Internet-Draft | Service Identity | October 2021 |
Saint-Andre, et al. | Expires 16 April 2022 | [Page] |
Many application technologies enable secure communication between two entities by means of Transport Layer Security (TLS) with Internet Public Key Infrastructure Using X.509 (PKIX) certificates. This document specifies procedures for representing and verifying the identity of application services in such interactions.¶
This document obsoletes RFC 6125.¶
This note is to be removed before publishing as an RFC.¶
Discussion of this document takes place on the Using TLS in Applications Working Group mailing list (uta@ietf.org), which is archived at https://mailarchive.ietf.org/arch/browse/uta/.¶
Source for this draft and an issue tracker can be found at https://github.com/richsalz/draft-ietf-uta-rfc6125bis.¶
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 16 April 2022.¶
Copyright (c) 2021 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 Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.¶
The visible face of the Internet largely consists of services that employ a client-server architecture in which an interactive or automated client communicates with an application service. When a client communicates with an application service using Transport Layer Security [TLS] or Datagram Transport Layer Security [DTLS], it has some notion of the server's identity (e.g., "the website at example.com") while attempting to establish secure communication. Likewise, during TLS negotiation, the server presents its notion of the service's identity in the form of a public-key certificate that was issued by a certification authority (CA) in the context of the Internet Public Key Infrastructure using X.509 [PKIX]. Informally, we can think of these identities as the client's "reference identity" and the server's "presented identity" (more formal definitions are given later). A client needs to verify that the server's presented identity matches its reference identity so it can authenticate the communication.¶
This document defines procedures for how clients do this verification. It therefore implicitly defines requirements on other parties, such as the CA's that issue certificates, the service administrators requesting them, and the protocol designers defining how things are named.¶
This document revises and obsoletes [VERIFY] based on the decade of experience and changes since it was first published. The major changes, in no particular order, include:¶
CN-ID
,
to represent the server identity; only the subjectAltNames extension
is used.¶
This document does not supersede the rules for certificate issuance or validation specified by [PKIX]. That document also governs any certificate-related topic on which this document is silent. This includes certificate syntax, certificate extensions such as name constraints or extended key usage, and handling of certification paths.¶
This document addresses only name forms in the leaf "end entity" server certificate. It does not address the name forms in the chain of certificates used to validate a cetrificate, let alone creating or checking the validity of such a chain. In order to ensure proper authentication, applications need to verify the entire certification path as per [PKIX].¶
The previous version of this specification, [VERIFY], surveyed the current practice from many IETF standards and tried to generalize best practices. This document takes the lessons learned in the past decade and codifies them. he rules are brief:¶
This document applies only to service identities associated with FQDNs only to TLS and DTLS, and only to PKIX-based systems.¶
TLS uses the words client and server, where the client is the entity that initiates the connection. In many cases, this models common practice, such as a browser connecting to a Web origin. For the sake of clarity, and to follow the usage in [TLS] and related specifications, we will continue to use to use the terms client and server in this document. Note that these are TLS-layer roles, and that the application protocol could support the TLS server making requests to the TLS client after the TLS handshake; these is no requirement that the roles at the application layer match the TLS-layer.¶
At the time of this writing, other protocols such as [QUIC] and Network Time Security ([NTS]) use TLS as a service to do the initial establishment of cryptographic key material. Such services MUST also follow the rules specified here.¶
The following topics are out of scope for this specification:¶
Certification authority policies. This includes items such as the following:¶
Because many concepts related to "identity" are often too vague to be actionable in application protocols, we define a set of more concrete terms for use in this specification.¶
A service on the Internet that enables interactive and automated clients to connect for the purpose of retrieving or uploading information, communicating with other entities, or connecting to a broader network of services.¶
An organization or individual that hosts or deploys an application service.¶
A formal identifier for the application protocol used to provide a particular kind of application service at a domain. This often apepars as a URI scheme [URI] or a DNS SRV Service [DNS-SRV].¶
A software agent or device that is not directly controlled by a human user.¶
A domain name or host name that is explicitly configured for communicating with the source domain, by either the human user controlling an interactive client or a trusted administrator. For example, a server at mailhost.example.com for connecting to an IMAP server hosting an email address of user@example.com.¶
A domain name or host name that a client has derived from the source domain in an automated fashion (e.g., by means of a [DNS-SRV] lookup).¶
A particular instance of an identifier type that is either presented by a server in a certificate or referenced by a client for matching purposes.¶
A formally defined category of identifier that can be included in a certificate and therefore that can also be used for matching purposes. For conciseness and convenience, we define the following identifier types of interest, which are based on those found in the PKIX specification [PKIX] and various PKIX extensions:¶
A software agent or device that is directly controlled by a human user, commonly known as a "user agent."¶
PKIX is a short name for the Internet Public Key Infrastructure using X.509 defined in [PKIX]. That document provides a profile of the X.509v3 certificate specifications and X.509v2 certificate revocation list (CRL) specifications for use in the Internet.¶
An identifier presented by a server to a client within a PKIX certificate when the client attempts to establish secure communication with the server. The certificate can include one or more presented identifiers of different types, and if the server hosts more than one domain then the certificate might present distinct identifiers for each domain.¶
An identifier used by the client when examining presented identifiers. It is constructed from the source domain, and optionally an application service type.¶
The ASN.1-based construction comprising a Relative Distinguished Name (RDN), which itself is a building-block component of Distinguished Names. See [LDAP-DN], Section 2.¶
The FQDN that a client expects an application service to present in the certificate. This is typically input by a human user, configured into a client, or provided by reference such as URL. The combination of a source domain and, optionally, an application service type enables a client to construct one or more reference identifiers.¶
An identifier placed in a subjectAltName extension.¶
A standard PKIX certificate extension [PKIX] enabling identifiers of various types to be bound to the certificate subject.¶
In this specification, the term refers to the name of a PKIX certificate's subject, encoded in a certificate's subject field (see [PKIX], Section 4.1.2.6).¶
Security-related terms used in this document, but not defined here or in [PKIX] should be understood in the the sense defined in [SECTERMS]. Such terms include "attack", "authentication", "identity", "trust", "validate", and "verify".¶
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.¶
This document assumes that the name of an application service is
based on a DNS domain name (e.g., example.com
) -- supplemented in
some circumstances by an application service type (e.g., "the IMAP
server at example.com").
The DNS name conforms to one of the following forms:¶
From the perspective of the application client or user, some names are direct because they are provided directly by a human user. This includes runtime input, prior configuration, or explicit acceptance of a client communication attempt. Other names are indirect because they are automatically resolved by the application based on user input, such as a target name resolved from a source name using DNS SRV or [NAPTR] records). The distinction matters most for certificate consumption, specifically verification as discussed in this document.¶
From the perspective of the application service, some names are unrestricted because they can be used in any type of service, such as a single certificate being used for both the HTTP and IMAP services at the host example.com. Other names are restricted because they can only be used for one type of service, such as a special-purpose certificate that can only be used for an IMAP service. This distinction matters most for certificate issuance.¶
We can categorize the supported identifier types as follows:¶
It is important to keep these distinctions in mind, as best practices for the deployment and use of the identifiers differ. As a further example, cross-protocol attacks such as [ALPACA] are possibile when two different protocol services use the same certificate. This can be addressed by using restricted identifiers, or telling services to not share certificates. Protocol specifications MUST specify which identifiers are mandatory-to-implement and SHOULD provide operational guidance when necessary.¶
The Common Name RDN MUST NOT be used to identify a service. Reasons for this include:¶
For similar reasons, other RDN's within the Subject Name MUST NOT be used to identify a service.¶
This section defines how protocol designers should reference this document, which MUST be a normative reference in their specification. The technology MUST only use the identifiers defined in this document. Its specification MAY choose to allow only one of them.¶
If the technology does not use DNS SRV records to resolve the DNS domain names of application services then its specification MUST state that SRV-ID as defined in this document is not supported. Note that many existing application technologies use DNS SRV records to resolve the DNS domain names of application services, but do not rely on representations of those records in PKIX certificates by means of SRV-IDs as defined in [SRVNAME].¶
If the technology does not use URI's to identify application services, then its specification MUST state that URI-ID as defined in this document is not supported. Note that many existing application technologies use URIs to identify application services, but do not rely on representation of those URIs in PKIX certificates by means of URI-IDs.¶
A technology MAY disallow the use of the wildcard character in DNS names. If it does so, then the specification MUST state that wildcard certificates as defined in this document are not supported.¶
This section provides instructions for issuers of certificates.¶
When a certification authority issues a certificate based on the FQDN at which the application service provider will provide the relevant application, the following rules apply to the representation of application service identities. Note that some of these rules are cumulative and can interact in important ways that are illustrated later in this document.¶
sip
but not sips
or tel
for SIP as described in [SIP-SIPS]).¶
Consider a simple website at www.example.com
, which is not discoverable via
DNS SRV lookups. Because HTTP does not specify the use of URIs in server
certificates, a certificate for this service might include only a DNS-ID of
www.example.com
.¶
Consider an IMAP-accessible email server at the host mail.example.net
servicing email addresses of the form user@example.net
and discoverable via
DNS SRV lookups on the application service name of example.net
. A
certificate for this service might include SRV-IDs of _imap.example.net
and
_imaps.example.net
(see [EMAIL-SRV]) along with DNS-IDs of example.net
and mail.example.net
.¶
Consider a SIP-accessible voice-over-IP (VoIP) server at the host
voice.example.edu
servicing SIP addresses of the form
user@voice.example.edu
and identified by a URI of <sip:voice.example.edu>.
A certificate for this service would include a URI-ID of
sip:voice.example.edu
(see [SIP-CERTS]) along with a DNS-ID of
voice.example.edu
.¶
Consider an XMPP-compatible instant messaging (IM) server at the host
im.example.org
servicing IM addresses of the form user@im.example.org
and
discoverable via DNS SRV lookups on the im.example.org
domain. A
certificate for this service might include SRV-IDs of
_xmpp-client.im.example.org
and _xmpp-server.im.example.org
(see
[XMPP]), a DNS-ID of im.example.org
. For backward compatibility, it may
also have an XMPP-specific XmppAddr
of im.example.org
(see [XMPP]).¶
This section provides instructions for service providers regarding the information to include in certificate signing requests (CSRs). In general, service providers SHOULD request certificates that include all of the identifier types that are required or recommended for the application service type that will be secured using the certificate to be issued.¶
If the certificate will be used for only a single type of application service, the service provider SHOULD request a certificate that includes a DNS-ID and, if appropriate for the application service type, an SRV-ID or URI-ID that limits the deployment scope of the certificate to only the defined application service type.¶
If the certificate might be used for any type of application service, then the service provider SHOULD to request a certificate that includes only a DNS-ID. Again, because of multi-protocol attacks this practice is discouraged; this can be mitigated by providing only one service on a host.¶
If a service provider offersmultiple application service types and wishes to
limit the applicability of certificates using SRV-IDs or URI-IDs, they SHOULD
request multiple certificates, rather than a single certificate containing
multiple SRV-IDs or URI-IDs each identifying ia different application service
type. This rule does not apply to application service type "bundles" that
identify distinct access methods to the same underlying application such as
an email application with access methods denoted by the application service
types of imap
, imaps
, pop3
, pop3s
, and submission
as described in
[EMAIL-SRV].¶
At a high level, the client verifies the application service's identity by performing the following actions:¶
Naturally, in addition to checking identifiers, a client should perform further checks, such as expiration and revocation, to ensure that the server is authorized to provide the requested service. Such checking is not a matter of verifying the application service identity presented in a certificate, however, and methods for doing so are therefore out of scope for this document.¶
The client MUST construct a list of acceptable reference identifiers, and MUST do so independently of the identifiers presented by the service.¶
The inputs used by the client to construct its list of reference identifiers might be a URI that a user has typed into an interface (e.g., an HTTPS URL for a website), configured account information (e.g., the domain name of a host for retrieving email, which might be different from the DNS domain name portion of a username), a hyperlink in a web page that triggers a browser to retrieve a media object or script, or some other combination of information that can yield a source domain and an application service type.¶
The client might need to extract the source domain and application service type from the input(s) it has received. The extracted data MUST include only information that can be securely parsed out of the inputs, such as parsing the FQDN out of the "host" component or deriving the application service type from the scheme of a URI. Other possibilities include pulling the data from a delegated domain that is explicitly established via client or system configuration, resolving the data via [DNSSEC], or obtaining the data from a third-party domain mapping service in which a human user has explicitly placed trust and with which the client communicates over a connection or association that provides both mutual authentication and integrity checking. These considerations apply only to extraction of the source domain from the inputs. Naturally, if the inputs themselves are invalid or corrupt (e.g., a user has clicked a link provided by a malicious entity in a phishing attack), then the client might end up communicating with an unexpected application service.¶
For example, given an input URI of <sips:alice@example.net>, a client
would derive the application service type sip
from the scheme
and parse the domain name example.net
from the host component.¶
Each reference identifier in the list MUST be based on the source domain and MUST NOT be based on a derived domain such as a domain name discovered through DNS resolution of the source domain. This rule is important because only a match between the user inputs and a presented identifier enables the client to be sure that the certificate can legitimately be used to secure the client's communication with the server.¶
Using the combination of FQDN(s) and application service type, the client MUST construct its list of reference identifiers in accordance with the following rules:¶
Which identifier types a client includes in its list of reference identifiers, and their priority, is a matter of local policy. For example, a client that is built to connect only to a particular kind of service might be configured to accept as valid only certificates that include an SRV-ID for that application service type. By contrast, a more lenient client, even if built to connect only to a particular kind of service, might include both SRV-IDs and DNS-IDs in its list of reference identifiers.¶
A web browser that is connecting via HTTPS to the website at www.example.com
would have a single reference identifier: a DNS-ID of www.example.com
.¶
A mail user agent that is connecting via IMAPS to the email service at
example.net
(resolved as mail.example.net
) might have three reference
identifiers: an SRV-ID of _imaps.example.net
(see [EMAIL-SRV]), and
DNS-IDs of example.net
and mail.example.net
. An email user agentthat
does not support [EMAIL-SRV] would probably be explicitly configured to
connect to mail.example.net
, whereas an SRV-aware user agent would derive
example.net
from an email address of the form user@example.net
but might
also accept mail.example.net
as the DNS domain name portion of reference
identifiers for the service.¶
A voice-over-IP (VoIP) user agent that is connecting via SIP to the voice
service at voice.example.edu
might have only one reference identifier:
a URI-ID of sip:voice.example.edu
(see [SIP-CERTS]).¶
An instant messaging (IM) client that is connecting via XMPP to the IM
service at im.example.org
might have three reference identifiers: an
SRV-ID of _xmpp-client.im.example.org
(see [XMPP]), a DNS-ID of
im.example.org
, and an XMPP-specific XmppAddr
of im.example.org
(see [XMPP]).¶
Once the client has constructed its list of reference identifiers and has received the server's presented identifiers in the form of a PKIX certificate, the client checks its reference identifiers against the presented identifiers for the purpose of finding a match. The search fails if the client exhausts its list of reference identifiers without finding a match. The search succeeds if any presented identifier matches one of the reference identifiers, at which point the client SHOULD stop the search.¶
Before applying the comparison rules provided in the following sections, the client might need to split the reference identifier into its DNS domain name portion and its application service type portion, as follows:¶
_imaps.example.net
has a DNS domain name portion
of example.net
and an application service type portion of
imaps
, which maps to the IMAP application protocol as explained in
[EMAIL-SRV].¶
sip:voice.example.edu
would be split
into a DNS domain name portion of voice.example.edu
and an application
service type of sip
(associated with an application protocol of SIP as
explained in [SIP-CERTS]).¶
A client MUST match the DNS name, and if an application service type is present it MUST also match the service type as well. These are described below.¶
This section describes how the client must determine if the the presented DNS name matches the reference DNS name. The rules differ depending on whether the domain to be checked is a "traditional domain name" or an "internationalized domain name" (as defined under Section 2). For clients that support names containing the wildcard character "*", this section also specifies a supplemental rule for such "wildcard certificates". This section uses the description of labels and domain names in [DNS-CONCEPTS].¶
If the DNS domain name portion of a reference identifier is a "traditional
domain name", then matching of the reference identifier against the presented
identifier MUST be performed by comparing the set of domain name labels using
a case-insensitive ASCII comparison, as clarified by [DNS-CASE]. For
example, WWW.Example.Com
would be lower-cased to www.example.com
for
comparison purposes. Each label MUST match in order for the names to be
considered to match, except as supplemented by the rule about checking of
wildcard labels given below.¶
If the DNS domain name portion of a reference identifier is an internationalized domain name, then the client MUST convert any U-labels [IDNA-DEFS] in the domain name to A-labels before checking the domain name. In accordance with [IDNA-PROTO], A-labels MUST be compared as case-insensitive ASCII. Each label MUST match in order for the domain names to be considered to match, except as supplemented by the rule about checking of wildcard labels given below.¶
If the technology specification supports wildcards, then the client MUST match the reference identifier against a presented identifier whose DNS domain name portion contains the wildcard character "*" in a label provided these requirements are met:¶
If the requirements are not met, the presented identifier is invalid and MUST be ignored.¶
A wildcard in a presented identifier can only match exactly one label in a reference identifier. Note that this is not the same as DNS wildcard matching, where the "*" label always matches at least one whole label and sometimes more. See [DNS-CONCEPTS], Section 4.3.3 and [DNS-WILDCARDS].¶
For information regarding the security characteristics of wildcard certificates, see Section 7.1.¶
The rules for matching the application service type deopend on whether the identifier is an SRV-ID or a URI-ID.¶
These identifiers provide an application service type portion to be checked,
but that portion is combined only with the DNS domain name portion of the
SRV-ID or URI-ID itself. For example, if a client's list of reference
identifiers includes an SRV-ID of _xmpp-client.im.example.org
and a DNS-ID
of apps.example.net
, the client would check both the combination of an
application service type of xmpp-client
and a DNS domain name of
im.example.org
and a DNS domain name of apps.example.net
. However, the
client would not check the combination of an application service type of
xmpp-client
and a DNS domain name of apps.example.net
because it does not
have an SRV-ID of _xmpp-client.apps.example.net
in its list of reference
identifiers.¶
If the identifier is an SRV-ID, then the application service name MUST
be matched in a case-insensitive manner, in accordance with [DNS-SRV].
Note that the _
character is prepended to the service identifier in
DNS SRV records and in SRV-IDs (per [SRVNAME]), and thus does not
need to be included in any comparison.¶
If the identifier is a URI-ID, then the scheme name portion MUST be
matched in a case-insensitive manner, in accordance with [URI].
Note that the :
character is a separator between the scheme name
and the rest of the URI, and thus does not need to be included in any
comparison.¶
If the client has found a presented identifier that matches a reference identifier, then the service identity check has succeeded. In this case, the client MUST use the matched reference identifier as the validated identity of the application service.¶
If the client does not find a presented identifier matching any of the reference identifiers then the client MUST proceed as described as follows.¶
If the client is an automated application not directly controlled by a human user, then it SHOULD terminate the communication attempt with a bad certificate error and log the error appropriately. The application MAY provide a configuration setting to disable this behavior, but it MUST enable it by default.¶
If the client is an interactive client that is directly controlled by a human user, then it SHOULD inform the user of the identity mismatch and automatically terminate the communication attempt with a bad certificate error in order to prevent users from inadvertently bypassing security protections in hostile situations.¶
Some interactive clients MAY give advanced users the option of proceeding with acceptance despite the identity mismatch. Although this behavior can be appropriate in certain specialized circumstances, it needs to be handled with extreme caution, for example by first encouraging even an advanced user to terminate the communication attempt and, if they choose to proceed anyway, by forcing the user to view the entire certification path before proceeding.¶
The application MAY also present the user with the ability to accept the presented certificate as valid for subsequent connections. Such ad-hoc "pinning" SHOULD NOT restrict future connections to just the pinned certificate. Local policy that statically enforces a given certificate for a given peer is best made available only as prior configuration, rather than a just-in-time override for a failed connection.¶
Wildcard certificates, those that have an identifier with "*" as the left-most DNS label, automatically vouch for any single-label host names within their domain, but not multiple levels of domains. This can be convenient for administrators but also poses the risk of vouching for rogue or buggy hosts. See for example [Defeating-SSL] (beginning at slide 91) and [HTTPSbytes] (slides 38-40).¶
Protection against a wildcard that identifies a public suffix
[Public-Suffix], such as *.co.uk
or *.com
, is beyond the scope of this
document.¶
Allowing internationalized domain names can lead to visually similar characters, also referred to as "confusables", being included within certificates. For discussion, see for example [IDNA-DEFS], Section 4.4 and [UTS-39].¶
A given application service might be addressed by multiple DNS domain names for a variety of reasons, and a given deployment might service multiple domains or protocols. TLS Extensions such as TLS Server Name Identification (SNI), discussed in [TLS], Section 4.4.2.2, and Application Layer Protocol Negotiation (ALPN), discussed in [ALPN], provide a way for the application to indicate the desired identifier and protocol to the server, which can be used to select the most appropriate certificate.¶
To accommodate the workaround that was needed before the development of the SNI extension, this specification allows multiple DNS-IDs, SRV-IDs, or URI-IDs in a certificate.¶
This document has no actions for IANA.¶
We gratefully acknowledge everyone who contributed to the previous
version of this document, [VERIFY].
Thanks also to Carsten Bormann for converting the previous document
to Markdown so that we could more easily use Martin Thomson's i-d-template
software.¶