Proof-of-Possession Key Semantics for JSON Web Tokens (JWTs)Microsoftmbj@microsoft.comhttp://self-issued.info/Ping Identityve7jtb@ve7jtb.comhttp://www.thread-safe.com/ARM LimitedAustriaHannes.Tschofenig@gmx.nethttp://www.tschofenig.priv.at
Security
OAuth Working GroupRFCRequest for CommentsI-DInternet-DraftJSON Web TokenJWTProof-of-PossessionHolder-of-Key
This specification defines how to declare in a JSON Web Token (JWT)
that the presenter of the JWT possesses a particular proof-of-possession key
and that the recipient can cryptographically confirm
proof-of-possession of the key by the presenter.
Being able to prove possession of a key is also sometimes described as
the presenter being a holder-of-key.
This specification defines how a JSON Web Token can declare
that the presenter of the JWT possesses a particular proof-of-possession (PoP) key
and that the recipient can cryptographically confirm
proof-of-possession of the key by the presenter.
Proof-of-possession of a key is also sometimes described as
the presenter being a holder-of-key.
The specification
describes key confirmation, among other confirmation mechanisms.
This specification defines how to communicate key confirmation key information
in JWTs.
Envision the following two use cases.
The first use case employs a symmetric proof-of-possession key and
the second use case employs an asymmetric proof-of-possession key.
In the case illustrated in ,
either the presenter generates a symmetric key and
privately sends it to the issuer (1)
or the issuer generates a symmetric key and
privately sends it to the presenter (1).
The issuer generates a JWT with an encrypted copy of this
symmetric key in the confirmation claim.
This symmetric key is encrypted with a key known
only to the issuer and the recipient,
which was previously established in step (0).
The entire JWT is integrity protected by the issuer.
The JWT is then (2) sent to the presenter.
Now, the presenter is in possession of the symmetric key as well as the
JWT (which includes the confirmation claim).
When the presenter (3) presents the JWT to the recipient,
it also needs to demonstrate possession of the symmetric key;
the presenter, for example, (4) uses the symmetric key
in a challenge/response protocol with the recipient.
The recipient is then able to verify that it is
interacting with the genuine presenter by decrypting the key
in the confirmation claim of the JWT.
By doing this, the recipient obtains the symmetric key, which it then uses to
verify cryptographically protected messages exchanged with the presenter (4).
This symmetric key mechanism described above is conceptually similar
to the use of Kerberos tickets.
Note that for simplicity, the diagram above and associated text describe
the direct use of symmetric keys without the use of derived keys.
A more secure practice is to derive the symmetric keys actually used
from secrets exchanged, such as the key exchanged in step (0),
using a Key Derivation Function (KDF) and use the derived keys,
rather than directly using the secrets exchanged.
In the case illustrated in ,
the presenter generates a public/private key pair
and (1) sends the public key to the issuer,
which creates a JWT that contains the public key (or an identifier for it)
in the confirmation claim.
The entire JWT is integrity protected using a digital
signature to protect it against modifications.
The JWT is then (2) sent to the presenter.
When the presenter (3) presents the JWT to the recipient,
it also needs to demonstrate possession of the private key.
The presenter, for example, (4) uses the private key in a TLS exchange
with the recipient or (4) signs a nonce with the private key.
The recipient is able to verify that it is interacting with
the genuine presenter by extracting the public key from
the confirmation claim of the JWT
(after verifying the digital signature of the JWT)
and utilizing it with the private key in the TLS exchange
or by checking the nonce signature.
In both cases, the JWT may contain other claims that are needed by the application.
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 RFC 2119 .
Unless otherwise noted, all the protocol parameter names and values are case sensitive.
This specification uses terms defined in
the JSON Web Token ,
JSON Web Key , and
JSON Web Encryption specifications.
These terms are defined by this specification:
Party that creates the JWT and binds the proof-of-possession key to it.
Party that proves possession of a private key (for asymmetric key cryptography)
or secret key (for symmetric key cryptography) to a recipient.
Party that receives the JWT containing the proof-of-possession key information from the presenter.
By including a cnf (confirmation) claim in a JWT,
the issuer of the JWT declares that the presenter possesses a particular key,
and that the recipient can cryptographically confirm that
the presenter has possession of that key.
The value of the cnf claim is a JSON object
and the members of that object identify the proof-of-possession key.
The presenter can be identified in one of several ways by the JWT,
depending upon the application requirements.
If the JWT contains a sub (subject) claim ,
the presenter is normally the subject identified by the JWT.
(In some applications, the subject identifier will be relative to
the issuer identified by the iss (issuer) claim .)
If the JWT contains no sub (subject) claim,
the presenter is normally the issuer identified by the JWT
using the iss (issuer) claim.
The case in which the presenter is the subject of the JWT is analogous to
SAML 2.0 SubjectConfirmation usage.
At least one of the sub and iss
claims MUST be present in the JWT.
Some use cases may require that both be present.
Another means used by some applications to identify the presenter
is an explicit claim,
such as the azp (authorized party) claim
defined by OpenID Connect .
Ultimately, the means of identifying the presenter is application-specific,
as is the means of confirming possession of the key that is communicated.
The cnf (confirmation) claim is used in the JWT
to contain members used to identify the proof-of-possession key.
Other members of the cnf object may be defined because
a proof-of-possession key may not be the only means of confirming
the authenticity of the token.
This is analogous to the SAML 2.0
SubjectConfirmation element, in which a number of different
subject confirmation methods can be included,
including proof-of-possession key information.
The set of confirmation members that a
JWT must contain to be considered valid is context dependent
and is outside the scope of this specification.
Specific applications of JWTs will require implementations
to understand and process some confirmation members in particular ways.
However, in the absence of such requirements, all confirmation members
that are not understood by implementations MUST be ignored.
This specification establishes the
IANA "JWT Confirmation Methods" registry for these members
in and registers the members defined by this specification.
Other specifications can register other members used for confirmation,
including other members for conveying proof-of-possession keys,
possibly using different key representations.
The cnf claim value MUST represent only a single
proof-of-possession key; thus, at most one of the
jwk,
jwe,
and jku
confirmation values defined below may be present.
Note that if an application needs to represent multiple proof-of-possession
keys in the same JWT, one way for it to achieve this is to
use other claim names, in addition to cnf,
to hold the additional proof-of-possession key information.
These claims could use the same syntax and semantics
as the cnf claim.
Those claims would be defined by applications or other specifications
and could be registered in
the IANA "JSON Web Token Claims" registry .
When the key held by the presenter is an asymmetric private key,
the jwk member
is a JSON Web Key
representing the corresponding asymmetric public key.
The following example demonstrates such a declaration
in the JWT Claims Set of a JWT:
The JWK MUST contain the required key members for a JWK of that key type
and MAY contain other JWK members,
including the kid (key ID) member.
The jwk member MAY also be used for a JWK
representing a symmetric key, provided that the JWT is encrypted
so that the key is not revealed to unintended parties.
If the JWT is not encrypted, the symmetric key MUST be encrypted as described below.
When the key held by the presenter is a symmetric key,
the jwe member
is an encrypted JSON Web Key
encrypted to a key known to the recipient
using the JWE Compact Serialization
containing the symmetric key.
The rules for encrypting a JWK are found in
Section 7 of the JSON Web Key specification.
The following example illustrates a symmetric key that could subsequently be
encrypted for use in the jwe member:
The UTF-8 encoding of this JWK
is used as the JWE Plaintext when encrypting the key.
The following example is a JWE Header that could be used
when encrypting this key:
The following example JWT Claims Set of a JWT illustrates
the use of an encrypted symmetric key as the
jwe member value:
The proof-of-possession key can also be identified by the use of
a Key ID instead of communicating the actual key,
provided the recipient is able to obtain the identified key
using the Key ID.
In this case,
the issuer of a JWT declares that the presenter possesses a particular key
and that the recipient can cryptographically confirm
proof-of-possession of the key by the presenter by including a
cnf (confirmation) claim in the JWT
whose value is a JSON object, with the JSON object containing a
kid (key ID) member
identifying the key.
The following example demonstrates such a declaration
in the JWT Claims Set of a JWT:
The content of the kid value is application specific.
For instance, some applications may choose to use a JWK Thumbprint
value as the kid value.
The proof-of-possession key can be passed by reference
instead of being passed by value.
This is done using the jku (JWK Set URL) member.
Its value is a URI that refers to a
resource for a set of JSON-encoded public keys represented as a JWK Set ,
one of which is the proof-of-possession key.
If there are multiple keys in the referenced JWK Set document,
a kid member MUST also be included,
with the referenced key's JWK also containing the same kid value.
The protocol used to acquire the resource MUST provide
integrity protection. An HTTP GET request to retrieve the
JWK Set MUST use Transport Layer Security (TLS)
and
the identity of the server MUST be validated, as per
Section 6 of RFC 6125.
The following example demonstrates such a declaration
in the JWT Claims Set of a JWT:
Proof-of-possession is typically demonstrated by having the presenter sign
a value determined by the recipient using the key possessed by the presenter.
This value is sometimes called a "nonce" or a "challenge".
The means of communicating the nonce and the nature of its contents
are intentionally not described in this specification,
as different protocols will communicate this information in different ways.
Likewise, the means of communicating the signed nonce is also not specified,
as this is also protocol-specific.
Note that another means of proving possession of the key
when it is a symmetric key is to encrypt the key to the recipient.
The means of obtaining a key for the recipient is likewise protocol-specific.
For examples using the mechanisms defined in this specification,
see .
All of the security considerations that
are discussed in also apply here.
In addition, proof-of-possession introduces its own unique security issues.
Possessing a key is only valuable if it is kept secret.
Appropriate means must be used to ensure that unintended parties
do not learn private key or symmetric key values.
Applications utilizing proof-of-possession should also utilize audience restriction,
as described in Section 4.1.3 of ,
as it provides different protections.
Proof-of-possession can be used by recipients to reject messages from unauthorized senders.
Audience restriction can be used by recipients to reject messages intended for different recipients.
A recipient might not understand the cnf claim.
Applications that require
the proof-of-possession keys communicated with it
to be understood and processed must ensure that
the parts of this specification that they use are implemented.
Proof-of-possession via encrypted symmetric secrets is subject to replay attacks.
This attack can be avoided when a signed nonce or challenge is used,
since the recipient can use a distinct nonce or challenge for each interaction.
Replay can also be avoided if a sub-key is derived from a shared secret
that is specific to the instance of the PoP demonstration.
Similarly to other information included in a JWT,
it is necessary to apply data origin authentication and integrity protection
(via a keyed message digest or a digital signature).
Data origin authentication ensures that the recipient of the JWT
learns about the entity that created the JWT,
since this will be important for any policy decisions.
Integrity protection prevents an adversary from changing
any elements conveyed within the JWT payload.
Special care has to be applied when carrying symmetric keys inside the JWT,
since those not only require integrity protection,
but also confidentiality protection.
A proof-of-possession key can be used as a correlation handle if the same key
is used with multiple parties.
Thus, for privacy reasons, it is recommended that different proof-of-possession keys
be used when interacting with different parties.
The following registration procedure is used for all the
registries established by this specification.
Values are registered on a Specification Required
basis after a three-week review period on the oauth-pop-reg-review@ietf.org mailing
list, on the advice of one or more Designated Experts. However, to allow for the
allocation of values prior to publication, the Designated Experts may approve
registration once they are satisfied that such a specification will be published.
[[ Note to the RFC Editor:
The name of the mailing list should be determined in consultation
with the IESG and IANA. Suggested name: oauth-pop-reg-review@ietf.org. ]]
Registration requests sent to the mailing list for review should use
an appropriate subject
(e.g., "Request to register JWT Confirmation Method: example").
Registration requests that are undetermined for
a period longer than 21 days can be brought to the IESG's attention
(using the iesg@ietf.org mailing list) for resolution.
Criteria that should be applied by the Designated Experts include
determining whether the proposed registration duplicates existing functionality,
determining whether it is likely to be of general applicability
or whether it is useful only for a single application,
evaluating the security properties of the item being registered,
and whether the registration makes sense.
It is suggested that multiple Designated Experts be appointed who are able to
represent the perspectives of different applications using this specification,
in order to enable broadly-informed review of registration decisions.
In cases where a registration decision could be perceived as
creating a conflict of interest for a particular Expert,
that Expert should defer to the judgment of the other Experts.
This specification registers the cnf claim in the IANA
"JSON Web Token Claims" registry
established by .
Claim Name: cnf
Claim Description: Confirmation
Change Controller: IESG
Specification Document(s): of [[ this document ]]
This specification establishes the
IANA "JWT Confirmation Methods" registry
for JWT cnf member values.
The registry records the confirmation method member
and a reference to the specification that defines it.
The name requested (e.g., "kid").
Because a core goal of this specification is for the resulting
representations to be compact, it is RECOMMENDED that the name be short
-- not to exceed 8 characters without a compelling reason to do so.
This name is case-sensitive.
Names may not match other registered names in a case-insensitive manner
unless the Designated Experts state that there is a compelling reason
to allow an exception.
Brief description of the confirmation method (e.g., "Key Identifier").
For Standards Track RFCs, list the "IESG". For others, give the name of the
responsible party. Other details (e.g., postal address, email address, home page
URI) may also be included.
Reference to the document or documents that specify the parameter,
preferably including URIs that
can be used to retrieve copies of the documents.
An indication of the relevant
sections may also be included but is not required.
Confirmation Method Value: jwk
Confirmation Method Description: JSON Web Key Representing Public Key
Change Controller: IESG
Specification Document(s): of [[ this document ]]
Confirmation Method Value: jwe
Confirmation Method Description: Encrypted JSON Web Key
Change Controller: IESG
Specification Document(s): of [[ this document ]]
Confirmation Method Value: kid
Confirmation Method Description: Key Identifier
Change Controller: IESG
Specification Document(s): of [[ this document ]]
Confirmation Method Value: jku
Confirmation Method Description: JWK Set URL
Change Controller: IESG
Specification Document(s): of [[ this document ]]
JSON Web Token (JWT)Microsoftmbj@microsoft.comhttp://self-issued.info/Ping Identityve7jtb@ve7jtb.comNomura Research Instituten-sakimura@nri.co.jpJSON Web Key (JWK)Microsoftmbj@microsoft.comhttp://self-issued.info/JSON Web Encryption (JWE)Microsoftmbj@microsoft.comhttp://self-issued.info/Cisco Systems, Inc.jhildebr@cisco.comJSON Web Token ClaimsIANAOpenID Connect Core 1.0Nomura Research Institute, Ltd.Ping IdentityMicrosoftGoogleSalesforceJSON Web Key (JWK) ThumbprintMicrosoftmbj@microsoft.comhttp://self-issued.info/Nomura Research Instituten-sakimura@nri.co.jphttp://nat.sakimura.org/
The authors wish to thank
Brian Campbell,
Stephen Farrell,
Barry Leiba,
Kepeng Li,
Chris Lonvick,
James Manger,
Kathleen Moriarty,
Justin Richer,
and Nat Sakimura
for their reviews of the specification.
[[ to be removed by the RFC Editor before publication as an RFC ]]
-11
Addressed Sec-Dir review comments by Chris Lonvick and
ballot comments by Stephen Farrell.
-10
Addressed ballot comments by Barry Leiba.
-09
Removed erroneous quotation marks around numeric exp claim values in examples.
-08
Added security consideration about also utilizing audience restriction.
-07
Addressed review comments by Hannes Tschofenig, Kathleen Moriarty, and Justin Richer.
Changes were:
Clarified that symmetric proof-of-possession keys
can be generated by either the presenter or the issuer.
Clarified that confirmation members that are not understood must be ignored
unless otherwise specified by the application.
-06
Added diagrams to the introduction.
-05
Addressed review comments by Kepeng Li.
-04
Allowed the use of jwk for symmetric keys
when the JWT is encrypted.
Added the jku (JWK Set URL) member.
Added privacy considerations.
Reordered sections so that the cnf (confirmation)
claim is defined before it is used.
Noted that applications can define new claim names, in addition to
cnf, to represent additional proof-of-possession keys,
using the same representation as cnf.
Applied wording clarifications suggested by Nat Sakimura.
-03
Separated the jwk and
jwe confirmation members;
the former represents a public key as a JWK and
the latter represents a symmetric key as a JWE encrypted JWK.
Changed the title to indicate that a proof-of-possession key
is being communicated.
Updated language that formerly assumed that the issuer was
an OAuth 2.0 authorization server.
Described ways that applications can choose to identify the presenter,
including use of the iss,
sub, and azp claims.
Harmonized the registry language with that used in JWT [RFC 7519].
Addressed other issues identified during working group last call.
Referenced the JWT and JOSE RFCs.
-02
Defined the terms Issuer, Presenter, and Recipient
and updated their usage within the document.
Added a description of a use case using
an asymmetric proof-of-possession key to the introduction.
Added the kid (key ID) confirmation method.
These changes address the open issues identified in the previous draft.
-01
Updated references.
-00
Created the initial working group draft from draft-jones-oauth-proof-of-possession-02.