The mechanisms
defined in [WS-Security] provide the basic mechanisms
on top of which secure messaging semantics can be defined for multiple message
exchanges. This specification defines extensions to allow security context
establishment and sharing, and session key derivation. This allows contexts to
be established and potentially more efficient keys or new key material to be
exchanged, thereby increasing the overall performance and security of the
subsequent exchanges.
The [WS-Security] specification focuses on the message
authentication model. This approach, while useful in many situations, is
subject to several forms of attack (see Security Considerations section of [WS-Security] specification).
Accordingly, this
specification introduces a security context and its usage. The context
authentication model authenticates a series of messages thereby addressing
these shortcomings, but requires additional communications if authentication
happens prior to normal application exchanges.
The security
context is defined as a new [WS-Security] token type
that is obtained using a binding of [WS-Trust].
The primary goals
of this specification are:
·
Define how
security contexts are established
·
Describe how
security contexts are amended
·
Specify how
derived keys are computed and passed
It is not a goal
of this specification to define how trust is established or determined.
This specification
is intended to provide a flexible set of mechanisms that can be used to support
a range of security protocols. Some protocols may require separate mechanisms
or restricted profiles of this specification.
The following list
identifies the key driving requirements:
·
Derived keys and
per-message keys
·
Extensible
security contexts
The [URI]
that MUST be used by implementations of this specification is:
http://docs.oasis-open.org/ws-sx/ws-secureconversation/200512
Table 1 lists XML
namespaces that are used in this specification. The choice of any namespace
prefix is arbitrary and not semantically significant.
Table 1: Prefixes and XML
Namespaces used in this specification.
The schema [XML-Schema1], [XML-Schema2]
for this specification can be located at:
http://docs.oasis-open.org/ws-sx/ws-secureconversation/200512/ws-secureconversation-1.3.xsd
In this document, reference is made to
the wsu:Id attribute in the utility schema. These were added to the
utility schema with the intent that other specifications requiring such an ID
or timestamp could reference it (as is done here).
Claim – A claim
is a statement made about a client, service or other resource (e.g. name,
identity, key, group, privilege, capability, etc.).
Security Token – A security token
represents a collection of claims.
Security
Context – A security context is an abstract concept that refers to an
established authentication state and negotiated key(s) that may have additional
security-related properties.
Security
Context Token – A security context token (SCT) is a wire representation of
that security context abstract concept, which allows a context to be named by a
URI and used with [WS-Security].
Signed
Security Token – A
signed security token is a security token that is
asserted and cryptographically endorsed by a specific authority (e.g. an X.509
certificate or a Kerberos ticket).
Proof-of-Possession Token – A proof-of-possession
(POP) token is a security token that contains secret data that can be
used to demonstrate authorized use of an associated security token. Typically,
although not exclusively, the proof-of-possession information is encrypted with
a key known only to the recipient of the POP token.
Digest – A digest
is a cryptographic checksum of an octet stream.
Signature - A signature [XML-Signature] is a value computed with a cryptographic
algorithm and bound to data in such a way that intended recipients of the data
can use the signature to verify that the data has not been altered and/or has
originated from the signer of the message, providing message integrity and
authentication. The signature can be computed and verified with symmetric key
algorithms, where the same key is used for signing and verifying, or with
asymmetric key algorithms, where different keys are used for signing and
verifying (a private and public key pair are used).
Security
Token Service - A security token service (STS) is a Web service that issues
security tokens (see [WS-Security]). That is, it
makes assertions based on evidence that it trusts, to whoever trusts it (or to
specific recipients). To communicate trust, a service requires proof, such as
a signature, to prove knowledge of a security token or set of security token. A
service itself can generate tokens or it can rely on a separate STS to issue a
security token with its own trust statement (note that for some security token
formats this can just be a re-issuance or co-signature). This forms the basis
of trust brokering.
Request
Security Token (RST)
– A RST is a message sent to a security token service
to request a security token.
Request
Security Token Response (RSTR) – A RSTR is a response to a request
for a security token. In many cases this is a direct response from a security
token service to a requestor after receiving an RST message. However, in
multi-exchange scenarios the requestor and security token service may exchange
multiple RSTR messages before the security token service issues a final RSTR
message. One or more RSTRs are contained within a single
RequestSecurityTokenResponseCollection (RSTRC).
The keywords
"MUST", "MUST NOT", "REQUIRED",
"SHALL", "SHALL NOT", "SHOULD", "SHOULD
NOT", "RECOMMENDED", "MAY", and "OPTIONAL"
in this document are to be interpreted as described in [RFC2119].
Namespace URIs of
the general form "some-URI" represents some application-dependent or
context-dependent URI as defined in [URI ].
This specification
uses the following syntax to define outlines for messages:
·
The syntax appears
as an XML instance, but values in italics indicate data types instead of
literal values.
·
Characters are
appended to elements and attributes to indicate cardinality:
o
"?" (0
or 1)
o
"*" (0
or more)
o
"+" (1
or more)
·
The character
"|" is used to indicate a choice between alternatives.
·
The characters
"(" and ")" are used to indicate that contained items are
to be treated as a group with respect to cardinality or choice.
·
The characters
"[" and "]" are used to call out references and property
names.
·
Ellipses (i.e.,
"...") indicate points of extensibility. Additional children and/or
attributes MAY be added at the indicated extension points but MUST NOT
contradict the semantics of the parent and/or owner, respectively. By default,
if a receiver does not recognize an extension, the receiver SHOULD ignore the
extension; exceptions to this processing rule, if any, are clearly indicated
below.
·
XML namespace prefixes
(see Table 1) are used to indicate the namespace of the element
being defined.
Elements and Attributes defined by
this specification are referred to in the text of this document using XPath 1.0
expressions. Extensibility points are referred to using an extended version of
this syntax:
- An element extensibility point is
referred to using {any} in place of the element name. This indicates that
any element name can be used, from any namespace other than the namespace
of this specification.
- An attribute extensibility point is
referred to using @{any} in place of the attribute name. This indicates
that any attribute name can be used, from any namespace other than the namespace
of this specification.
In this document
reference is made to the wsu:Id
attribute and the wsu:Created and wsu:Expires elements in a utility schema (http://docs.oasis-open.org/wss/2004/01/oasis-200401-wss-wssecurity-utility-1.0.xsd).
The wsu:Id attribute and the wsu:Created and
wsu:Expires elements were added to the utility schema with the
intent that other specifications requiring such an ID type attribute or
timestamp element could reference it (as is done here).
1.6 Normative References
[RFC2119] S. Bradner, "Key words
for use in RFCs to Indicate Requirement Levels", RFC 2119, Harvard University, March 1997.
http://www.ietf.org/rfc/rfc2119.txt
.
[RFC2246] IETF Standard,
"The TLS Protocol", January 1999. http://www.ietf.org/rfc/rfc2246.txt
[SOAP] W3C Note,
"SOAP: Simple Object Access Protocol 1.1", 08 May 2000.
http://www.w3.org/TR/2000/NOTE-SOAP-20000508/.
[SOAP12] W3C Recommendation, "SOAP 1.2
Part 1: Messaging Framework", 24 June 2003.
http://www.w3.org/TR/2003/REC-soap12-part1-20030624/
[URI] T.
Berners-Lee, R. Fielding, L. Masinter, "Uniform Resource Identifiers
(URI): Generic Syntax", RFC 3986, MIT/LCS, Day Software, Adobe Systems,
January 2005.
http://www.ietf.org/rfc/rfc3986.txt
[WS-Addressing] W3C
Recommendation, "Web Services Addressing (WS-Addressing)", 9 May
2006.
http://www.w3.org/TR/2006/REC-ws-addr-core-20060509.
[WS-Security] OASIS Standard, "OASIS Web
Services Security: SOAP Message Security 1.0 (WS-Security 2004)", March
2004.
http://docs.oasis-open.org/wss/2004/01/oasis-200401-wss-soap-message-security-1.0.pdf
OASIS Standard, "OASIS Web Services
Security: SOAP Message Security 1.1 (WS-Security 2004)", February 2006.
http://www.oasis-open.org/committees/download.php/16790/wss-v1.1-spec-os-SOAPMessageSecurity.pdf
[WS-Trust] OASIS
Committee Draft, “WS-Trust 1.4”, 2008
http://docs.oasis-open.org/ws-sx/ws-trust/200802
[XML-Encrypt] W3C
Recommendation, "XML Encryption Syntax and Processing", 10 December
2002.
http://www.w3.org/TR/2002/REC-xmlenc-core-20021210/.
[XML-Schema1] W3C Recommendation,
"XML Schema Part 1: Structures Second Edition", 28 October 2004.
http://www.w3.org/TR/2004/REC-xmlschema-1-20041028/.
[XML-Schema2] W3C Recommendation,
"XML Schema Part 2: Datatypes Second Edition", 28 October 2004.
http://www.w3.org/TR/2004/REC-xmlschema-2-20041028/.
[XML-Signature] W3C
Recommendation, "XML-Signature Syntax and Processing", 12 February
2002.
http://www.w3.org/TR/2002/REC-xmlenc-core-20021210/
[WS-MEX] "Web
Services Metadata Exchange (WS-MetadataExchange)", BEA, Computer
Associates, IBM, Microsoft, SAP, Sun Microsystems, Inc., webMethods, September
2004.
[WS-SecurityPolicy] OASIS Standard,
“WS-SecurityPolicy 1.3”, 2008
http://docs.oasis-open.org/ws-sx/ws-securitypolicy/200802
While message
authentication is useful for simple or one-way messages, parties that wish to
exchange multiple messages typically establish a security context in which to
exchange multiple messages.
A security context is shared among the communicating parties for the lifetime
of a communications session.
In this
specification, a security context is represented by the <wsc:SecurityContextToken> security token. In the [WS-Security] and [WS-Trust]
framework, the following URI is used to represent the token type:
http://docs.oasis-open.org/ws-sx/ws-secureconversation/200512/sct
The Security
Context Token does not support references to it using key identifiers or key
names. All references MUST either use an ID (to a wsu:Id
attribute) or a <wsse:Reference> to the <wsc:Identifier>
element.
Once the context
and secret have been established (authenticated), the mechanisms described in Derived
Keys can be used to
compute derived keys for each
key usage in the secure
context.
The following
illustration represents an overview of the syntax of the <wsc:SecurityContextToken> element. It should be noted that this
token supports an open content model to allow context-specific data to be
passed.
<wsc:SecurityContextToken
wsu:Id="..." xmlns:wsc="..." xmlns:wsu="..." ...>
<wsc:Identifier>...</wsc:Identifier>
<wsc:Instance>...</wsc:Instance>
...
</wsc:SecurityContextToken>
The following
describes elements and attributes used in a <wsc:SecurityContextToken> element.
/wsc:SecurityContextToken
This element is a security token that describes a security
context.
/wsc:SecurityContextToken/wsc:Identifier
This REQUIRED element identifies the security context using
an absolute URI. Each security context URI MUST be unique to both the sender
and recipient. It is RECOMMENDED that the value be globally unique in time and
space.
/wsc:SecurityContextToken/wsc:Instance
When contexts are renewed and given different keys it is
necessary to identify the different key instances without revealing the actual
key. When present this OPTIONAL element contains a string that is unique for a
given key value for this wsc:Identifier.
The initial issuance need not contain a wsc:Instance
element, however, all subsequent issuances with different keys MUST have a wsc:Instance element with a unique
value.
/wsc:SecurityContextToken/@wsu:Id
This OPTIONAL attribute specifies a string label for this element.
/wsc:SecurityContextToken/@{any}
This is an extensibility mechanism to allow additional
attributes, based on schemas, to be added to the element.
/wsc:SecurityContextToken/{any}
This is an extensibility mechanism to allow additional
elements (arbitrary content) to be used.
The
<wsc:SecurityContextToken> token elements MUST be preserved.
That is, whatever elements contained within the tag on creation MUST be
preserved wherever the token is used. A consumer of a <wsc:SecurityContextToken> token MAY extend the token by
appending information. Consequently producers of <wsc:SecurityContextToken> tokens should consider this fact when
processing previously generated tokens. A service consuming (processing) a <wsc:SecurityContextToken> token MAY fault if it discovers an
element or attribute inside the token that it doesn't understand, or it MAY
ignore it. The fault code wsc:UnsupportedContextToken is RECOMMENDED if a fault is raised. The behavior is specified by the
services policy [WS-SecurityPolicy]. Care should be taken when adding
information to tokens to ensure that relying parties can ensure the information
has not been altered since the SCT definition does not require a specific way
to secure its contents (which as noted above can be appended to).
Security contexts,
like all security tokens, can be referenced using the mechanisms described in [WS-Security] (the <wsse:SecurityTokenReference> element referencing the wsu:Id
attribute relative to the XML base document or referencing using the <wsc:Identifier> element's absolute URI). When a token is referenced, the
associated key is used. If a token provides multiple keys then specific
bindings and profiles MUST describe how to reference the separate keys. If a
specific key instance needs to be referenced, then the global attribute wsc:Instance is included in the <wsse:Reference> sub-element (only when using <wsc:Identifier>
references) of the <wsse:SecurityTokenReference> element as illustrated below:
<wsse:SecurityTokenReference
xmlns:wsse="..." xmlns:wsc="...">
<wsse:Reference URI="uuid:... " wsc:Instance="..."/>
</wsse:SecurityTokenReference>
The following
sample message illustrates the use of a security context token. In this
example a context has been established and the secret is known to both
parties. This secret is used to sign the message body.
(001)
<?xml version="1.0" encoding="utf-8"?>
(002)
<S11:Envelope xmlns:S11="..." xmlns:ds="..." xmlns:wsse="..."
xmlns:wsu="..." xmlns:wsc="...">
(003)
<S11:Header>
(004)
...
(005)
<wsse:Security>
(006)
<wsc:SecurityContextToken wsu:Id="MyID">
(007) <wsc:Identifier>uuid:...</wsc:Identifier>
(008)
</wsc:SecurityContextToken>
(009)
<ds:Signature>
(010)
...
(011)
<ds:KeyInfo>
(012)
<wsse:SecurityTokenReference>
(013)
<wsse:Reference URI="#MyID"/>
(014)
</wsse:SecurityTokenReference>
(015)
</ds:KeyInfo>
(016)
</ds:Signature>
(017)
</wsse:Security>
(018)
</S11:Header>
(019)
<S11:Body wsu:Id="MsgBody">
(020)
<tru:StockSymbol
xmlns:tru="http://fabrikam123.com/payloads">
QQQ
</tru:StockSymbol>
(021)
</S11:Body>
(022)
</S11:Envelope>
Let's review some
of the key sections of this example:
Lines (003)-(018)
contain the SOAP message headers.
Lines (005)-(017)
represent the <wsse:Security> header block. This contains the
security-related information for the message.
Lines (006)-(008)
specify a security token that is associated
with the message. In this case it is a security context token. Line (007)
specifies the unique ID of the context.
Lines (009)-(016)
specify the digital signature. In this example, the signature is based on the
security context (specifically the secret/key associated with the context). Line
(010) represents the typical contents of an XML Digital Signature which, in
this case, references the body and potentially some of the other headers
expressed by line (004).
Lines (012)-(014)
indicate the key that was used for the signature. In this case, it is the
security context token included in the message. Line (013) provides a URI link
to the security context token specified in Lines (006)-(008).
The body of the
message is represented by lines (019)-(021).
A security context
needs to be created and shared by the communicating parties before being used.
This specification defines three different ways of establishing a security
context among the parties of a secure communication.
Security
context token created by a security token service – The context initiator asks a
security token service to create a new security context token. The newly
created security context token is distributed to the parties through the mechanisms
defined here and in [WS-Trust]. For this scenario the
initiating party sends a <wst:RequestSecurityToken>
request to the token
service and a <wst:RequestSecurityTokenResponseCollection> containing a
<wst:RequestSecurityTokenResponse> is returned. The response contains a
<wst:RequestedSecurityToken> containing (or pointing to) the new
security context token and a <wst:RequestedProofToken> pointing to the "secret"
for the returned context. The requestor then uses the security context token
(with [WS-Security]) when securing messages to
applicable services.
Security
context token created by one of the communicating parties and
propagated with a message
– The initiator creates a security context token and sends it to the other
parties on a message using the mechanisms described in this specification and
in [WS-Trust]. This model works when the sender is
trusted to always create a new security context token. For this scenario the
initiating party creates a security context token and issues a signed
unsolicited <wst:RequestSecurityTokenResponse> to the other party. The message
contains a <wst:RequestedSecurityToken> containing (or pointing to) the new
security context token and a <wst:RequestedProofToken> pointing to the "secret"
for the security context token. The recipient can then choose whether or not
to accept the security context token. As described in [WS-Trust],
the <wst:RequestSecurityTokenResponse> element MAY be in the <wst:RequestSecurityTokenResponseCollection>
within a body or
inside a header block. It should be noted that unless delegation tokens are
used, this scenario requires that parties trust each other to share a secret
key (and non-repudiation is probably not possible). As receipt of these
messages may be expensive, and because a recipient may receive multiple
messages, the
…/wst:RequestSecurityTokenResponse/@Context attribute in [WS-Trust]
allows the initiator to specify a URI to indicate the intended usage (allowing
processing to be optimized).
Security
context token created through negotiation/exchanges – When there is a need to negotiate
or participate in a sequence of message exchanges among the participants on the
contents of the security context token, such as the shared secret, this
specification allows the parties to exchange data to establish a security context.
For this scenario the initiating party sends a <wst:RequestSecurityToken> request to the other party and a <wst:RequestSecurityTokenResponse> is returned. It is RECOMMENDED that
the framework described in [WS-Trust] be used; however,
the type of exchange will likely vary. If appropriate, the basic
challenge-response definition in [WS-Trust] is
RECOMMENDED. Ultimately (if successful), a final response contains a <wst:RequestedSecurityToken> containing (or pointing to) the new
security context and a <wst:RequestedProofToken> pointing to the "secret"
for the context.
If an SCT is
received, but the key sizes are not supported, then a fault SHOULD be generated
using the wsc:UnsupportedContextToken fault code unless another more
specific fault code is available.
This binding
describes how to use [WS-Trust] to request and return
SCTs. This binding builds on the issuance binding for [WS-Trust]
(note that other sections of this specification define new separate bindings of
[WS-Trust]). Consequently, aspects of the issuance
binding apply to this binding unless otherwise stated. For example, the token
request type is the same as in the issuance binding.
When requesting
and returning security context tokens the following Action URIs [WS-Addressing] are used (note that a specialized
action is used here because of the specialized semantics of SCTs):
As with all token
services, the options supported may be limited. This is especially true of
SCTs because the issuer may only be able to issue tokens for itself and quite
often will only support a specific set of algorithms and parameters as
expressed in its policy.
SCTs are not
required to have lifetime semantics. That is, some SCTs may have specific
lifetimes and others may be bound to other resources rather than have their own
lifetimes.
Since the SCT
binding builds on the issuance binding, it allows the optional extensions
defined for the issuance binding including the use of exchanges. Subsequent
profiles MAY restrict the extensions and types and usage of exchanges.
The following
illustrates a request for a SCT from a security token service. The request in
this example contains no information concerning the Web Service with whom the
requestor wants to communicate securely (e.g. using the wsp:AppliesTo parameter
in the RST). In order for the security token service to process this request it
MSUT have prior knowledge for which Web Service the requestor needs a token.
This may be preconfigured although it is typically passed in the RST. In this
example the key is encrypted for the recipient (security token service) using
the token service's X.509 certificate as per XML Encryption [XML-Encrypt].
The encrypted data (using the encrypted key) contains a <wsse:UsernameToken> token that the recipient uses to authorize the request.
The request is secured (integrity) using the X.509 certificate of the
requestor. The response encrypts the proof information using the requestor's
X.509 certificate and secures the message (integrity) using the token service's
X.509 certificate. Note that the details of XML Signature and XML Encryption
have been omitted; refer to [WS-Security] for
additional details. It should be noted that if the requestor doesn't have an
X.509 certificate this scenario could be achieved using a TLS [RFC2246]
connection or by creating an ephemeral key.
<S11:Envelope
xmlns:S11="..." xmlns:wsse="..." xmlns:wsu="..."
xmlns:wst="..." xmlns:xenc="...">
<S11:Header>
...
<wsa:Action xmlns:wsa="...">
http://docs.oasis-open.org/ws-sx/ws-trust/200512/RST/SCT
</wsa:Action>
...
<wsse:Security>
<xenc:EncryptedKey>
...
</xenc:EncryptedKey>
<xenc:EncryptedData Id="encUsernameToken">
... encrypted username token (whose id is myToken) ...
</xenc:EncryptedData>
<ds:Signature xmlns:ds="...">
...
<ds:KeyInfo>
<wsse:SecurityTokenReference>
<wsse:Reference URI="#myToken"/>
</wsse:SecurityTokenReference>
</ds:KeyInfo>
</ds:Signature>
</wsse:Security>
...
</S11:Header>
<S11:Body
wsu:Id="req">
<wst:RequestSecurityToken>
<wst:TokenType>
http://docs.oasis-open.org/ws-sx/ws-secureconversation/200512/sct
</wst:TokenType>
<wst:RequestType>
http://docs.oasis-open.org/ws-sx/ws-trust/200512/Issue
</wst:RequestType>
</wst:RequestSecurityToken>
</S11:Body>
</S11:Envelope>
<S11:Envelope
xmlns:S11="..."
xmlns:wst="..." xmlns:wsc="..."
xmlns:xenc="...">
<S11:Header>
...
<wsa:Action xmlns:wsa="...">
http://docs.oasis-open.org/ws-sx/ws-trust/200512/RSTR/SCT
</wsa:Action>
...
</S11:Header>
<S11:Body>
<wst:RequestSecurityTokenResponseCollection>
<wst:RequestSecurityTokenResponse>
<wst:RequestedSecurityToken>
<wsc:SecurityContextToken>
<wsc:Identifier>uuid:...</wsc:Identifier>
</wsc:SecurityContextToken>
</wst:RequestedSecurityToken>
<wst:RequestedProofToken>
<xenc:EncryptedKey
Id="newProof">
...
</xenc:EncryptedKey>
</wst:RequestedProofToken>
</wst:RequestSecurityTokenResponse>
</wst:RequestSecurityTokenResponseCollection>
</S11:Body>
</S11:Envelope>
There are
scenarios where a security token service is used to broker trust using SCT
tokens between requestors and Web Services endpoints. In these cases it is
typical for requestors to identify the target Web Service in the RST.
In the example
below the requestor uses the element <wsp:AppliesTo> with an endpoint
reference as described in [WS-Trust]
in the SCT request to indicate the Web Service the token is needed for.
In the request
example below the <wst:TokenType> element is omitted. This requires that
the security token service know what type of token the endpoint referenced in
the <wsp:AppliesTo> element expects.
<S11:Envelope
xmlns:S11="..." xmlns:wsse="..." xmlns:wsu="..."
xmlns:wst="..." xmlns:xenc="..." xmlns:wsp="..."
xmlns:wsa="...">
<S11:Header>
...
<wsa:Action xmlns:wsa="...">
http://docs.oasis-open.org/ws-sx/ws-trust/200512/RST/SCT
</wsa:Action>
...
<wsse:Security>
...
</wsse:Security>
...
</S11:Header>
<S11:Body wsu:Id="req">
<wst:RequestSecurityToken>
<wst:RequestType>
http://docs.oasis-open.org/ws-sx/ws-trust/200512/Issue
</wst:RequestType>
<wsp:AppliesTo>
<wsa:EndpointReference>
<wsa:Address>http://example.org/webservice</wsa:Address>
</wsa:EndpointReference>
</wsp:AppliesTo>
</wst:RequestSecurityToken>
</S11:Body>
</S11:Envelope>
<S11:Envelope
xmlns:S11="..."
xmlns:wst="..." xmlns:wsc="..." xmlns:xenc="..." xmlns:wsp="..."
xmlns:wsa="...">
<S11:Header>
<wsa:Action xmlns:wsa="...">
http://docs.oasis-open.org/ws-sx/ws-trust/200512/RSTR/SCT
</wsa:Action>
...
</S11:Header>
<S11:Body>
<wst:RequestSecurityTokenResponseCollection>
<wst:RequestSecurityTokenResponse>
<wst:RequestedSecurityToken>
<wsc:SecurityContextToken>
<wsc:Identifier>uuid:...</wsc:Identifier>
</wsc:SecurityContextToken>
</wst:RequestedSecurityToken>
<wst:RequestedProofToken>
<xenc:EncryptedKey Id="newProof">
...
</xenc:EncryptedKey>
</wst:RequestedProofToken>
<wsp:AppliesTo>
<wsa:EndpointReference>
<wsa:Address>http://example.org/webservice</wsa:Address>
</wsa:EndpointReference>
</wsp:AppliesTo>
</wst:RequestSecurityTokenResponse>
</wst:RequestSecurityTokenResponseCollection>
</S11:Body>
</S11:Envelope>
The following
illustrates propagating a context to another party. This example does not
contain any information regarding the Web Service the SCT is intended for (e.g.
using the wsp:AppliesTo parameter in the RST).
<S11:Envelope
xmlns:S11="..."
xmlns:wst="..." xmlns:wsc="..."
xmlns:xenc="..." >
<S11:Header>
...
</S11:Header>
<S11:Body>
<wst:RequestSecurityTokenResponse>
<wst:RequestedSecurityToken>
<wsc:SecurityContextToken>
<wsc:Identifier>uuid:...</wsc:Identifier>
</wsc:SecurityContextToken>
</wst:RequestedSecurityToken>
<wst:RequestedProofToken>
<xenc:EncryptedKey
Id="newProof">
...
</xenc:EncryptedKey>
</wst:RequestedProofToken>
</wst:RequestSecurityTokenResponse>
</S11:Body>
</S11:Envelope>
When an SCT is
created, a set of claims is associated with it. There are times when an
existing SCT needs to be amended to carry additional claims (note that the
decision as to who is authorized to amend a context is a service-specific
decision). This is done using the SCT Amend binding. In such cases an
explicit request is made to amend the claims associated with an SCT. It should
be noted that using the mechanisms described in [WS-Trust],
an issuer MAY, at any time, return an amended SCT by issuing an unsolicited
(not explicitly requested) SCT inside an RSTR (either as a separate message or
in a header).
The following
Action URIs are used with this binding:
This binding
allows optional extensions but DOES NOT allow key semantics to be altered.
Proof of possession of the key associated with the security
context MUST be proven in order for context to be amended. It is RECOMMENDED
that the proof of possession is done by creating a signature over the message body
and crucial headers using the key associated with the security context.
Additional claims to amend the security context with MUST be
indicated by providing signatures over the security context signature created
using the key associated with the security context. Those additional signatures
are used to prove additional security tokens that carry claims to augment the
security context.
This binding uses
the request type from the issuance binding.
<S11:Envelope
xmlns:S11="..." xmlns:wsse="..." xmlns:wsu="..."
xmlns:wst="..." xmlns:wsc="...">
<S11:Header>
...
<wsa:Action xmlns:wsa="...">
http://docs.oasis-open.org/ws-sx/ws-trust/200512/RST/SCT/Amend
</wsa:Action>
...
<wsse:Security>
<xx:CustomToken wsu:Id="cust" xmlns:xx="...">
...
</xx:CustomToken>
<ds:Signature xmlns:ds="...">
...signature over #sig1 using #cust...
</ds:Signature>
<wsc:SecurityContextToken wsu:Id="sct">
<wsc:Identifier>uuid:...UUID1...</wsc:Identifier>
</wsc:SecurityContextToken>
<ds:Signature xmlns:ds="..." Id="sig1">
...signature over body and key headers using #sct...
<ds:KeyInfo>
<wsse:SecurityTokenReference>
<wsse:Reference URI="#sct"/>
</wsse:SecurityTokenReference>
</ds:KeyInfo>
...
</ds:Signature>
</wsse:Security>
...
</S11:Header>
<S11:Body wsu:Id="req">
<wst:RequestSecurityToken>
<wst:RequestType>
http://docs.oasis-open.org/ws-sx/ws-trust/200512/Issue
</wst:RequestType>
</wst:RequestSecurityToken>
</S11:Body>
</S11:Envelope>
<S11:Envelope
xmlns:S11="..." xmlns:wst="..."
xmlns:wsc="...">
<S11:Header>
...
<wsa:Action xmlns:wsa="...">
http://docs.oasis-open.org/ws-sx/ws-trust/200512/RSTR/SCT/Amend
</wsa:Action>
...
</S11:Header>
<S11:Body>
<wst:RequestSecurityTokenResponseCollection>
<wst:RequestSecurityTokenResponse>
<wst:RequestedSecurityToken>
<wsc:SecurityContextToken>
<wsc:Identifier>uuid:...UUID1...</wsc:Identifier>
</wsc:SecurityContextToken>
</wst:RequestedSecurityToken>
</wst:RequestSecurityTokenResponse>
</wst:RequestSecurityTokenResponseCollection>
</S11:Body>
</S11:Envelope>
When a security
context is created it typically has an associated expiration. If a requestor
desires to extend the duration of the token it uses this specialized binding of
the renewal mechanism defined in WS-Trust. The following Action URIs are used
with this binding:
This binding
allows optional extensions but DOES NOT allow key semantics to be altered.
A renewal MUST include re-authentication of the original
claims because the original claims might have an expiration time that conflicts
with the requested expiration time in the renewal request. Because the security
context token issuer is not required to cache such information from the
original issuance request, the requestor is REQUIRED to re-authenticate the
original claims in every renewal request. It is RECOMMENDED that the original claims
re-authentication is done in the same way as in the original token issuance
request.
Proof of possession of the key associated with the security
context MUST be proven in order for security context to be renewed. It is
RECOMMENDED that this is done by creating the original claims signature over
the signature that signs message body and crucial headers.
During renewal, new key material MAY be exchanged. Such key
material MUST NOT be protected using the existing session key.
This binding uses
the request type from the renewal binding.
The following
example illustrates a renewal which re-proves the original claims.
<S11:Envelope
xmlns:S11="..." xmlns:wsse="..." xmlns:wsu="..."
xmlns:wst="..." xmlns:wsc="...">
<S11:Header>
...
<wsa:Action xmlns:wsa="...">
http://docs.oasis-open.org/ws-sx/ws-trust/200512/RST/SCT/Renew
</wsa:Action>
...
<wsse:Security>
<xx:CustomToken wsu:Id="cust" xmlns:xx="...">
...
</xx:CustomToken>
<ds:Signature xmlns:ds="..." Id="sig1">
... signature over body and key headers using #cust...
</ds:Signature>
<wsc:SecurityContextToken wsu:Id="sct">
<wsc:Identifier>uuid:...UUID1...</wsc:Identifier>
</wsc:SecurityContextToken>
<ds:Signature xmlns:ds="..." Id="sig2">
... signature over #sig1 using #sct ...
</ds:Signature>
</wsse:Security>
...
</S11:Header>
<S11:Body
wsu:Id="req">
<wst:RequestSecurityToken>
<wst:RequestType>
http://docs.oasis-open.org/ws-sx/ws-trust/200512/Renew
</wst:RequestType>
<wst:RenewTarget>
<wsse:SecurityTokenReference>
<wsse:Reference URI="uuid:...UUID1..."/>
</wsse:SecurityTokenReference>
</wst:RenewTarget>
<wst:Lifetime>...</wst:Lifetime>
</wst:RequestSecurityToken>
</S11:Body>
</S11:Envelope>
<S11:Envelope
xmlns:S11="..." xmlns:wst="..."
xmlns:wsc="...">
<S11:Header>
...
<wsa:Action xmlns:wsa="...">
http://docs.oasis-open.org/ws-sx/ws-trust/200512/RSTR/SCT/Renew
</wsa:Action>
...
</S11:Header>
<S11:Body>
<wst:RequestSecurityTokenResponseCollection>
<wst:RequestSecurityTokenResponse>
<wst:RequestedSecurityToken>
<wsc:SecurityContextToken>
<wsc:Identifier>uuid:...UUID1...</wsc:Identifier>
<wsc:Instance>UUID2</wsc:Instance>
</wsc:SecurityContextToken>
</wst:RequestedSecurityToken>
<wst:Lifetime>...</wst:Lifetime>
</wst:RequestSecurityTokenResponse>
</wst:RequestSecurityTokenResponseCollection>
</S11:Body>
</S11:Envelope>
It is not uncommon
for a requestor to be done with a security context token before it expires. In
such cases the requestor can explicitly cancel the security context using this
specialized binding based on the WS-Trust Cancel binding.
The following
Action URIs are used with this binding:
Once a security
context has been cancelled it MUST NOT be allowed for authentication or
authorization or allow renewal.
Proof
of possession of the key associated with the security context MUST be proven in
order for security context to be cancelled. It is RECOMMENDED that this is done
by creating a signature over the message body and crucial headers using the key
associated with the security context.
This binding uses
the Cancel request type from WS-Trust.
As described in
WS-Trust the RSTR cancel message is informational and the context is cancelled
once the cancel RST is processed even if the cancel RSTR is never received by
the requestor.
The following
example illustrates canceling a context.
<S11:Envelope
xmlns:S11="..." xmlns:wsse="..." xmlns:wsu="..."
xmlns:wst="..." xmlns:wsc="...">
<S11:Header>
...
<wsa:Action xmlns:wsa="...">
http://docs.oasis-open.org/ws-sx/ws-trust/200512/RST/SCT/Cancel
</wsa:Action>
...
<wsse:Security>
<wsc:SecurityContextToken wsu:Id="sct">
<wsc:Identifier>uuid:...UUID1...</wsc:Identifier>
</wsc:SecurityContextToken>
<ds:Signature xmlns:ds="..." Id="sig1">
...signature over body and key headers using #sct...
</ds:Signature>
</wsse:Security>
...
</S11:Header>
<S11:Body wsu:Id="req">
<wst:RequestSecurityToken>
<wst:RequestType>
http://docs.oasis-open.org/ws-sx/ws-trust/200512/Cancel
</wst:RequestType>
<wst:CancelTarget>
<wsse:SecurityTokenReference>
<wsse:Reference URI="uuid:...UUID1..."/>
</wsse:SecurityTokenReference>
</wst:CancelTarget>
</wst:RequestSecurityToken>
</S11:Body>
</S11:Envelope>
<S11:Envelope
xmlns:S11="..." xmlns:wst="..." >
<S11:Header>
...
<wsa:Action xmlns:wsa="...">
http://docs.oasis-open.org/ws-sx/ws-trust/200512/RSTR/SCT/Cancel
</wsa:Action>
...
</S11:Header>
<S11:Body>
<wst:RequestSecurityTokenResponseCollection>
<wst:RequestSecurityTokenResponse>
<wst:RequestedTokenCancelled/>
</wst:RequestSecurityTokenResponse>
</wst:RequestSecurityTokenResponseCollection>
</S11:Body>
</S11:Envelope>
A security context
token implies or contains a shared secret. This secret MAY be used for signing
and/or encrypting messages, but it is RECOMMENDED that derived keys be used for
signing and encrypting messages associated only with the security context.
Using a common
secret, parties MAY define different key derivations to use. For example, four
keys may be derived so that two parties can sign and encrypt using separate
keys. In order to keep the keys fresh (prevent providing too much data for
analysis), subsequent derivations MAY be used. We introduce the <wsc:DerivedKeyToken> token as a mechanism for indicating which derivation is
being used within a given message.
The derived key
mechanism can use different algorithms for deriving keys. The algorithm is
expressed using a URI. This specification defines one such algorithm.
As well, while
presented here using security context tokens, the <wsc:DerivedKeyToken> token can be used to derive keys from any security token
that has a shared secret, key, or key material.
We use a subset of
the mechanism defined for TLS in RFC 2246. Specifically, we use the P_SHA-1
function to generate a sequence of bytes that can be used to generate security
keys. We refer to this algorithm as:
http://docs.oasis-open.org/ws-sx/ws-secureconversation/200512/dk/p_sha1
This function is
used with three values – secret, label,
and seed. The secret is the shared secret that is
exchanged (note that if two secrets were securely exchanged, possibly as part
of an initial exchange, they are concatenated in the order they were
sent/received). Secrets are processed as octets representing their binary
value (value prior to encoding). The label is the concatenation of the
client's label and the service's label. These labels can be discovered in each
party's policy (or specifically within a <wsc:DerivedKeyToken> token). Labels are processed as
UTF-8 encoded octets. If additional information is not specified as explicit
elements, then a default value of "WS-SecureConversation"
(represented as UTF-8 octets) is used. The seed is the concatenation of nonce
values (if multiple were exchanged) that were exchanged (initiator +
receiver). The nonce is processed as a binary octet sequence (the value prior
to base64 encoding). The nonce seed is REQUIRED, and MUST be generated by one
or more of the communicating parties. The P_SHA-1 function has two parameters
– secret and value. We
concatenate the label and the seed
to create the value. That is:
P_SHA1 (secret, label + seed)
At this point,
both parties can use the P_SHA-1 function to generate shared keys as needed.
For this protocol, we don't define explicit derivation uses.
The <wsc:DerivedKeyToken> element is used to indicate that the key for a specific
reference is generated from the function. This is so that explicit security
tokens, secrets, or key material need not be exchanged as often thereby
increasing efficiency and overall scalability. However, parties MUST mutually
agree on specific derivations (e.g. the first 128 bits is the client's
signature key, the next 128 bits in the client's encryption key, and so on).
The policy presents a method for specifying this information. The RECOMMENDED
approach is to use separate nonces and have independently generated keys for
signing and encrypting in each direction. Furthermore, it is RECOMMENDED that
new keys be derived for each message (i.e., previous nonces are not re-used).
Once the parties
determine a shared secret to use as the basis of a key generation sequence, an
initial key is generated using this sequence. When a new key is required, a
new <wsc:DerivedKeyToken> MAY be passed referencing the
previously generated key. The recipient then knows to use the sequence to
generate a new key, which will match that specified in the security token. If
both parties pre-agree on key sequencing, then additional token exchanges are not
required.
For keys derived
using a shared secret from a security context, the <wsse:SecurityTokenReference> element SHOULD be used to reference
the <wsc:SecurityContextToken>. Basically, a signature or
encryption references a <wsc:DerivedKeyToken> in the <wsse:Security> header
that, in turn, references the <wsc:SecurityContextToken>.
Derived keys are
expressed as security tokens. The
following URI is used to represent the token type:
http://docs.oasis-open.org/ws-sx/ws-secureconversation/200512/dk
The derived key
token does not support references using key identifiers or key names. All
references MUST use an ID (to a wsu:Id attribute) or
a URI reference to the <wsc:Identifier> element in the SCT.
The following illustrates the syntax for <wsc:DerivedKeyToken>:
<wsc:DerivedKeyToken wsu:Id="..."
Algorithm="..." xmlns:wsc="..." xmlns:wsse="..."
xmlns:wsu="...">
<wsse:SecurityTokenReference>...</wsse:SecurityTokenReference>
<wsc:Properties>...</wsc:Properties>
<wsc:Generation>...</wsc:Generation>
<wsc:Offset>...</wsc:Offset>
<wsc:Length>...</wsc:Length>
<wsc:Label>...</wsc:Label>
<wsc:Nonce>...</wsc:Nonce>
</wsc:DerivedKeyToken>
The following
describes the attributes and tags listed in the schema overview above:
/wsc:DerivedKeyToken
This specifies a key that is derived from a shared secret.
/wsc:DerivedKeyToken/@wsu:Id
This OPTIONAL attribute specifies an XML ID that can be
used locally to reference this element.
/wsc:DerivedKeyToken/@Algorithm
This OPTIONAL URI attribute specifies key derivation
algorithm to use. This specification predefines the P_SHA1
algorithm described above. If this attribute isn't specified, this algorithm
is assumed.
/wsc:DerivedKeyToken/wsse:SecurityTokenReference
This OPTIONAL element is used to specify security context
token, security token, or shared key/secret used for the derivation. If not
specified, it is assumed that the recipient can determine the shared key from
the message context. If the context cannot be determined, then a fault such as
wsc:UnknownDerivationSource SHOULD be raised.
/wsc:DerivedKeyToken/wsc:Properties
This OPTIONAL element allows metadata to be associated with
this derived key. For example, if the <wsc:Name>
property is defined, this derived key is given a URI name that can then be used
as the source for other derived keys. The <wsc:Nonce>
and <wsc:Label> elements can be
specified as properties and indicate the nonce and label to use (defaults) for
all keys derived from this key.
/wsc:DerivedKeyToken/wsc:Properties/wsc:Name
This OPTIONAL element is used to give this derived key a
URI name that can then be used as the source for other derived keys.
/wsc:DerivedKeyToken/wsc:Properties/wsc:Label
This OPTIONAL element defines a label to use for all keys
derived from this key. See /wsc:DerivedKeyToken/wsc:Label defined below.
/wsc:DerivedKeyToken/wsc:Properties/wsc:Nonce
This OPTIONAL element defines a nonce to use for all keys
derived from this key. See /wsc:DerivedKeyToken/wsc:Nonce defined below.
/wsc:DerivedKeyToken/wsc:Properties/{any}
This is an extensibility mechanism to allow additional
elements (arbitrary content) to be used.
/wsc:DerivedKeyToken/wsc:Generation
If fixed-size keys (generations) are being generated, then
this OPTIONAL element can be used to specify which generation of the key to
use. The value of this element is an unsigned long value indicating the
generation number to use (beginning with zero). This element MUST NOT be used
if the <wsc:Offset> element is specified.
Specifying this element is equivalent to specifying the <wsc:Offset> and <wsc:Length>
elements having multiplied out the values. That is, offset = (generation) *
fixed_size and length = fixed_size.
/wsc:DerivedKeyToken/wsc:Offset
If fixed-size keys are not being generated, then the <wsc:Offset> and <wsc:Length> elements indicate where in the
byte stream to find the generated key. This specifies the ordering (in bytes)
of the generated output. The value of this OPTIONAL element is an unsigned
long value indicating the byte position (starting at 0). For example, 0
indicates the first byte of output and 16 indicates the 17th byte of
generated output. This element MUST NOT be used if the <wsc:Generation> element is specified. It
should be noted that not all algorithms will support the <wsc:Offset> and <wsc:Length> elements.
/wsc:DerivedKeyToken/wsc:Length
This element specifies the length (in bytes) of the derived
key. This OPTIONAL element can be specified in conjunction with <wsc:Offset> or <wsc:Generation>. If this isn't specified, it
is assumed that the recipient knows the key size to use. The value of this
element is an unsigned long value indicating the size of the key in bytes
(e.g., 16).
/wsc:DerivedKeyToken/wsc:Label
The label can be specified within a
<wsc:DerivedKeyToken> using the wsc:Label element. If the label isn't
specified then a default value of
"WS-SecureConversationWS-SecureConversation" (represented as UTF-8
octets) is used. Labels are processed as UTF-8 encoded octets.
/wsc:DerivedKeyToken/wsc:Nonce
If specified, this OPTIONAL element specifies a base64
encoded nonce that is used in the key derivation function for this derived
key. If this isn't specified, it is assumed that the recipient knows the nonce
to use. Note that once a nonce is used for a derivation sequence, the same
nonce SHOULD NOT be used for all subsequent derivations.
If additional
information is not specified as explicit elements, then the following defaults
apply:
·
The offset is 0
·
The length is 32
bytes (256 bits)
It is RECOMMENDED
that separate derived keys be used to strengthen the cryptography. If multiple
keys are used, then care should be taken not to derive too many times and risk
key attacks.
The following
example illustrates a message sent using two derived keys, one for signing and
one for encrypting:
<S11:Envelope
xmlns:S11="..." xmlns:wsse="..." xmlns:wsu="..."
xmlns:xenc="..." xmlns:wsc="..."
xmlns:ds="...">
<S11:Header>
<wsse:Security>
<wsc:SecurityContextToken wsu:Id="ctx2">
<wsc:Identifier>uuid:...UUID2...</wsc:Identifier>
</wsc:SecurityContextToken>
<wsc:DerivedKeyToken wsu:Id="dk2">
<wsse:SecurityTokenReference>
<wsse:Reference URI="#ctx2"/>
</wsse:SecurityTokenReference>
<wsc:Nonce>KJHFRE...</wsc:Nonce>
</wsc:DerivedKeyToken>
<xenc:ReferenceList>
...
<ds:KeyInfo>
<wsse:SecurityTokenReference>
<wsse:Reference URI="#dk2"/>
</wsse:SecurityTokenReference>
</ds:KeyInfo>
...
</xenc:ReferenceList>
<wsc:SecurityContextToken wsu:Id="ctx1">
<wsc:Identifier>uuid:...UUID1...</wsc:Identifier>
</wsc:SecurityContextToken>
<wsc:DerivedKeyToken wsu:Id="dk1">
<wsse:SecurityTokenReference>
<wsse:Reference URI="#ctx1"/>
</wsse:SecurityTokenReference>
<wsc:Nonce>KJHFRE...</wsc:Nonce>
</wsc:DerivedKeyToken>
<xenc:ReferenceList>
...
<ds:KeyInfo>
<wsse:SecurityTokenReference>
<wsse:Reference URI="#dk1"/>
</wsse:SecurityTokenReference>
</ds:KeyInfo>
...
</xenc:ReferenceList>
</wsse:Security>
...
</S11:Header>
<S11:Body>
...
</S11:Body>
</S11:Envelope>
The following
illustrates the syntax for a derived key based on the 3rd generation of the
shared key identified in the specified security context:
<wsc:DerivedKeyToken xmlns:wsc="..." xmlns:wsse="...">
<wsse:SecurityTokenReference>
<wsse:Reference URI="#ctx1"/>
</wsse:SecurityTokenReference>
<wsc:Generation>2</wsc:Generation>
</wsc:DerivedKeyToken>
The following
illustrates the syntax for a derived key based on the 1st generation of a key
derived from an existing derived key (4th generation):
<wsc:DerivedKeyToken xmlns:wsc="...">
<wsc:Properties>
<wsc:Name>.../derivedKeySource</wsc:Name>
<wsc:Label>NewLabel</wsc:Label>
<wsc:Nonce>FHFE...</wsc:Nonce>
</wsc:Properties>
<wsc:Generation>3</wsc:Generation>
</wsc:DerivedKeyToken>
<wsc:DerivedKeyToken wsu:Id="newKey" xmlns:wsc="..." xmlns:wsse="..."
>
<wsse:SecurityTokenReference>
<wsse:Reference URI=".../derivedKeySource"/>
</wsse:SecurityTokenReference>
<wsc:Generation>0</wsc:Generation>
</wsc:DerivedKeyToken>
In the example
above we have named a derived key so that other keys can be derived from it.
To do this we use the <wsc:Properties> element name tag to assign a global
name attribute. Note that in this example, the ID attribute could have been
used to name the base derived key if we didn't want it to be a globally named
resource. We have also included the <wsc:Label> and <wsc:Nonce> elements as metadata properties
indicating how to derive sequences of this derivation.
This specification
also defines a shortcut mechanism for referencing certain types of derived
keys. Specifically, a @wsc:Nonce attribute can also
be added to the security token reference (STR) defined in the [WS-Security] specification. When present, it indicates
that the key is not in the referenced token, but is a key derived from the
referenced token's key/secret. The @wsc:Length attribute can be used in
conjunction with @wsc:Nonce in the security token reference (STR) to indicate
the length of the derived key. The value of this attribute is an unsigned long
value indicating the size of the key in bytes. If this attribute isn't
specified, the default derived key length value is 32.
Consequently, the
following two illustrations are functionally equivalent:
<wsse:Security xmlns:wsc="..." xmlns:wsse="..." xmlns:xx="..."
xmlns:ds="..." xmlns:wsu="...">
<xx:MyToken wsu:Id="base">...</xx:MyToken>
<wsc:DerivedKeyToken wsu:Id="newKey">
<wsse:SecurityTokenReference>
<wsse:Reference URI="#base"/>
</wsse:SecurityTokenReference>
<wsc:Nonce>...</wsc:Nonce>
</wsc:DerivedKeyToken>
<ds:Signature>
...
<ds:KeyInfo>
<wsse:SecurityTokenReference>
<wsse:Reference URI="#newKey"/>
</wsse:SecurityTokenReference>
</ds:KeyInfo>
</ds:Signature>
</wsse:Security>
This is
functionally equivalent to the following:
<wsse:Security xmlns:wsc="..." xmlns:wsse="..." xmlns:xx="..."
xmlns:ds="..." xmlns:wsu="...">
<xx:MyToken wsu:Id="base">...</xx:MyToken>
<ds:Signature>
...
<ds:KeyInfo>
<wsse:SecurityTokenReference wsc:Nonce="...">
<wsse:Reference URI="#base"/>
</wsse:SecurityTokenReference>
</ds:KeyInfo>
</ds:Signature>
</wsse:Security>
For a variety of
reasons it may be necessary to reference a Security Context Token. These
references can be broken into two general categories: references from within
the <wsse:Security> element, generally used to indicate
the key used in a signature or encryption operation and references from other
parts of the SOAP envelope, for example to specify a token to be used in some
particular way. References within the <wsse:Security>
element can further be
divided into reference to an SCT found within the message and references to a
SCT not present in the message.
The Security Context Token does not support references to it
using key identifiers or key names. All references MUST either use an ID (to a
wsu:Id attribute) or a <wsse:Reference> to the <wsc:Identifier> element.
References using an ID are message-specific. References
using the <wsc:Identifier>
element value are message independent.
If the SCT is referenced from within the <wsse:Security> element or from
an RST or RSTR, it is RECOMMENDED that these references be message independent,
but these references MAY be message-specific. A reference from the RST/RSTR is
treated differently than other references from the SOAP Body as the RST/RSTR is
exclusively dealing with security related information similar to the
<wsse:Security> element.
When an SCT located in the <wsse:Security>
element is referenced from outside the <wsse:Security>
element, a message independent referencing mechanisms MUST be used, to enable a
cleanly layered processing model unless there is a prior agreement between the
involved parties to use message-specific referencing mechanism.
When an SCT is referenced from within the <wsse:Security> element, but the
SCT is not present in the message, (presumably because it was transmitted in a
previous message) a message independent referencing mechanism MUST be used.
The following
example illustrates associating a specific security context with an action.
<S11:Envelope
xmlns:S11="..." xmlns:wsse="..." xmlns:wsu="..."
xmlns:wsc="...">
<S11:Header>
...
<wsse:Security>
<wsc:SecurityContextToken wsu:Id="sct1">
<wsc:Identifier>uuid:...UUID1...</wsc:Identifier>
</wsc:SecurityContextToken>
<ds:Signature xmlns:ds="...">
...signature over body and crucial headers using #sct1...
</ds:Signature>
<wsc:SecurityContextToken wsu:Id="sct2">
<wsc:Identifier>uuid:...UUID2...</wsc:Identifier>
</wsc:SecurityContextToken>
<ds:Signature xmlns:ds="...">
...signature over body and crucial headers using #sct2...
</ds:Signature>
</wsse:Security>
...
</S11:Header>
<S11:Body wsu:Id="req">
<xx:Custom xmlns:xx="http://example.com/custom"
xmlns:wsse="...">
...
<wsse:SecurityTokenReference>
<wsse:Reference URI="uuid:...UUID2..."/>
</wsse:SecurityTokenReference>
</xx:Custom>
</S11:Body>
</S11:Envelope>
There are many
circumstances where an error can occur while
processing security information. Errors use the SOAP Fault mechanism. Note
that the reason text provided below is RECOMMENDED, but alternative text MAY be
provided if more descriptive or preferred by the implementation. The tables
below are defined in terms of SOAP 1.1. For SOAP 1.2, the Fault/Code/Value is
env:Sender (as defined in SOAP 1.2) and the Fault/Code/Subcode/Value is the faultcode below and the Fault/Reason/Text is the faultstring below. It should be noted that profiles MAY
provide second-level details fields, but they should be careful not to
introduce security vulnerabilities when doing so (e.g. by providing too
detailed information).
|
Error that
occurred (faultstring)
|
Fault code
(faultcode)
|
|
The requested context elements are insufficient or
unsupported.
|
wsc:BadContextToken
|
|
Not all of the values associated with the SCT are
supported.
|
wsc:UnsupportedContextToken
|
|
The specified source for the derivation is unknown.
|
wsc:UnknownDerivationSource
|
|
The provided context token has
expired
|
wsc:RenewNeeded
|
|
The specified context token could not be renewed.
|
wsc:UnableToRenew
|
As stated in the
Goals section of this document, this specification is meant to provide
extensible framework and flexible syntax, with which one could implement
various security mechanisms. This framework and syntax by itself does not provide any guarantee of security. When
implementing and using this framework and syntax, one must make every effort to
ensure that the result is not vulnerable to any one of a wide range of attacks.
It is not feasible
to provide a comprehensive list of security considerations for such an
extensible set of mechanisms. A complete security analysis must be conducted on
specific solutions based on this specification. Below we illustrate some of the
security concerns that often come up with protocols of this type, but we stress
that this is not an exhaustive list of concerns.
It is critical
that all relevant elements of a message be included in signatures. As well,
the signatures for security context establishment must include a timestamp,
nonce, or sequence number depending on the degree of replay prevention
required. Security context establishment should include full policies to
prevent possible attacks (e.g. downgrading attacks).
Authenticating
services are susceptible to denial of service attacks. Care should be taken to
mitigate such attacks as is warranted by the service.
There are many
other security concerns that one may need to consider in security protocols.
The list above should not be used as a "check list" instead of a
comprehensive security analysis.
In addition to the
consideration identified here, readers should also review the security
considerations in [WS-Security] and [WS-Trust].
An implementation
conforms to this specification if it satisfies all of the MUST or REQUIRED
level requirements defined within this specification. A SOAP Node MUST NOT use
the XML namespace identifier for this specification (listed in Section 1.3)
within SOAP Envelopes unless it is compliant with this specification.
This specification
references a number of other specifications (see the table above). In order to
comply with this specification, an implementation MUST implement the portions
of referenced specifications necessary to comply with the required provisions
of this specification. Additionally, the implementation of the portions of the
referenced specifications that are specifically cited in this specification
MUST comply with the rules for those portions as established in the referenced
specification.
Additionally
normative text within this specification takes precedence over normative
outlines (as described in section 1.5.1), which in turn take precedence over
the XML Schema [XML Schema Part 1, Part 2] and WSDL [WSDL 1.1] descriptions.
That is, the normative text in this specification further constrains the
schemas and/or WSDL that are part of this specification; and this specification
contains further constraints on the elements defined in referenced schemas.
Compliant services
are NOT REQUIRED to implement everything defined in this specification.
However, if a service implements an aspect of the specification, it MUST comply
with the requirements specified (e.g. related "MUST" statements). If
an OPTIONAL message is not supported, then the implementation SHOULD Fault just
as it would for any other unrecognized/unsupported message. If an OPTIONAL
message is supported, then the implementation MUST satisfy all of the MUST and
REQUIRED sections of the message.
This non-normative
appendix illustrates several sample usage patterns of [WS-Trust]
and this document. Specifically, it illustrates different patterns that could
be used to parallel, at an end-to-end message level, the selected TLS/SSL
scenarios. This is not intended to be the definitive method for the scenarios,
nor is it fully inclusive. Its purpose is simply to illustrate, in a context
familiar to readers, how this specification might be used.
The following
sections are based on a scenario where the client wishes to authenticate the
server prior to sharing any of its own credentials.
It should be noted
that the following sample usages are illustrative; any implementation of the
examples illustrated below should be carefully reviewed for potential security
attacks. For example, multi-leg exchanges such as those below should be
careful to prevent man-in-the-middle attacks or downgrade attacks. It may be
desirable to use running hashes as challenges that are signed or a similar
mechanism to ensure continuity of the exchange.
The examples below
assume that both parties understand the appropriate security policies in use
and can correctly construct signatures and encryption that the other party can
process.
A.1 Anonymous SCT
In this scenario
the requestor wishes to remain anonymous while authenticating the recipient and
establishing an SCT for secure communication.
This scenario
assumes that the requestor has a key for the recipient. If this isn't the
case, they can use [WS-MEX] or the mechanisms described in
a later section or obtain one from another security token service.
There are two
basic patterns that can apply, which only vary slightly. The first is as
follows:
1. The requestor sends an RST to the
recipient requesting an SCT. The request contains key material encrypted for
the recipient. The request is not authenticated.
2. The recipient, if it accepts such
requests, returns an RSTRC with one or more RSTRs with the SCT as the requested
token and does not return any proof information indicating that the requestor's
key is the proof.
A slight variation
on this is as follows:
1. The requestor sends an RST to the
recipient requesting an SCT. The request contains key material encrypted for
the recipient. The request is not authenticated.
2. The recipient, if it accepts such
requests, returns an RSTRC with one or more RSTR and with the SCT as the
requested token and returns its own key material encrypted using the
requestor's key.
Another slight
variation is to return a new key encrypted using the requestor's provided key.
It should be noted
that the variations that involve encrypting data using the requestor's key
material might be subject to certain types of key attacks.
Yet another
approach is to establish a secure channel (e.g. TLS/SSL IP/Sec) between the
requestor and the recipient. Key material can then safely flow in either
direction. In some circumstances, this provides greater protection than the
approach above when returning key information to the requestor.
A.2 Mutual Authentication SCT
In this scenario
the requestor is willing to authenticate, but wants the recipient to
authenticate first. The following steps outline the message flow:
1. The requestor sends an RST requesting
an SCT. The request contains key material encrypted for the recipient. The
request is not authenticated.
2. The recipient returns an RSTRC with
one or more RSTRs including a challenge for the requestor. The RSTRC is
secured by the recipient so that the requestor can authenticate it.
3. The requestor, after authenticating
the recipient's RSTRC, sends an RSTRC responding to the challenge.
4. The recipient, after authenticating
the requestor's RSTRC, sends a secured RSTRC containing the token and either
proof information or partial key material (depending on whether or not the
requestor provided key material).
Another variation
exists where step 1 includes a specific challenge for the service. Depending
on the type of challenge used this may not be necessary because the message may
contain enough entropy to ensure a fresh response from the recipient.
In other
variations the requestor doesn't include key information until step 3 so that
it can first verify the signature of the recipient in step 2.
If the recipient's
security token is not known, the RST/RSTR mechanism can still be used. The
following example illustrates one possible sequence of messages:
1. The requestor sends an RST requesting
an SCT. This request does not contain any key material, nor is the request
authenticated.
2. The recipient sends an RSTRC with one
or more RSTRs to the requestor with an embedded challenge. The RSTRC is
secured by the recipient so that the requestor can authenticate it.
3. The requestor sends an RSTRC to the
recipient and includes key information protected for the recipient. This
request may or may not be secured depending on whether or not the request is
anonymous.
4. The final issuance step depends on the
exact scenario. Any of the final legs from above might be used.
Note that step 1
might include a challenge for the recipient. Please refer to the comment in
the previous section on this scenario.
Also note that in
response to step 1 the recipient might issue a fault secured with [WS-Security] providing the requestor with information
about the recipient's security token.
The following individuals have participated in the creation
of this specification and are gratefully acknowledged:
Original Authors of the
initial contribution:
Steve
Anderson, OpenNetwork
Jeff Bohren, OpenNetwork
Toufic Boubez, Layer 7
Marc Chanliau, Computer Associates
Giovanni Della-Libera, Microsoft
Brendan Dixon, Microsoft
Praerit Garg, Microsoft
Martin Gudgin (Editor), Microsoft
Satoshi Hada, IBM
Phillip Hallam-Baker, VeriSign
Maryann Hondo, IBM
Chris Kaler, Microsoft
Hal Lockhart, BEA
Robin Martherus, Oblix
Hiroshi Maruyama, IBM
Anthony Nadalin (Editor), IBM
Nataraj Nagaratnam, IBM
Andrew Nash, Reactivity
Rob Philpott, RSA Security
Darren Platt, Ping Identity
Hemma Prafullchandra, VeriSign
Maneesh Sahu, Actional
John Shewchuk, Microsoft
Dan Simon, Microsoft
Davanum Srinivas, Computer Associates
Elliot Waingold, Microsoft
David Waite, Ping Identity
Doug Walter, Microsoft
Riaz Zolfonoon, RSA Security
Original Acknoledgements
of the initial contribution:
Paula Austel, IBM
Keith Ballinger, Microsoft
John Brezak, Microsoft
Tony Cowan, IBM
HongMei Ge, Microsoft
Slava Kavsan, RSA Security
Scott Konersmann, Microsoft
Leo Laferriere, Computer Associates
Paul Leach, Microsoft
Richard Levinson, Computer Associates
John Linn, RSA Security
Michael McIntosh, IBM
Steve Millet, Microsoft
Birgit Pfitzmann, IBM
Fumiko Satoh, IBM
Keith Stobie, Microsoft
T.R. Vishwanath, Microsoft
Richard Ward, Microsoft
Hervey Wilson, Microsoft
TC Members during the development of this specification:
Don Adams, Tibco Software Inc.
Jan Alexander, Microsoft Corporation
Steve Anderson, BMC Software
Donal Arundel, IONA Technologies
Howard Bae, Oracle Corporation
Abbie Barbir, Nortel Networks Limited
Charlton Barreto, Adobe Systems
Mighael Botha, Software AG, Inc.
Toufic Boubez, Layer 7 Technologies Inc.
Norman Brickman, Mitre Corporation
Melissa Brumfield, Booz Allen Hamilton
Geoff Bullen, Microsoft Corporation
Lloyd Burch, Novell
Scott Cantor, Internet2
Greg Carpenter, Microsoft Corporation
Steve Carter, Novell
Ching-Yun (C.Y.) Chao, IBM
Martin Chapman, Oracle Corporation
Kate Cherry, Lockheed Martin
Henry (Hyenvui) Chung, IBM
Luc Clement, Systinet Corp.
Paul Cotton, Microsoft Corporation
Glen Daniels, Sonic Software Corp.
Peter Davis, Neustar, Inc.
Martijn de Boer, SAP AG
Duane DeCouteau, Veterans Health
Administration
Werner Dittmann, Siemens AG
Abdeslem DJAOUI, CCLRC-Rutherford Appleton Laboratory
Fred Dushin, IONA Technologies
Petr Dvorak, Systinet Corp.
Colleen Evans, Microsoft Corporation
Ruchith Fernando, WSO2
Mark Fussell, Microsoft Corporation
Vijay Gajjala, Microsoft Corporation
Marc Goodner, Microsoft Corporation
Hans Granqvist, VeriSign
Martin Gudgin, Microsoft Corporation
Tony Gullotta, SOA Software Inc.
Jiandong Guo, Sun Microsystems
Phillip Hallam-Baker, VeriSign
Patrick Harding, Ping Identity Corporation
Heather Hinton, IBM
Frederick Hirsch, Nokia Corporation
Jeff Hodges, Neustar, Inc.
Will Hopkins, BEA Systems, Inc.
Alex Hristov, Otecia Incorporated
John Hughes, PA Consulting
Diane Jordan, IBM
Venugopal K, Sun Microsystems
Chris Kaler, Microsoft Corporation
Dana Kaufman, Forum Systems, Inc.
Paul Knight, Nortel Networks Limited
Ramanathan Krishnamurthy, IONA Technologies
Christopher Kurt, Microsoft Corporation
Kelvin Lawrence, IBM
Hubert Le Van Gong, Sun Microsystems
Jong Lee, BEA Systems, Inc.
Rich Levinson, Oracle Corporation
Tommy Lindberg, Dajeil Ltd.
Mark Little, JBoss Inc.
Hal Lockhart, BEA Systems, Inc.
Mike Lyons, Layer 7 Technologies Inc.
Eve Maler, Sun Microsystems
Ashok Malhotra, Oracle Corporation
Anand Mani, CrimsonLogic Pte Ltd
Jonathan Marsh, Microsoft Corporation
Robin Martherus, Oracle Corporation
Miko Matsumura, Infravio, Inc.
Gary McAfee, IBM
Michael McIntosh, IBM
John Merrells, Sxip Networks SRL
Jeff Mischkinsky, Oracle Corporation
Prateek Mishra, Oracle Corporation
Bob Morgan, Internet2
Vamsi Motukuru, Oracle Corporation
Raajmohan Na, EDS
Anthony Nadalin, IBM
Andrew Nash, Reactivity, Inc.
Eric Newcomer, IONA Technologies
Duane Nickull, Adobe Systems
Toshihiro Nishimura, Fujitsu Limited
Rob Philpott, RSA Security
Denis Pilipchuk, BEA Systems, Inc.
Darren Platt, Ping Identity Corporation
Martin Raepple, SAP AG
Nick Ragouzis, Enosis Group LLC
Prakash Reddy, CA
Alain Regnier, Ricoh Company, Ltd.
Irving Reid, Hewlett-Packard
Bruce Rich, IBM
Tom Rutt, Fujitsu Limited
Maneesh Sahu, Actional Corporation
Frank Siebenlist, Argonne National Laboratory
Joe Smith, Apani Networks
Davanum Srinivas, WSO2
David Staggs, Veterans Health Administration
Yakov Sverdlov, CA
Gene Thurston, AmberPoint
Victor Valle, IBM
Asir Vedamuthu, Microsoft Corporation
Greg Whitehead, Hewlett-Packard
Ron Williams, IBM
Corinna Witt, BEA Systems, Inc.
Kyle Young, Microsoft Corporation
