending DNS updates, the service includes a KEY record containing the public portion of the key in each Host Description Instruction and each Service Description Instruction. Each KEY record MUST contain the same public key. The update is signed using SIG(0), using the private key that corresponds to the public key in the KEY record. The lifetimes of the records in the update is set using the EDNS(0) Update Lease option . The KEY record in Service Description updates MAY be omitted for brevity; if it is omitted, the SRP server MUST behave as if the same KEY record that is given for the Host Description is also given for each Service Description for which no KEY record is provided. Omitted KEY records are not used when computing the SIG(0) signature.

Name Conflict Handling Both Host Description records and Service Description Records can have names that result in name conflicts. Service Discovery records cannot have name conflicts. If any Host Description or Service Description record is found by the server to have a conflict with an existing name, the server will respond to the SRP Update with a YXDOMAIN rcode. In this case, the Service MUST either abandon the service registration attempt, or else choose a new name. There is no specific requirement for how this is done; typically, however, the service will append a number to the preferred name. This number could be sequentially increasing, or could be chosen randomly. One existing implementation attempts several sequential numbers before choosing randomly. So for instance, it might try host.service.arpa, then host-1.service.arpa, then host-2.service.arpa, then host-31773.service.arpa.

Record Lifetimes The lifetime of the DNS&nbhy;SD PTR, SRV, A, AAAA and TXT records uses the LEASE field of the Update Lease option, and is typically set to two hours. This means that if a device is disconnected from the network, it does not appear in the user interfaces of devices looking for services of that type for too long. The lifetime of the KEY records is set using the KEY-LEASE field of the Update Lease Option, and should be set to a much longer time, typically 14 days. The result of this is that even though a device may be temporarily unplugged, disappearing from the network for a few days, it makes a claim on its name that lasts much longer. This means that even if a device is unplugged from the network for a few days, and its services are not available for that time, no other device can come along and claim its name the moment it disappears from the network. In the event that a device is unplugged from the network and permanently discarded, then its name is eventually cleaned up and made available for re-use.

Compression in SRV records Although requires that the target name in the SRV record not be compressed, an SRP client SHOULD compress the target in the SRV record. The motivation for not compressing in RFC2782 is not stated, but is assumed to be because a caching resolver that does not understand the format of the SRV record might store it as binary data and thus return an invalid pointer in response to a query. This does not apply in the case of SRP: an SRP server needs to understand SRV records in order to validate the SRP Update. Compression of the target potentially saves substantial space in the SRP Update.

Removing published services

Removing all published services To remove all the services registered to a particular host, the SRP client retransmits its most recent update with an Update Lease option that has a LEASE value of zero. If the registration is to be permanently removed, KEY-LEASE should also be zero. Otherwise, it should have the same value it had previously; this holds the name in reserve for when the SRP client is once again able to provide the service. SRP clients are normally expected to remove all service instances when removing a host. However, in some cases a SRP client may not have retained sufficient state to know that some service instance is pointing to a host that it is removing. This method of removing services is intended for the case where the client is going offline and does not want its services advertised. Therefore, it is sufficient for the client to send the Host Description Instruction. To support this, when removing services based on the lease time being zero, an SRP server MUST remove all service instances pointing to a host when a host is removed, even if the SRP client doesn't list them explicitly. If the key lease time is nonzero, the SRP server MUST NOT delete the KEY records for these SRP clients.

Removing some published services In some use cases a client may need to remove some specific service, without removing its other services. This can be accomplished in one of two ways. To simply remove a specific service, the client sends a valid SRP Update where the Service Discovery Instruction contains a single Delete an RR from an RRset () update that deletes the PTR record whose target is the service instance name. The Service Description Instruction in this case contains a single Delete all RRsets from a Name () update to the service instance name. The second alternative is used when some service is being replaced by a different service with a different service instance name. In this case, the old service is deleted as in the first alternative. The new service is added, just as it would be in an update that wasn't deleting the old service. Because both the removal of the old service and the add of the new service consist of a valid Service Discovery Instruction and a valid Service Description Instruction, the update as a whole is a valid SRP Update, and will result in the old service being removed and the new one added, or, to put it differently, in the old service being replaced by the new service. It is perhaps worth noting that if a service is being updated without the service instance name changing, that will look very much like the second alternative above. The difference is that because the target for the PTR record in the Service Discovery Instruction is the same for both the Delete An RR From An RRset update and the Add To An RRSet update, these will be seen as a single Service Description Instruction, not as two Instructions. The same would be true of the Service Description Instruction. Whichever of these two alternatives is used, the host lease will be updated with the lease time provided in the SRP update. In neither of these cases is it permissible to delete the host. All services must point to a host. If a host is to be deleted, this must be done using the method described in , which deletes the host and all services that have that host as their target.

Validation and Processing of SRP Updates

Validation of Adds and Deletes The SRP server first validates that the DNS Update is a syntactically and semantically valid DNS Update according to the rules specified in RFC2136. SRP Updates consist of a set of instructions that together add or remove one or more services. Each instruction consists of some combination of delete updates and add updates. When an instruction contains a delete and an add, the delete MUST precede the add. The SRP server checks each instruction in the SRP Update to see that it is either a Service Discovery Instruction, a Service Description Instruction, or a Host Description Instruction. Order matters in DNS updates. Specifically, deletes must precede adds for records that the deletes would affect; otherwise the add will have no effect. This is the only ordering constraint; aside from this constraint, updates may appear in whatever order is convenient when constructing the update. Because the SRP Update is a DNS update, it MUST contain a single question that indicates the zone to be updated. Every delete and update in an SRP Update MUST be within the zone that is specified for the SRP Update.

Service Discovery Instruction An instruction is a Service Discovery Instruction if it contains

• exactly one "Add to an RRSet" or exactly one "Delete an RR from an RRSet" () RR update,
• which updates a PTR RR,
• the target of which is a Service Instance Name
• for which name a Service Description Instruction is present in the SRP Update
• if the Service Discovery Instruction is an "Add to an RRSet" instruction, the Service Description Instruction does not match if it does not contain an "Add to an RRset" update for the SRV RR describing that service.
• if the Service Discovery Instruction is a "Delete an RR from an RRSet" update, the Service Description Instruction does not match if it contains an "Add to an RRset" update.
• Service Discovery Instructions do not contain any other add or delete updates.

Service Description Instruction An instruction is a Service Description Instruction if, for the appropriate Service Instance Name, it contains

• exactly one "Delete all RRsets from a name" update for the service instance name (),
• zero or one "Add to an RRset" SRV RR,
• zero or one "Add to an RRset" KEY RR that, if present, contains the public key corresponding to the private key that was used to sign the message (if present, the KEY MUST match the KEY RR given in the Host Description),
• zero or more "Add to an RRset" TXT RRs,
• If there is one "Add to an RRset" SRV update, there MUST be at least one "Add to an RRset" TXT update.
• the target of the SRV RR Add, if present points to a hostname for which there is a Host Description Instruction in the SRP Update, or
• if there is no "Add to an RRset" SRV RR, then either
• • the name to which the "Delete all RRsets from a name" applies does not exist, or
  • there is an existing KEY RR on that name, which matches the key with which the SRP Update was signed.
• Service Descriptions Instructions do not modify any other resource records.

An SRP server MUST correctly handle compressed names in the SRV target.

Host Description Instruction An instruction is a Host Description Instruction if, for the appropriate hostname, it contains

• exactly one "Delete all RRsets from a name" RR,
• one or more "Add to an RRset" RRs of type A and/or AAAA,
• A and/or AAAA records must be of sufficient scope to be reachable by all hosts that might query the DNS. If a link-scope address or IPv4 autoconfiguration address is provided by the SRP client, the SRP server MUST treat this as if no address records were received; that is, the Host Description is not valid.
• exactly one "Add to an RRset" RR that adds a KEY RR that contains the public key corresponding to the private key that was used to sign the message,
• there is a Service Instance Name Instruction in the SRP Update for which the SRV RR that is added points to the hostname being updated by this update.
• Host Description Instructions do not modify any other resource records.

Valid SRP Update Requirements An SRP Update MUST include zero or more Service Discovery Instructions. For each Service Discovery Instruction, there MUST be at least one Service Description Instruction. Note that in the case of SRP subtypes (), it's quite possible that two Service Discovery Instructions might reference the same Service Description Instruction. For each Service Description Instruction there MUST be at least one Service Discovery Instruction with its service instance name as the target of its PTR record. There MUST be exactly one Host Description Instruction. Every Service Description Instruction must have that Host Description Instruction as the target of its SRV record. A DNS Update that does not meet these constraints is not an SRP Update. A DNS Update that contains any additional adds or deletes that cannot be identified as Service Discovery, Service Description or Host Description Instructions is not an SRP Update. A DNS update that contains any prerequisites is not an SRP Update. Such messages should either be processed as regular RFC2136 updates, including access control checks and constraint checks, if supported, or else rejected with RCODE=REFUSED. In addition, in order for an update to be a valid SRP Update, the target of every Service Discovery Instruction MUST be a Service Description Instruction that is present in the SRP Update. There MUST NOT be any Service Description Instruction to which no Service Discovery Instruction points. The target of the SRV record in every Service Description Instruction MUST be the single Host Description Instruction. If the definitions of each of these instructions are followed carefully and the update requirements are validated correctly, many DNS Updates that look very much like SRP Updates nevertheless will fail to validate. For example, a DNS update that contains an Add to an RRset instruction for a Service Name and an Add to an RRset instruction for a Service Instance Name, where the PTR record added to the Service Name does not reference the Service Instance Name, is not a valid SRP Update message, but may be a valid RFC2136 update.

FCFS Name And Signature Validation Assuming that a DNS Update message has been validated with these conditions and is a valid SRP Update, the server checks that the name in the Host Description Instruction exists. If so, then the server checks to see if the KEY record on that name is the same as the KEY record in the Host Description Instruction. The server performs the same check for the KEY records in any Service Description Instructions. For KEY records that were omitted from Service Description Instructions, the KEY from the Host Description Instruction is used. If any existing KEY record corresponding to a KEY record in the SRP Update does not match the KEY record in the SRP Update (whether provided or taken from the Host Description Instruction), then the server MUST reject the SRP Update with the YXDOMAIN RCODE. Otherwise, the server validates the SRP Update using SIG(0) against the public key in the KEY record of the Host Description Instruction. If the validation fails, the server MUST reject the SRP Update with the REFUSED RCODE. Otherwise, the SRP Update is considered valid and authentic, and is processed according to the method described in RFC2136. KEY record updates omitted from Service Description Instruction are processed as if they had been explicitly present: every Service Description that is updated MUST, after the SRP Update has been applied, have a KEY RR, and it must be the same KEY RR that is present in the Host Description to which the Service Description refers.

Handling of Service Subtypes SRP servers MUST treat the update instructions for a service type and all its subtypes as atomic. That is, when a service and its subtypes are being updated, whatever information appears in the SRP Update is the entirety of information about that service and its subtypes. If any subtype appeared in a previous update but does not appear in the current update, then the DNS server MUST remove that subtype. Similarly, there is no mechanism for deleting subtypes. A delete of a service deletes all of its subtypes. To delete an individual subtype, an SRP Update must be constructed that contains the service type and all subtypes for that service.

SRP Update response The status that is returned depends on the result of processing the update, and can be either SUCCESS or SERVFAIL: all other possible outcomes should already have been accounted for when applying the constraints that qualify the update as an SRP Update.

Optional Behavior The server MAY add a Reverse Mapping that corresponds to the Host Description. This is not required because the Reverse Mapping serves no protocol function, but it may be useful for debugging, e.g. in annotating network packet traces or logs. In order for the server to add a reverse mapping update, it must be authoritative for the zone or have credentials to do the update. The SRP client MAY also do a reverse mapping update if it has credentials to do so. The server MAY apply additional criteria when accepting updates. In some networks, it may be possible to do out-of-band registration of keys, and only accept updates from pre-registered keys. In this case, an update for a key that has not been registered should be rejected with the REFUSED RCODE. There are at least two benefits to doing this rather than simply using normal SIG(0) DNS updates. First, the same registration protocol can be used in both cases, so both use cases can be addressed by the same service implementation. Second, the registration protocol includes maintenance functionality not present with normal DNS updates. Note that the semantics of using SRP in this way are different than for typical RFC2136 implementations: the KEY used to sign the SRP Update only allows the SRP client to update records that refer to its Host Description. RFC2136 implementations do not normally provide a way to enforce a constraint of this type. The server may also have a dictionary of names or name patterns that are not permitted. If such a list is used, updates for Service Instance Names that match entries in the dictionary are rejected with YXDOMAIN.

TTL Consistency All RRs within an RRset are required to have the same TTL ( Clarifications to the DNS Specification). In order to avoid inconsistencies, SRP places restrictions on TTLs sent by services and requires that SRP servers enforce consistency. Services sending SRP Updates MUST use consistent TTLs in all RRs within the SRP Update. SRP servers MUST check that the TTLs for all RRs within the SRP Update are the same. If they are not, the SRP update MUST be rejected with a REFUSED RCODE. Additionally, when adding RRs to an RRset, for example when processing Service Discovery records, the server MUST use the same TTL on all RRs in the RRset. How this consistency is enforced is up to the implementation. TTLs sent in SRP Updates are advisory: they indicate the SRP client's guess as to what a good TTL would be. SRP servers may override these TTLs. SRP servers SHOULD ensure that TTLs are reasonable: neither too long nor too short. The TTL should never be longer than the lease time (). Shorter TTLs will result in more frequent data refreshes; this increases latency on the DNS-SD client side, increases load on any caching resolvers and on the authoritative server, and also increases network load, which may be an issue for constrained networks. Longer TTLs will increase the likelihood that data in caches will be stale. TTL minimums and maximums SHOULD be configurable by the operator of the SRP server.

Maintenance

Cleaning up stale data Because the DNS&nbhy;SD registration protocol is automatic, and not managed by humans, some additional bookkeeping is required. When an update is constructed by the SRP client, it MUST include an EDNS(0) Update Lease Option . The Update Lease Option contains two lease times: the Lease Time and the Key Lease Time. These leases are promises, similar to DHCP leases, from the SRP client that it will send a new update for the service registration before the lease time expires. The Lease time is chosen to represent the time after the update during which the registered records other than the KEY record should be assumed to be valid. The Key Lease time represents the time after the update during which the KEY record should be assumed to be valid. The reasoning behind the different lease times is discussed in the section on first-come, first-served naming (). SRP servers may be configured with limits for these values. A default limit of two hours for the Lease and 14 days for the SIG(0) KEY are currently thought to be good choices. Constrained devices with limited battery that wake infrequently are likely to request longer leases; servers that support such devices may need to set higher limits. SRP clients that are going to continue to use names on which they hold leases should update well before the lease ends, in case the registration service is unavailable or under heavy load. The lease time applies specifically to the host. All service instances, and all service entries for such service instances, depend on the host. When the lease on a host expires, the host and all services that reference it MUST be removed at the same time—it is never valid for a service instance to remain when the host it references has been removed. If the KEY record for the host is to remain, the KEY record for any services that reference it MUST also remain. However, the service PTR record MUST be removed, since it has no key associated with it, and since it is never valid to have a service PTR record for which there is no service instance on the target of the PTR record. SRP Servers SHOULD also track a lease time per service instance. The reason for doing this is that a client may re-register a host with a different set of services, and not remember that some different service instance had previously been registered. In this case, when that service instance lease expires, the SRP server SHOULD remove the service instance (although the KEY record for the service instance SHOULD be retained until the key lease on that service expires). This is beneficial because if the SRP client continues to renew the host, but never mentions the stale service again, the stale service will continue to be advertised. The SRP server MUST include an EDNS(0) Update Lease option in the response if the lease time proposed by the service has been shortened or lengthened. The service MUST check for the EDNS(0) Update Lease option in the response and MUST use the lease times from that option in place of the options that it sent to the server when deciding when to update its registration. The times may be shorter or longer than those specified in the SRP Update; the SRP client must honor them in either case. SRP clients should assume that each lease ends N seconds after the update was first transmitted, where N is the lease duration. Servers should assume that each lease ends N seconds after the update that was successfully processed was received. Because the server will always receive the update after the SRP client sent it, this avoids the possibility of misunderstandings. SRP servers MUST reject updates that do not include an EDNS(0) Update Lease option. Dual-use servers MAY accept updates that don't include leases, but SHOULD differentiate between SRP Updates and other updates, and MUST reject updates that would otherwise be SRP Updates if they do not include leases. Lease times have a completely different function than TTLs. On an authoritative DNS server, the TTL on a resource record is a constant: whenever that RR is served in a DNS response, the TTL value sent in the answer is the same. The lease time is never sent as a TTL; its sole purpose is to determine when the authoritative DNS server will delete stale records. It is not an error to send a DNS response with a TTL of 'n' when the remaining time on the lease is less than 'n'.

Security Considerations

Source Validation SRP Updates have no authorization semantics other than first-come, first-served. This means that if an attacker from outside of the administrative domain of the server knows the server's IP address, it can in principle send updates to the server that will be processed successfully. Servers should therefore be configured to reject updates from source addresses outside of the administrative domain of the server. For updates sent to an anycast IP address of an SRP server, this validation must be enforced by every router on the path from the Constrained-Device Network to the unconstrained portion of the network. For TCP updates, the initial SYN-SYN+ACK handshake prevents updates being forged by an off-network attacker. In order to ensure that this handshake happens, SRP servers relying on three-way-handshake validation MUST NOT accept TCP Fast Open payloads. If the network infrastructure allows it, an SRP server MAY accept TCP Fast Open payloads if all such packets are validated along the path, and the network is able to reject this type of spoofing at all ingress points. Note that these rules only apply to the validation of SRP Updates. A server that accepts updates from SRP clients may also accept other DNS updates, and those DNS updates may be validated using different rules. However, in the case of a DNS service that accepts SRP updates, the intersection of the SRP Update rules and whatever other update rules are present must be considered very carefully. For example, a normal, authenticated DNS update to any RR that was added using SRP, but that is authenticated using a different key, could be used to override a promise made by the SRP Server to an SRP client, by replacing all or part of the service registration information with information provided by an authenticated DNS update client. An implementation that allows both kinds of updates should not allow DNS Update clients that are using different authentication and authorization credentials to to update records added by SRP clients.

SRP Server Authentication This specification does not provide a mechanism for validating responses from DNS servers to SRP clients. In the case of Constrained Network/Constrained Node clients, such validation isn't practical because there's no way to establish trust. In principle, a KEY RR could be used by a non-constrained SRP client to validate responses from the server, but this is not required, nor do we specify a mechanism for determining which key to use.

Required Signature Algorithm For validation, SRP servers MUST implement the ECDSAP256SHA256 signature algorithm. SRP servers SHOULD implement the algorithms specified in , in the validation column of the table, that are numbered 13 or higher and have a "MUST", "RECOMMENDED", or "MAY" designation in the validation column of the table. SRP clients MUST NOT assume that any algorithm numbered lower than 13 is available for use in validating SIG(0) signatures.

Privacy Considerations Because DNSSD SRP Updates can be sent off-link, the privacy implications of SRP are different than for multicast DNS responses. Host implementations that are using TCP SHOULD also use TLS if available. Server implementations MUST offer TLS support. The use of TLS with DNS is described in and . Hosts that implement TLS support SHOULD NOT fall back to TCP; since servers are required to support TLS, it is entirely up to the host implementation whether to use it. Public keys can be used as identifiers to track hosts. SRP servers MAY elect not to return KEY records for queries for SRP registrations.

Delegation of 'service.arpa.' In order to be fully functional, the owner of the 'arpa.' zone must add a delegation of 'service.arpa.' in the '.arpa.' zone . This delegation should be set up as was done for 'home.arpa', as a result of the specification in .

IANA Considerations

Registration and Delegation of 'service.arpa' as a Special-Use Domain Name IANA is requested to record the domain name 'service.arpa.' in the Special-Use Domain Names registry . IANA is requested, with the approval of IAB, to implement the delegation requested in . IANA is further requested to add a new entry to the "Transport-Independent Locally-Served Zones" subregistry of the the "Locally-Served DNS Zones" registry . The entry will be for the domain 'service.arpa.' with the description "DNS&nbhy;SD Registration Protocol Special-Use Domain", listing this document as the reference.

'dnssd-srp' Service Name IANA is also requested to add a new entry to the Service Names and Port Numbers registry for dnssd-srp with a transport type of tcp. No port number is to be assigned. The reference should be to this document, and the Assignee and Contact information should reference the authors of this document. The Description should be as follows: Availability of DNS Service Discovery Service Registration Protocol Service for a given domain is advertised using the "_dnssd&nbhy;srp._tcp.<domain>" SRV record gives the target host and port where DNSSD Service Registration Service is provided for the named domain.

'dnssd-srp-tls' Service Name IANA is also requested to add a new entry to the Service Names and Port Numbers registry for dnssd-srp with a transport type of tcp. No port number is to be assigned. The reference should be to this document, and the Assignee and Contact information should reference the authors of this document. The Description should be as follows: Availability of DNS Service Discovery Service Registration Protocol Service for a given domain over TLS is advertised using the "_dnssd&nbhy;srp&nbhy;tls._tcp.<domain>." SRV record gives the target host and port where DNSSD Service Registration Service is provided for the named domain.

Anycast Address IANA is requested to allocate an IPv6 Anycast address from the IPv6 Special-Purpose Address Registry, similar to the Port Control Protocol anycast address, 2001:1::1. The value TBD should be replaced with the actual allocation in the table that follows. The values for the registry are:

Attribute	value
Address Block	2001:1::TBD/128
Name	DNS-SD Service Registration Protocol Anycast Address
RFC	[this document]
Allocation Date	[date of allocation]
Termination Date	N/A
Source	True
Destination	True
Forwardable	True
Global	True
Reserved-by-protocol	False

Implementation Status [Note to the RFC Editor: please remove this section prior to publication.] This section records the status of known implementations of the protocol defined by this specification at the time of posting of this Internet-Draft, and is based on a proposal described in RFC 7942. The description of implementations in this section is intended to assist the IETF in its decision processes in progressing drafts to RFCs. 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. According to RFC 7942, "this will allow reviewers and working groups to assign due consideration to documents that have the benefit of running code, which may serve as evidence of valuable experimentation and feedback that have made the implemented protocols more mature. It is up to the individual working groups to use this information as they see fit". There are two known independent implementations of SRP clients:

• SRP Client for OpenThread: https://github.com/openthread/openthread/pull/6038
• mDNSResponder open source project: https://github.com/Abhayakara/mdnsresponder

There are two related implementations of an SRP server. One acts as a DNS Update proxy, taking an SRP Update and applying it to the specified DNS zone using DNS update. The other acts as an Advertising Proxy . Both are included in the mDNSResponder open source project mentioned above.

Acknowledgments Thanks to , Jonathan Hui, Esko Dijk, Kangping Dong and Abtin Keshavarzian for their thorough technical reviews. Thanks to Kangping and Abtin as well for testing the document by doing an independent implementation. Thanks to Tamara Kemper for doing a nice developmental edit, Tim Wattenberg for doing a SRP client proof-of-concept implementation at the Montreal Hackathon at IETF 102, and Tom Pusateri for reviewing during the hackathon and afterwards.

Normative References Informative References

Testing using standard RFC2136-compliant servers It may be useful to set up a DNS server for testing that does not implement SRP. This can be done by configuring the server to listen on the anycast address, or advertising it in the _dnssd&nbhy;srp._tcp.<zone> SRV and _dnssd&nbhy;srp&nbhy;tls._tcp.<zone> record. It must be configured to be authoritative for "default.service.arpa", and to accept updates from hosts on local networks for names under "default.service.arpa" without authentication, since such servers will not have support for FCFS authentication (). A server configured in this way will be able to successfully accept and process SRP Updates from services that send SRP updates. However, no prerequisites will be applied, and this means that the test server will accept internally inconsistent SRP Updates, and will not stop two SRP Updates, sent by different services, that claim the same name(s), from overwriting each other. Since SRP Updates are signed with keys, validation of the SIG(0) algorithm used by the client can be done by manually installing the client public key on the DNS server that will be receiving the updates. The key can then be used to authenticate the client, and can be used as a requirement for the update. An example configuration for testing SRP using BIND 9 is given in .

How to allow services to update standard RFC2136-compliant servers Ordinarily SRP Updates will fail when sent to an RFC 2136-compliant server that does not implement SRP because the zone being updated is "default.service.arpa", and no DNS server that is not an SRP server should normally be configured to be authoritative for "default.service.arpa". Therefore, a service that sends an SRP Update can tell that the receiving server does not support SRP, but does support RFC2136, because the RCODE will either be NOTZONE, NOTAUTH or REFUSED, or because there is no response to the update request (when using the anycast address) In this case a service MAY attempt to register itself using regular RFC2136 DNS updates. To do so, it must discover the default registration zone and the DNS server designated to receive updates for that zone, as described earlier, using the _dns&nbhy;update._udp SRV record. It can then make the update using the port and host pointed to by the SRV record, and should use appropriate prerequisites to avoid overwriting competing records. Such updates are out of scope for SRP, and a service that implements SRP MUST first attempt to use SRP to register itself, and should only attempt to use RFC2136 backwards compatibility if that fails. Although the owner name for the SRV record specifies the UDP protocol for updates, it is also possible to use TCP, and TCP should be required to prevent spoofing.

Sample BIND9 configuration for default.service.arpa.