Network Working Group J. Rosenberg Request for Comments: 5025 Cisco Category: Standards Track December 2007
Presence Authorization Rules
Status of This Memo
This document specifies an Internet standards track protocol for the Internet community, and requests discussion and suggestions for improvements. Please refer to the current edition of the "Internet Official Protocol Standards" (STD 1) for the standardization state and status of this protocol. Distribution of this memo is unlimited.
Abstract
Authorization is a key function in presence systems. Authorization policies, also known as authorization rules, specify what presence information can be given to which watchers, and when. This specification defines an Extensible Markup Language (XML) document format for expressing presence authorization rules. Such a document can be manipulated by clients using the XML Configuration Access Protocol (XCAP), although other techniques are permitted.
Table of Contents
1. Introduction ....................................................2 2. Terminology .....................................................3 3. Structure of Presence Authorization Documents ...................3 3.1. Conditions .................................................4 3.1.1. Identity ............................................4 3.1.1.1. Acceptable Forms of Authentication .........4 3.1.1.2. Computing a URI for the Watcher ............5 3.1.2. Sphere ..............................................6 3.2. Actions ....................................................7 3.2.1. Subscription Handling ...............................7 3.3. Transformations ............................................9 3.3.1. Providing Access to Data Component Elements .........9 3.3.1.1. Device Information .........................9 3.3.1.2. Person Information ........................10 3.3.1.3. Service Information .......................11 3.3.2. Providing Access to Presence Attributes ............12 3.3.2.1. Provide Activities ........................12 3.3.2.2. Provide Class .............................12 3.3.2.3. Provide DeviceID ..........................13 3.3.2.4. Provide Mood ..............................13 3.3.2.5. Provide Place-is ..........................13
Rosenberg Standards Track [Page 1]
RFC 5025 Presence Authorization December 2007
3.3.2.6. Provide Place-type ........................13 3.3.2.7. Provide Privacy ...........................13 3.3.2.8. Provide Relationship ......................14 3.3.2.9. Provide Sphere ............................14 3.3.2.10. Provide Status-Icon ......................14 3.3.2.11. Provide Time-Offset ......................14 3.3.2.12. Provide User-Input .......................14 3.3.2.13. Provide Note .............................15 3.3.2.14. Provide Unknown Attribute ................15 3.3.2.15. Provide All Attributes ...................16 4. When to Apply the Authorization Policies .......................17 5. Implementation Requirements ....................................17 6. Example Document ...............................................18 7. XML Schema .....................................................19 8. Schema Extensibility ...........................................21 9. XCAP Usage .....................................................22 9.1. Application Unique ID .....................................22 9.2. XML Schema ................................................22 9.3. Default Namespace .........................................22 9.4. MIME Type .................................................22 9.5. Validation Constraints ....................................22 9.6. Data Semantics ............................................22 9.7. Naming Conventions ........................................23 9.8. Resource Interdependencies ................................23 9.9. Authorization Policies ....................................23 10. Security Considerations .......................................23 11. IANA Considerations ...........................................24 11.1. XCAP Application Usage ID ................................24 11.2. URN Sub-Namespace Registration ...........................25 11.3. XML Schema Registrations .................................25 12. Acknowledgements ..............................................26 13. References ....................................................26 13.1. Normative References .....................................26 13.2. Informative References ...................................27
The Session Initiation Protocol (SIP) for Instant Messaging and Presence (SIMPLE) specifications allow a user, called a watcher, to subscribe to another user, called a presentity [17], in order to learn their presence information [18]. This subscription is handled by a presence agent. However, presence information is sensitive, and a presence agent needs authorization from the presentity prior to handing out presence information. As such, a presence authorization document format is needed. This specification defines a format for such a document, called a presence authorization document.
Rosenberg Standards Track [Page 2]
RFC 5025 Presence Authorization December 2007
[8] specifies a framework for representing authorization policies, and is applicable to systems such as geo-location and presence. This framework is used as the basis for presence authorization documents. In the framework, an authorization policy is a set of rules. Each rule contains conditions, actions, and transformations. The conditions specify under what conditions the rule is to be applied to presence server processing. The actions element tells the server what actions to take. The transformations element indicates how the presence data is to be manipulated before being presented to that watcher, and as such, defines a privacy filtering operation. [8] identifies a small number of specific conditions common to presence and location services, and leaves it to other specifications, such as this one, to fill in usage specific details.
A presence authorization document can be manipulated by clients using several means. One such mechanism is the XML Configuration Access Protocol (XCAP) [2]. This specification defines the details necessary for using XCAP to manage presence authorization documents.
In this document, the key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" are to be interpreted as described in RFC 2119 [1] and indicate requirement levels for compliant implementations.
A presence authorization document is an XML document, formatted according to the schema defined in [8]. Presence authorization documents inherit the MIME type of common policy documents, application/auth-policy+xml. As described in [8], this document is composed of rules that contain three parts - conditions, actions, and transformations. Each action or transformation, which is also called a permission, has the property of being a positive grant of information to the watcher. As a result, there is a well-defined mechanism for combining actions and transformations obtained from several sources. This mechanism is privacy safe, since the lack of any action or transformation can only result in less information being presented to a watcher.
This section defines the new conditions, actions, and transformations defined by this specification.
Although the <identity> element is defined in [8], that specification indicates that the specific usages of the framework document need to define details that are protocol and usage specific. In particular, it is necessary for a usage of the common policy framework to:
o Define acceptable means of authentication.
o Define the procedure for representing the identity of the WR (Watcher/Requestor) as a URI or Internationalized Resource Identifier (IRI) [13].
This sub-section defines those details for systems based on [18]. It does so in general terms, so that the recommendations defined here apply to existing and future authentication mechanisms in SIP.
When used with SIP, a request is considered authenticated if one of the following is true:
The watcher proves its identity to the server through a form of cryptographic authentication, including authentication based on a shared secret or a certificate, and that authentication yields an identity for the watcher.
The request comes from a sender that is asserting the identity of the watcher, and:
1. the assertion includes a claim that the asserting party used a form of cryptographic authentication (as defined above) to determine the identity of the watcher, and
3. the assertion provides an identity in the form of a URI.
Based on this definition, examples of valid authentication techniques include SIP [5], digest authentication [4], cryptographically verified identity assertions (RFC 4474 [15]), and identity assertions made in closed network environments (RFC 3325 [16]).
However, the anonymous authentication described on page 194 of RFC 3261 [5] is not considered a valid mechanism for authentication
Rosenberg Standards Track [Page 4]
RFC 5025 Presence Authorization December 2007
because it does not produce an identity for the watcher. However, an anonymous From header field, when used in conjunction with RFC 4474 [15], is considered an acceptable mechanism for authentication, since it still implies that the asserting node performed authentication that produced the identity of the watcher.
Computing the URI for the watcher depends on whether the identity is being ascertained through authentication or through an asserted identity.
If an identity assertion is being utilized, the asserted identity itself (which is in the form of a URI for acceptable forms of identity assertion) is utilized as the URI. If the identity assertion mechanism asserts multiple URIs for the watcher, then each of them is used for the comparisons outlined in [8], and if any of them match a <one> or <except> element, the watcher is considered a match.
If an identity is being determined directly by a cryptographic authentication, that authentication must produce a URI, or must produce some form of identifier that can be linked, through provisioning, to a URI that is bound to that identifier.
For example, in the case of SIP Digest authentication, the authentication process produces a username scoped within a realm. That username and realm are bound to an Address of Record (AOR) through provisioning, and the resulting AOR is used as the watcher URI. Consider the following "user record" in a database:
If the presence server receives a SUBSCRIBE request, challenges it with the realm set to "example.com", and the subsequent SUBSCRIBE contains an Authorization header field with a username of "ali" and a digest response generated with the password "f779ajvvh8a6s6", the identity used in matching operations is "sip:alice@example.com".
In SIP systems, it is possible for a user to have aliases - that is, there are multiple SIP AORs "assigned" to a single user. In terms of this specification, there is no relationship between those aliases. Each would look like a different user. This will be the consequence for systems where the watcher is in a different domain than the presentity. However, even if the watcher and presentity are in the
Rosenberg Standards Track [Page 5]
RFC 5025 Presence Authorization December 2007
same domain, and the presence server knows that there are aliases for the watcher, these aliases are not mapped to each other or used in any way.
SIP also allows for anonymous requests. If a request is anonymous because the watcher utilized an authentication mechanism that does not provide an identity to the presence server (such as the SIP digest "anonymous" username), the request is considered unauthenticated (as discussed above) and will match only an empty <identity> element. If a request is anonymous because it contains a Privacy header field [14], but still contains an asserted identity meeting the criteria defined above, that identity is utilized, and the fact that the request was anonymous has no impact on the identity processing.
It is important to note that SIP frequently uses both SIP URI and tel URI [12] as identifiers, and to make matters more confusing, a SIP URI can contain a phone number in its user part, in the same format used in a tel URI. A WR identity that is a SIP URI with a phone number will NOT match the <one> and <except> conditions whose 'id' is a tel URI with the same number. The same is true in the reverse. If the WR identity is a tel URI, this will not match a SIP URI in the <one> or <except> conditions whose user part is a phone number. URIs of different schemes are never equivalent.
The <sphere> element is defined in [8]. However, each application making use of the common policy specification needs to determine how the presence server computes the value of the <sphere> to be used in the evaluation of the condition.
To compute the value of <sphere>, the presence agent examines all published presence documents for the presentity. If at least one of them includes the <sphere> element [9] as part of the person data component [10], and all of those containing the element have the same value for it, which is the value used for the <sphere> in presence policy processing. If, however, the <sphere> element was not present in any of the published documents, or it was present but had inconsistent values, its value is considered undefined in terms of presence policy processing.
Care must be taken in using <sphere> as a condition for determining the subscription handling. Since the value of <sphere> changes dynamically, a state change can cause a subscription to be suddenly terminated. The watcher has no way to know, aside from polling, when their subscription would be reinstated as the value of <sphere>
Rosenberg Standards Track [Page 6]
RFC 5025 Presence Authorization December 2007
changes. For this reason, <sphere> is primarily useful for matching on rules that define transformations.
The <sub-handling> element specifies the subscription authorization decision that the server should make. It also specifies whether or not the presence document for the watcher should be constructed using "polite blocking". Usage of polite blocking and the subscription authorization decision are specified jointly since proper privacy handling requires a correlation between them. As discussed in [8], since the combination algorithm runs independently for each permission, if correlations exist between permissions, they must be merged into a single variable. That is what is done here. The <sub-handling> element is an enumerated Integer type. The defined values are:
block: This action tells the server to reject the subscription, placing it in the "terminated" state. It has the value of zero, and it represents the default value. No value of the <sub- handling> element can ever be lower than this. Strictly speaking, it is not necessary for a rule to include an explicit block action, since the default in the absence of any action will be block. However, it is included for completeness.
confirm: This action tells the server to place the subscription in the "pending" state, and await input from the presentity to determine how to proceed. It has a value of ten.
polite-block: This action tells the server to place the subscription into the "active" state, and to produce a presence document that indicates that the presentity is unavailable. A reasonable document would exclude device and person information elements, and include only a single service whose basic status is set to closed [3]. This action has a value of twenty.
allow: This action tells the server to place the subscription into the "active" state. This action has a value of thirty.
NOTE WELL: Placing a value of block for this element does not guarantee that a subscription is denied! If any matching rule has any other value for this element, the subscription will receive treatment based on the maximum of those other values. This is based on the combining rules defined in [8].
Rosenberg Standards Track [Page 7]
RFC 5025 Presence Authorization December 2007
Future specifications that wish to define new types of actions MUST define an entirely new action (separate from <sub-handling>), and define their own set of values for that action. A document could contain both <sub-handling> and a subscription handling action defined by a future specification; in that case, since each action is always a positive grant of information, the resulting action is the least restrictive one across both elements.
The exact behavior of a presence server upon a change in the sub- handling value can be described by utilizing the subscription processing state machine in Figure 1 of RFC 3857 [6].
If the <sub-handling> permission changes value to "block", this causes a "rejected" event to be generated into the subscription state machine for all affected subscriptions. This will cause the state machine to move into the "terminated" state, resulting in the transmission of a NOTIFY to the watcher with a Subscription-State header field with value "terminated" and a reason of "rejected" [7], which terminates their subscription. If a new subscription arrives later on, and the value of <sub-handling> that applies to that subscription is "block", the subscription processing follows the "subscribe, policy=reject" branch from the "init" state, and a 403 response to the SUBSCRIBE is generated.
If the <sub-handling> permission changes value to "confirm", the processing depends on the states of the affected subscriptions. Unfortunately, the state machine in RFC 3857 does not define an event corresponding to an authorization decision of "pending". If the subscription is in the "active" state, it moves back into the "pending" state. This causes a NOTIFY to be sent, updating the Subscription-State [7] to "pending". No reason is included in the Subscription-State header field (none are defined to handle this case). No further documents are sent to this watcher. There is no change in state if the subscription is in the "pending", "waiting", or "terminated" states. If a new subscription arrives later on, and the value of <sub-handling> that applies to that subscription is "confirm", the subscription processing follows the "subscribe, no policy" branch from the "init" state, and a 202 response to the SUBSCRIBE is generated, followed by a NOTIFY with Subscription-State of "pending". No presence document is included in that NOTIFY.
If the <sub-handling> permission changes value from "blocked" or "confirm" to "polite-block" or "allow", this causes an "approved" event to be generated into the state machine for all affected subscriptions. If the subscription was in the "pending" state, the state machine will move to the "active" state, resulting in the transmission of a NOTIFY with a Subscription-State header field of "active", and the inclusion of a presence document in that NOTIFY.
Rosenberg Standards Track [Page 8]
RFC 5025 Presence Authorization December 2007
If the subscription was in the "waiting" state, it will move into the "terminated" state. If a new subscription arrives later on, and the value of <sub-handling> that applies to that subscription is "polite-block" or "allow", the subscription processing follows the "subscribe, policy=accept" branch from the "init" state, and a 200 OK response to the SUBSCRIBE is generated, followed by a NOTIFY with Subscription-State of "active" with a presence document in the body of the NOTIFY.
The transformations defined here are used to drive the behavior of the privacy filtering operation. Each transformation defines the visibility a watcher is granted to a particular component of the presence document. One group of transformations grants visibility to person, device, and service data elements based on identifying information for those elements. Another group of transformations provides access to particular data elements in the presence document.
3.3.1. Providing Access to Data Component Elements
The transformations in this section provide access to person, device, and service data component elements. Once access has been granted to such an element, access to specific presence attributes for that element is controlled by the permissions defined in Section 3.3.2.
The <provide-devices> permission allows a watcher to see <device> information present in the presence document. It is a set variable. Each member of the set provides a way to identify a device or group of devices. This specification defines three types of elements in the set - <class>, which identifies a device occurrence by class; <deviceID>, which identifies a device occurrence by device ID; and <occurrence-id>, which identifies a device occurrence by occurrence ID. The device ID and occurrence ID are defined in [10]. Each member of the set is identified by its type (class, deviceID, or occurrence-id) and value (value of the class, value of the deviceID, or value of the occurrence-id).
For example, consider the following <provide-devices> element:
The <provide-devices> element can also take on the special value <all-devices>, which is a short-hand notation for all device occurrences present in the presence document.
Permission is granted to see a particular device occurrence if one of the device identifiers in the set identifies that device occurrence. If a <class> permission is granted to the watcher, and the <class> of the device occurrence matches the value of the <class> permission based on case-sensitive equality, the device occurrence is included in the presence document. If a <deviceID> permission is granted to the watcher, and the <deviceID> of the device occurrence matches the value of the <deviceID> permission based on URI equivalence, the device occurrence is included in the presence document. If an <occurrence-id> permission is granted to the watcher, and the <occurrence-id> of the device occurrence matches the value of the <occurrence-id> permission based on case-sensitive equality, the device occurrence is included in the presence document. In addition, a device occurrence is included in the presence document if the <all-devices> permission was granted to the watcher.
The <provide-persons> permission allows a watcher to see the <person> information present in the presence document. It is a set variable. Each member of the set provides a way to identify a person occurrence. This specification defines two types of elements in the set - <class>, which identifies a person occurrence by class, and <occurrence-id>, which identifies an occurrence by its occurrence ID. Each member of the set is identified by its type (class or occurrence-id) and value (value of the class or value of the occurrence-id). The <provide-persons> element can also take on the
Rosenberg Standards Track [Page 10]
RFC 5025 Presence Authorization December 2007
special value <all-persons>, which is a short-hand notation for all person occurrences present in the presence document. The set combination is done identically to the <provide-devices> element.
Permission is granted to see a particular person occurrence if one of the person identifiers in the set identifies that person occurrence. If a <class> permission is granted to the watcher, and the <class> of the person occurrence matches the value of the <class> permission based on case-sensitive equality, the person occurrence is included in the presence document. If an <occurrence-id> permission is granted to the watcher, and the <occurrence-id> of the person occurrence matches the value of the <occurrence-id> permission based on case-sensitive equality, the person occurrence is included in the presence document. In addition, a person occurrence is included in the presence document if the <all-persons> permission was granted to the watcher.
The <provide-services> permission allows a watcher to see service information present in <tuple> elements in the presence document. Like <provide-devices>, it is a set variable. Each member of the set provides a way to identify a service occurrence. This specification defines four types of elements in the set - <class>, which identifies a service occurrence by class; <occurrence-id>, which identifies a service by its occurrence ID; <service-uri>, which identifies a service by its service URI; and <service-uri-scheme>, which identifies a service by its service URI scheme. Each member of the set is identified by its type (class, occurrence-id, service-uri, or service-uri-scheme) and value (value of the class, value of the occurrence-id, value of the service-uri, or value of the service- uri-scheme). The <provide-services> element can also take on the special value <all-services>, which is a short-hand notation for all service occurrences present in the presence document. The set combination is done identically to the <provide-persons> element.
Permission is granted to see a particular service occurrence if one of the service identifiers in the set identifies that service occurrence. If a <class> permission is granted to the watcher, and the <class> of the service occurrence matches the value of the <class> permission based on case-sensitive equality, the service occurrence is included in the presence document. If a <service-uri> permission is granted to the watcher, and the <service-uri> of the service occurrence matches the value of the <service-uri> permission based on URI equivalence, the service occurrence is included in the presence document. If an <occurrence-id> permission is granted to the watcher, and the <occurrence-id> of the service occurrence matches the value of the <occurrence-id> permission based on case-
Rosenberg Standards Track [Page 11]
RFC 5025 Presence Authorization December 2007
sensitive equality, the service occurrence is included in the presence document. If a <service-uri-scheme> permission is granted to the watcher, and the scheme of the service URI for the service occurrence matches the value of <service-uri-scheme> based on case- sensitive equality, the service occurrence is included in the presence document. In addition, a service occurrence is included in the presence document if the <all-services> permission was granted to the watcher.
The permissions of Section 3.3.1 provide coarse-grained access to presence data by allowing or blocking specific services or devices, and allowing or blocking person information.
Once person, device, or service information is included in the document, the permissions in this section define which presence attributes are reported there. Certain information is always reported. In particular, the <contact>, <service-class> [9], <basic> status, and <timestamp> elements in all <tuple> elements, if present, are provided to watchers. The <timestamp> element in all <person> elements, if present, is provided to watchers. The <timestamp> and <deviceID> elements in all <device> elements, if present, are provided to all watchers.
This permission controls access to the <activities> element defined in [9]. The name of the element providing this permission is <provide-activities>, and it is a Boolean type. If its value is TRUE, then the <activities> element in the person data element is reported to the watcher. If FALSE, this presence attribute is removed if present.
This permission controls access to the <class> element defined in [9]. The name of the element providing this permission is <provide- class>, and it is a Boolean type. If its value is TRUE, then any <class> element in a person, service, or device data element is reported to the watcher. If FALSE, this presence attribute is removed if present.
This permission controls access to the <deviceID> element in a <tuple> element, as defined in [9]. The name of the element providing this permission is <provide-deviceID>, and it is a Boolean type. If its value is TRUE, then the <deviceID> element in the service data element is reported to the watcher. If FALSE, this presence attribute is removed if present. Note that the <deviceID> in a device data element is always included, and not controlled by this permission.
This permission controls access to the <mood> element defined in [9]. The name of the element providing this permission is <provide-mood>, and it is a Boolean type. If its value is TRUE, then the <mood> element in the person data element is reported to the watcher. If FALSE, this presence attribute is removed if present.
This permission controls access to the <place-is> element defined in [9]. The name of the element providing this permission is <provide- place-is>, and it is a Boolean type. If its value is TRUE, then the <place-is> element in the person data element is reported to the watcher. If FALSE, this presence attribute is removed if present.
This permission controls access to the <place-type> element defined in [9]. The name of the element providing this permission is <provide-place-type>, and it is a Boolean type. If its value is TRUE, then the <place-type> element in the person data element is reported to the watcher. If FALSE, this presence attribute is removed if present.
This permission controls access to the <privacy> element defined in [9]. The name of the element providing this permission is <provide- privacy>, and it is a Boolean type. If its value is TRUE, then the <privacy> element in the person or service data element is reported to the watcher. If FALSE, this presence attribute is removed if present.
This permission controls access to the <relationship> element defined in [9]. The name of the element providing this permission is <provide-relationship>, and it is a Boolean type. If its value is TRUE, then the <relationship> element in the service data element is reported to the watcher. If FALSE, this presence attribute is removed if present.
This permission controls access to the <sphere> element defined in [9]. The name of the element providing this permission is <provide- sphere>, and it is a Boolean type. If its value is TRUE, then the <sphere> element in the person data element is reported to the watcher. If FALSE, this presence attribute is removed if present.
This permission controls access to the <status-icon> element defined in [9]. The name of the element providing this permission is <provide-status-icon>, and it is a Boolean type. If its value is TRUE, then any <status-icon> element in the person or service data element is reported to the watcher. If FALSE, this presence attribute is removed if present.
This permission controls access to the <time-offset> element defined in [9]. The name of the element providing this permission is <provide-time-offset>, and it is a Boolean type. If its value is TRUE, then the <time-offset> element in the person data element is reported to the watcher. If FALSE, this presence attribute is removed if present.
This permission controls access to the <user-input> element defined in [9]. The name of the element providing this permission is <provide-user-input>, and it is an enumerated integer type. Its value defines what information is provided to watchers in person, device, or service data elements:
false: This value indicates that the <user-input> element is removed from the document. It is assigned the numeric value of 0.
Rosenberg Standards Track [Page 14]
RFC 5025 Presence Authorization December 2007
bare: This value indicates that the <user-input> element is to be retained. However, any "idle-threshold" and "since" attributes are to be removed. This value is assigned the numeric value of 10.
thresholds: This value indicates that the <user-input> element is to be retained. However, only the "idle-threshold" attribute is to be retained. This value is assigned the numeric value of 20.
full: This value indicates that the <user-input> element is to be retained, including any attributes. This value is assigned the numeric value of 30.
This permission controls access to the <note> element defined in [3] for <tuple> and [10] for <person> and <device>. The name of the element providing this permission is <provide-note>, and it is a Boolean type. If its value is TRUE, then any <note> elements in the person, service, or device data elements are reported to the watcher. If FALSE, this presence attribute is removed if present.
This permission has no bearing on any <note> values present within <activities>, <mood>, <place-is>, <place-type>, <privacy>, <relationship>, or <service-class> elements. Notes within these elements are essentially values for their respective elements, and are present if the respective element is permitted in the presence document. For example, if an <activities> element is present in a presence document, and there is a <note> value for it, that note is present in the document sent to the watcher if the <provide- activities> permission is given, regardless of whether the <provide- note> permission is given.
It is important that systems be allowed to include proprietary or new presence information and that users be able to set permissions for that information, without requiring an upgrade of the presence server and authorization system. For this reason, the <provide-unknown- attribute> permission is defined. This permission indicates that the unknown presence attribute with the given name and namespace (supplied as mandatory attributes of the <provide-unknown-attribute> element) should be included in the document. Its type is Boolean.
The value of the name attribute MUST be an unqualified element name (meaning that a namespace prefix MUST NOT be included), and the value of the ns attribute MUST be a namespace URI. The two are combined to form a qualified element name, which will be matched to all unknown
Rosenberg Standards Track [Page 15]
RFC 5025 Presence Authorization December 2007
child elements of the Presence Information Data Format (PIDF) <tuple>, <device>, or <person> elements with the same qualified name. In this context, "unknown" means that the presence server is not aware of any schemas that define authorization policies for that element. By definition, this will exclude the <provide-unknown- attribute> rule from being applied to any of the presence status extensions defined by RPID, since authorization policies for those are defined here.
Another consequence of this definition is that the interpretation of the <provide-unknown-attribute> element can change should the presence server be upgraded. For example, consider a server that, prior to the upgrade, had an authorization document that used <provide-unknown-attribute> with a value of TRUE for some attribute, say foo. This attribute was from a namespace and schema unknown to the server, and so the attribute was provided to watchers. However, after upgrade, the server is now aware of a new namespace and schema for a permission that grants access to the foo attribute. Now, the <provide-unknown-attribute> permission for the foo attribute will be ignored, resulting in a removal of those elements from presence documents sent to watchers. The system remains privacy safe, but behavior might not be as expected. Developers of systems that allow clients to set policies are advised to check the capabilities of the server (using the mechanism described in Section 8) before uploading a new authorization document, to make sure that the behavior will be as expected.
This permission grants access to all presence attributes in all of the person, device, and tuple elements that are present in the document (the ones present in the document are determined by the <provide-persons>, <provide-devices>, and <provide-services> permissions). It is effectively a macro that expands into a set of provide-activities, provide-class, provide-deviceID, provide-mood, provide-place-is, provide-place-type, provide-privacy, provide- relationship, provide-sphere, provide-status-icon, provide-time- offset, provide-user-input, provide-note, and provide-unknown- attribute permissions such that each presence attribute in the document has a permission for it. This implies that, so long as an entire person, service, or device occurrence is provided, every single presence attribute, including ones not known to the server and/or defined in future presence document extensions, is granted to the watcher.
This specification does not mandate at what point in the processing of presence data the privacy filtering aspects of the authorization policy are applied. However, they must be applied such that the final presence document sent to the watcher is compliant to the privacy policy described in the authorization documents that apply to the user (there can be more than one; the rules for combining them are described in [8]). More concretely, if the presence document sent to a watcher is D, and the privacy filtering operation applied do a presence document x is F(x), then D MUST have the property that D = F(D). In other words, further applications of the privacy filtering operation would not result in any further changes of the presence document, making further application of the filtering operation a no-op. A corollary of this is that F(F(D)) = D for all D.
The subscription processing aspects of the document get applied by the server when it decides to accept or reject the subscription.
The rules defined by the document in this specification form a "contract" of sorts between a client that creates this document and the server that executes the policies it contains. Consequently, presence servers implementing this specification MUST support all of the conditions, actions, and transformations defined in this specification. If servers were to implement a subset of these, clients would need a mechanism to discover which subset is supported. No such mechanism is defined.
It is RECOMMENDED that clients support all of the actions, transformations, and conditions defined in this specification. If a client supports a subset, it is possible that a user might manipulate their authorization rules from a different client, supporting a different subset, and store those results on the server. When the user goes back to the first client and views their presence authorization rules there, the client may not be able to properly render or manipulate the document retrieved from the server, since it may contain conditions, actions, or transformations not supported by the client. The only reason that this normative requirement is not a MUST is that there are valid conditions in which a user manipulates their presence authorization rules from a single client, in which case this problem does not occur.
This specification makes no normative recommendations on the mechanism used to transport presence authorization documents from
Rosenberg Standards Track [Page 17]
RFC 5025 Presence Authorization December 2007
clients to their servers. Although Section 9 defines how to utilize XCAP, XCAP is not normatively required by this specification.
The following presence authorization document specifies permissions for the user "user@example.com". The watcher is allowed to access presence information (the 'allow' value for <sub-handling>). They will be granted access to the presence data of all services whose contact URI schemes are sip and mailto. Person information is also provided. However, since there is no <provide-devices>, no device information will be given to the watcher. Within the service and person information provided to the watcher, the <activities> element will be shown, as will the <user-input> element. However, any "idle-threshold" and "since" attributes in the <user-input> element will be removed. Finally, the presence attribute <foo> will be shown to the watcher. Any other presence attributes will be removed.
It is anticipated that future changes to this specification are accomplished through extensions that define new types of permissions. These extensions MUST exist within a different namespace. Furthermore, the schema defined above and the namespace for elements defined within it MUST NOT be altered by future specifications. Changes in the basic schema, or in the interpretation of elements within that schema, may result in violations of user privacy due to misinterpretation of documents.
When extensions are made to the set of permissions, it becomes necessary for the agent constructing the permission document (typically a SIP user agent, though not necessarily) to know which permissions are supported by the server. This allows the agent to know how to build a document that results in the desired behavior, since unknown permissions would be ignored by the server. To handle this, when presence authorization documents are transported using
Rosenberg Standards Track [Page 21]
RFC 5025 Presence Authorization December 2007
XCAP, the XCAP capabilities document stored at the server SHOULD contain the namespaces for the permissions supported by the presence server. This way, an agent can query for this list prior to constructing a document.
XCAP requires application usages to define a unique application usage ID (AUID) in either the IETF tree or a vendor tree. This specification defines the "pres-rules" AUID within the IETF tree, via the IANA registration in Section 11.
XCAP requires application usages to define the MIME type for documents they carry. Presence authorization documents inherit the MIME type of common policy documents, application/auth-policy+xml.
When a presence agent receives a subscription for some user foo within a domain, it will look for all documents within http://[xcap root]/pres-rules/users/foo, and use all documents found beneath that point to guide authorization policy. If only a single document is used, it SHOULD be called "index".
This application usage does not modify the default XCAP authorization policy, which is that only a user can read, write, or modify their own documents. A server can allow privileged users to modify documents that they don't own, but the establishment and indication of such policies are outside the scope of this document.
Presence authorization policies contain very sensitive information. They indicate which other users are "liked" or "disliked" by a user. As such, when these documents are transported over a network, they SHOULD be encrypted.
Modification of these documents by an attacker can disrupt the service seen by a user, often in subtle ways. As a result, when these documents are transported, the transport SHOULD provide authenticity and message integrity.
In the case where XCAP is used to transfer the document, both clients and servers MUST implement HTTP over Transport Layer Security (TLS) and HTTP Digest authentication. Sites SHOULD authenticate clients using digest authentication over TLS, and sites SHOULD define the root services URI as an https URI.
Authorization documents themselves exist for the purposes of providing a security function - privacy. The SIP presence specifications [18] require the usage of an authorization function prior to the granting of presence information, and this specification meets that need. Presence authorization documents inherit the privacy properties of the common policy format on which they are based. This format has been designed to be privacy-safe, which means that failure of the presence server to obtain or understand an authorization document can never reveal more information than is
Rosenberg Standards Track [Page 23]
RFC 5025 Presence Authorization December 2007
desired about the user, only less. This is a consequence of the fact that all permissions are positive grants of information, and not negative grants.
A consequence of this design is that the results of combining several authorization documents can be non-obvious to end users. For example, if one authorization document grants permission for all users from the example.com domain to see their presence, and another document blocks joe@example.com, the combination of these will still provide presence to joe@example.com. Designers of user interfaces are encouraged to carefully pay attention to the results of combining multiple rules.
Another concern is cases where a user sets their privacy preferences from one client, uploads their presence authorization document to a server, and then modifies them from a different client. If the clients support different subsets of the document format, users may be confused about what information is being revealed. Clients retrieving presence authorization documents from a server SHOULD render, to the users, information about rules that they do not understand, so that users can be certain what rules are in place.
[1] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997.
[2] Rosenberg, J., "The Extensible Markup Language (XML) Configuration Access Protocol (XCAP)", RFC 4825, May 2007.
[3] Sugano, H., Fujimoto, S., Klyne, G., Bateman, A., Carr, W., and J. Peterson, "Presence Information Data Format (PIDF)", RFC 3863, August 2004.
[4] Franks, J., Hallam-Baker, P., Hostetler, J., Lawrence, S., Leach, P., Luotonen, A., and L. Stewart, "HTTP Authentication: Basic and Digest Access Authentication", RFC 2617, June 1999.
[5] Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston, A., Peterson, J., Sparks, R., Handley, M., and E. Schooler, "SIP: Session Initiation Protocol", RFC 3261, June 2002.
[6] Rosenberg, J., "A Watcher Information Event Template-Package for the Session Initiation Protocol (SIP)", RFC 3857, August 2004.
[7] Roach, A., "Session Initiation Protocol (SIP)-Specific Event Notification", RFC 3265, June 2002.
[8] Schulzrinne, H., Tschofenig, H., Morris, J., Cuellar, J., Polk, J., and J. Rosenberg, "Common Policy: A Document Format for Expressing Privacy Preferences", RFC 4745, February 2007.
[9] Schulzrinne, H., Gurbani, V., Kyzivat, P., and J. Rosenberg, "RPID: Rich Presence Extensions to the Presence Information Data Format (PIDF)", RFC 4480, July 2006.
[10] Rosenberg, J., "A Data Model for Presence", RFC 4479, July 2006.
[11] Mealling, M., "The IETF XML Registry", BCP 81, RFC 3688, January 2004.
[12] Schulzrinne, H., "The tel URI for Telephone Numbers", RFC 3966, December 2004.
Rosenberg Standards Track [Page 26]
RFC 5025 Presence Authorization December 2007
[13] Duerst, M. and M. Suignard, "Internationalized Resource Identifiers (IRIs)", RFC 3987, January 2005.
[14] Peterson, J., "A Privacy Mechanism for the Session Initiation Protocol (SIP)", RFC 3323, November 2002.
[15] Peterson, J. and C. Jennings, "Enhancements for Authenticated Identity Management in the Session Initiation Protocol (SIP)", RFC 4474, August 2006.
[16] Jennings, C., Peterson, J., and M. Watson, "Private Extensions to the Session Initiation Protocol (SIP) for Asserted Identity within Trusted Networks", RFC 3325, November 2002.
[17] Day, M., Rosenberg, J., and H. Sugano, "A Model for Presence and Instant Messaging", RFC 2778, February 2000.
[18] Rosenberg, J., "A Presence Event Package for the Session Initiation Protocol (SIP)", RFC 3856, August 2004.
This document is subject to the rights, licenses and restrictions contained in BCP 78, and except as set forth therein, the authors retain all their rights.
This document and the information contained herein are provided on an "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST AND THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
Intellectual Property
The IETF takes no position regarding the validity or scope of any Intellectual Property Rights or other rights that might be claimed to pertain to the implementation or use of the technology described in this document or the extent to which any license under such rights might or might not be available; nor does it represent that it has made any independent effort to identify any such rights. Information on the procedures with respect to rights in RFC documents can be found in BCP 78 and BCP 79.
Copies of IPR disclosures made to the IETF Secretariat and any assurances of licenses to be made available, or the result of an attempt made to obtain a general license or permission for the use of such proprietary rights by implementers or users of this specification can be obtained from the IETF on-line IPR repository at http://www.ietf.org/ipr.
The IETF invites any interested party to bring to its attention any copyrights, patents or patent applications, or other proprietary rights that may cover technology that may be required to implement this standard. Please address the information to the IETF at ietf-ipr@ietf.org.