Internet Engineering Task Force (IETF) J. Field Request for Comments: 8322 Pivotal Category: Standards Track S. Banghart ISSN: 2070-1721 D. Waltermire NIST February 2018
Resource-Oriented Lightweight Information Exchange (ROLIE)
Abstract
This document defines a resource-oriented approach for security automation information publication, discovery, and sharing. Using this approach, producers may publish, share, and exchange representations of software descriptors, security incidents, attack indicators, software vulnerabilities, configuration checklists, and other security automation information as web-addressable resources. Furthermore, consumers and other stakeholders may access and search this security information as needed, establishing a rapid and on-demand information exchange network for restricted internal use or public access repositories. This specification extends the Atom Publishing Protocol and Atom Syndication Format to transport and share security automation resource representations.
Status of This Memo
This is an Internet Standards Track document.
This document is a product of the Internet Engineering Task Force (IETF). It represents the consensus of the IETF community. It has received public review and has been approved for publication by the Internet Engineering Steering Group (IESG). Further information on Internet Standards is available in Section 2 of RFC 7841.
Information about the current status of this document, any errata, and how to provide feedback on it may be obtained at https://www.rfc-editor.org/info/rfc8322.
Field, et al. Standards Track [Page 1]
RFC 8322 ROLIE February 2018
Copyright Notice
Copyright (c) 2018 IETF Trust and the persons identified as the document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.
Table of Contents
1. Introduction ....................................................3 2. Terminology .....................................................4 3. XML-Related Conventions .........................................5 3.1. XML Namespaces .............................................5 3.2. RELAX NG Compact Schema ....................................5 4. Background and Motivation .......................................5 5. ROLIE Requirements for the Atom Publishing Protocol .............7 5.1. AtomPub Service Documents ..................................7 5.1.1. Use of the "app:workspace" Element ..................8 5.1.2. Use of the "app:collection" Element .................8 5.1.3. Service Document Discovery ..........................9 5.2. Category Documents .........................................9 5.3. Transport Layer Security ..................................10 5.4. User Authentication and Authorization .....................10 5.5. "/" (Forward Slash) Resource URL ..........................11 5.6. HTTP Methods ..............................................11 6. ROLIE Requirements for the Atom Syndication Format .............11 6.1. Use of the "atom:feed" Element ............................11 6.1.1. Use of the "atom:category" Element .................13 6.1.2. Use of the "atom:link" Element .....................14 6.1.3. Use of the "atom:updated" Element ..................15 6.2. Use of the "atom:entry" Element ...........................16 6.2.1. Use of the "atom:content" Element ..................17 6.2.2. Use of the "atom:link" Element .....................17 6.2.3. Use of the "rolie:format" Element ..................18 6.2.4. Use of the "rolie:property" Element ................19 6.2.5. Requirements for a Standalone Entry ................20
Field, et al. Standards Track [Page 2]
RFC 8322 ROLIE February 2018
7. Available Extension Points Provided by ROLIE ...................21 7.1. The Category Extension Point ..............................21 7.1.1. General Use of the "atom:category" Element .........22 7.1.2. Identification of Security Automation Information Types ..................................22 7.2. The "rolie:format" Extension Point ........................24 7.3. The Link Relation Extension Point .........................24 7.4. The "rolie:property" Extension Point ......................24 8. IANA Considerations ............................................26 8.1. XML Namespaces and Schema URNs ............................26 8.2. ROLIE URN Sub-namespace ...................................26 8.3. ROLIE URN Parameters ......................................27 8.4. ROLIE Information Types Registry ..........................29 9. Security Considerations ........................................29 10. Privacy Considerations ........................................31 11. References ....................................................32 11.1. Normative References .....................................32 11.2. Informative References ...................................34 Appendix A. RELAX NG Compact Schema for ROLIE .....................37 Appendix B. Examples of Use .......................................37 B.1. Service Discovery ..........................................37 B.2. Feed Retrieval .............................................40 B.3. Entry Retrieval ............................................42 Acknowledgements ..................................................43 Authors' Addresses ................................................43
This document defines a resource-oriented approach to security automation information sharing that follows the Representational State Transfer (REST) architectural style [REST]. In this approach, computer security resources are maintained in web-accessible repositories structured as Atom Syndication Format [RFC4287] Feeds. Within a given Feed, which may be requested by the consumer, representations of specific types of security automation information are organized, categorized, and described. Furthermore, all collections available to a given user are discoverable, allowing the consumer to search all available content they are authorized to view, and to locate and request the desired information resources. Through the use of granular authentication and access controls, only authorized consumers may be permitted the ability to read or write to a given Feed.
The goal of this approach is to increase the communication and sharing of security information between providers and consumers that can be used to automate security processes (e.g., incident reports, vulnerability assessments, configuration checklists, and other security automation information). Such sharing allows human
Field, et al. Standards Track [Page 3]
RFC 8322 ROLIE February 2018
operators and computer systems to leverage this standardized communication system to gather information that supports the automation of security processes.
To support new types of security automation information being used as time goes on, this specification defines a number of extension points that can be used either privately or globally. These global extensions are IANA-registered by Resource-Oriented Lightweight Information Exchange (ROLIE) extension specifications and provide enhanced interoperability for new use cases and domains. Sections 5 and 6 of this document define the requirements for XML representations of ROLIE; other equivalent representations (e.g. JSON) may be described by other documents. An overview of the extension system is provided in Section 7. Implementers seeking to provide support for specific security automation information types should refer to the specification for that domain as described by the IANA registry found in Section 8.4.
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.
The previous key words are used in this document to define only the requirements for implementations of this specification and are not used for recommendations or requirements for the usage of ROLIE. (In other words, a programmer of a ROLIE server MUST implement a given feature, but a user of that ROLIE server needn't use that feature.)
Definitions for some of the common computer-security-related terminology used in this document can be found in Section 2 of [RFC7970].
The following term is unique to this specification:
Information Type: A class of security automation information having one or more associated data models. Often, such security automation information is used in the automation of a security process. See Section 7.1.2 for more information.
This specification uses XML namespaces [W3C.REC-xml-names-20091208] to uniquely identify XML element names. It uses the following namespace prefix mappings for the indicated namespace URI:
In order to automate security processes, tools need access to sufficient sources of structured security information that can be used to drive security processes. Thus, security information sharing is one of the core components of automating security processes. Vulnerabilities, configurations, software identification, security incidents, and patch data are just a few of the classes of information that are shared today to enable effective security on a wide scale. However, as the scale of defense broadens as networks become larger and more complex, and the volume of information to process makes humans-in-the-loop difficult to scale, the need for automation and machine-to-machine communication becomes increasingly critical.
Field, et al. Standards Track [Page 5]
RFC 8322 ROLIE February 2018
ROLIE seeks to address this need by providing four major information- sharing benefits:
Extensible information type categories and format agnosticism: ROLIE is not bound to any given data format or category of information. Instead, information categories are extensible, and Entries declare the format of the referenced data. In cases where several formats or serializations are available, ROLIE can use link relations to communicate how a consumer can access these formats. For example, clients may request that a given resource representation be returned as XML, JSON, or in some other format or serialization. This approach allows the provider to support multiple isomorphic formats, allowing the consumer to select the most suitable version.
Open and distributed information sharing: Using the Atom Publishing Protocol (AtomPub), ROLIE Feeds can easily aggregate Feeds and accept information posted to them from other sources. Webs of communicating ROLIE servers form ad hoc sharing communities, increasing data availability and the ability to correlate linked data across sources for participating consumers. ROLIE servers needn't be distributed, however, as large ROLIE repositories can function as a central collection or federated collections.
Stateless communication model: ROLIE, as a RESTful system, is stateless. That is, the server doesn't keep track of client sessions but rather uses link relations for state transitions. In practice, this means that any consumer can find and share information at any organizational level and at any time without needing to execute a long series of requests.
Information discovery and navigation: ROLIE provides a number of mechanisms to allow clients to programmatically discover and navigate collections of information in order to dynamically discover new or revised content. Extensible information types and other categories provide one way of determining content that is desirable. Link elements, each with a target URI and an established relationship type, provide a means for ROLIE providers to link other information that is relevant to the current Entry or Feed.
These benefits result in an information-sharing protocol that is lightweight, interactive, open, and, most importantly, machine readable.
The requirements in this specification are broken into two major sections: extensions to AtomPub [RFC5023] and extensions to the Atom Syndication Format [RFC4287]. All normative requirements in AtomPub
Field, et al. Standards Track [Page 6]
RFC 8322 ROLIE February 2018
and Atom Syndication are inherited from their respective specifications and apply here unless the requirement is explicitly overridden in this document. In this way, this document may upgrade the requirement (e.g., make a "SHOULD" a "MUST") but will never downgrade a given requirement (e.g., make a "MUST" a "SHOULD").
5. ROLIE Requirements for the Atom Publishing Protocol
This section describes a number of restrictions of, and extensions to, AtomPub [RFC5023] that define the use of AtomPub in the context of a ROLIE-based solution. The normative requirements in this section are generally oriented towards client and server implementations. An understanding of the AtomPub specification [RFC5023] is helpful to understand the requirements in this section.
As described in Section 8 of [RFC5023], a Service Document is an XML-based document format that allows a client to dynamically discover the Collections provided by a publisher. A Service Document consists of one or more "app:workspace" elements that may each contain a number of "app:collection" elements.
The general structure of a Service Document is as follows (from Section 4.2 of [RFC5023]):
In AtomPub, a workspace, represented by the "app:workspace" element, describes a group of one or more Collections. Building on the AtomPub concept of a workspace, in ROLIE a workspace represents an aggregation of Collections pertaining to security automation information resources. This specification does not restrict the number of workspaces that may be in a Service Document or the specific Collections to be provided within a given workspace.
A ROLIE implementation can host Collections containing both public and private information Entries. It is suggested that implementations segregate Collections into different "app:workspace" elements by their client access requirements. With proper naming of workspaces, this reduces the amount of trial and error a human user would need to utilize to discover accessible Collections.
In AtomPub, a Collection in a Service Document, represented by the "app:collection" element, provides metadata that can be used to point to a specific Atom Feed that contains information Entries that may be of interest to a client. The association between a Collection and a Feed is provided by the "href" attribute of the "app:collection" element. Building on the AtomPub concept of a Collection, in ROLIE a Collection represents a pointer to a group of security automation information resources pertaining to a given type of security automation information. Collections are represented as Atom Feeds as per RFC 5023. Requirements specific to Atom Feed are defined in Section 6.1.
ROLIE defines specialized data requirements for Collections, Feeds, and Entries containing data related to security automation. The difference between a ROLIE Collection and a non-ROLIE Collection defined in a Service Document can be determined as follows:
ROLIE Collection: An app:collection is considered a ROLIE Collection when it contains an "app:categories" element that contains only one "atom:category" element with a "scheme" attribute value of "urn:ietf:params:rolie:category:information-type". Further, this category has an appropriate "term" attribute value as defined in Section 7.1.1. This ensures that a given Collection corresponds to a specific type of security automation information.
Non-ROLIE Collection: An app:collection is considered a non-ROLIE Collection when it does not contain an "atom:category" element with a "scheme" attribute value of "urn:ietf:params:rolie:category:information-type".
Field, et al. Standards Track [Page 8]
RFC 8322 ROLIE February 2018
By distinguishing between ROLIE and non-ROLIE Collections in this way, implementations supporting ROLIE can host Collections pertaining to security automation information alongside Collections of other non-ROLIE information within the same AtomPub instance.
The following are additional requirements on the use of the "app:collection" element for a ROLIE Collection:
o The child "atom:category" elements contained in the "app:categories" element MUST be the same set of "atom:category" elements used in the Atom Feed resource referenced by the "app:collection" element's "href" attribute value. This ensures that the category metadata associated with the Collection and the associated Feed is discoverable in both of these resources.
o The "app:categories" element in an app:collection MAY include additional "atom:category" elements using a scheme other than "urn:ietf:params:rolie:category:information-type". This allows other category metadata to be included.
The Service Document serves as the "head" of a given ROLIE repository: from the Service Document, all other repository content can be discovered. A client will need to determine the URL of this Service Document to discover the Collections provided by the repository. The client might determine the URL from a web page, based on out-of-band communication, or through a "service" link relation in a Feed or Entry Document that the client has already retrieved. The latter is a typical scenario if the client learns of a specific Feed or Entry through an out-of-band mechanism and wishes to discover additional information provided by the repository.
This document does not provide a fully automated discovery mechanism. A mechanism may be defined in the future that allows automated clients to discover the URL to use to retrieve a ROLIE Service Document representing the head of the ROLIE repository.
As described in Section 7 of [RFC5023], a Category Document is an XML-based document format that allows a client to dynamically discover the categories used within AtomPub Service Documents, Atom Syndication Feeds, and Entry Documents provided by a publisher. A Category Document consists of one "app:categories" element that contains a number of inline "atom:category" elements, or a URI referencing a Category Document.
ROLIE is intended to be handled with Transport Layer Security (TLS). TLS version 1.2 MUST be supported. TLS 1.2 SHOULD be implemented according to all recommendations and best practices presented in [RFC7525].
It is RECOMMENDED that the most recent published version of TLS be supported. If this version is TLS 1.3 [TLS-1.3], it is suggested that 0-RTT (Zero Round-Trip Time Resumption) not be used, in order to prevent replay attacks. Replay attacks on PUT, POST, or DELETE requests can disrupt repository operation by modifying data unexpectedly.
For example, an automated ROLIE repository that updates very frequently may receive a PUT request against a given resource a few times an hour (or more). An attacker may store an early PUT request, and at the end of the resumption window replay the PUT request, reverting the resource to an old version. Not only could an attacker be doing this replay continuously to cause havoc on the server, but the client is completely unaware of the attack taking place.
Given the potentially sensitive nature of data handled by ROLIE, all appropriate precautions should be taken at the transport layer to protect forward secrecy and user privacy.
The server MUST implement certificate-based client authentication. This MAY be enabled on a workspace-by-workspace basis.
Implementations MUST support user authentication. However, a given implementation MAY allow user authentication to be disabled on a Feed-by-Feed or workspace-by-workspace basis.
It is recommended that servers participating in an information- sharing consortium and supporting interactive user logins by members of the consortium support client authentication via a federated identity scheme.
This document does not mandate the use of any specific user authorization mechanisms. However, service implementers SHOULD support appropriate authorization checking for all resource accesses, including individual Atom Entries, Atom Feeds, and Atom Service Documents.
The "/" resource MAY be supported for compatibility with existing deployments that are using [RFC6546] ("Transport of Real-time Inter-network Defense (RID) Messages over HTTP/TLS"). The following requirements apply only to implementations that support both RFC 6546 and the "/" resource as described above:
o Consistent with Erratum ID 3267 [Err3267] for [RFC6546], a client requesting a GET on the "/" resource SHOULD receive an HTTP status code 405 ("Method Not Allowed").
o An implementation MAY provide full support for [RFC6546] such that a POST to the "/" resource containing a recognized RID message is handled correctly as a RID request. Alternatively, a client requesting a POST to "/" MAY receive an HTTP status code 307 ("Temporary Redirect"). In this case, the location header in the HTTP response will provide the URL of the appropriate RID endpoint, and the client may repeat the POST method at the indicated location.
If RFC 6546 is unsupported, then a request for the "/" resource may be handled as deemed appropriate by the server.
Servers MAY accept request methods beyond those specified in this document.
Clients MUST be capable of recognizing and processing any standard HTTP status code, as defined in Section 5 of [RFC5023].
6. ROLIE Requirements for the Atom Syndication Format
This section describes a number of restrictions of, and extensions to, the Atom Syndication Format [RFC4287] that define the valid use of the format in the context of a ROLIE implementation. An understanding of the Atom Syndication Format specification [RFC4287] is helpful to understand the requirements in this section.
As described in Section 4.1.1 of [RFC4287], an Atom Feed is an XML-based document format that describes a list of related information items. The Atom Feeds provided by a ROLIE service are listed in the service's Service Document through one or more "app:collection" elements. Each Feed Document, represented using the "atom:feed" element, contains a listing of zero or more Entries.
Field, et al. Standards Track [Page 11]
RFC 8322 ROLIE February 2018
When applied to the problem domain of security automation information sharing, an Atom Feed may be used to represent any meaningful collection of security automation information resources. Each Entry in a Feed represents an individual resource (e.g., a specific checklist, a software vulnerability record). Additional Feeds can be used to represent other collections of security automation resources.
As discussed in Section 5.1.2, ROLIE defines specialized data requirements for Feeds containing data related to security automation. The difference between a ROLIE Feed and a non-ROLIE Feed can be determined as follows:
ROLIE Feed: For an "atom:feed" to be considered a ROLIE Feed, the "atom:feed" MUST contain only one child "atom:category" element with a "scheme" attribute value of "urn:ietf:params:rolie:category:information-type". This category MUST have an appropriate "term" attribute value as defined in Section 7.1.1. This ensures that a given Feed corresponds to a specific type of security automation information.
Non-ROLIE Feed: For an "atom:feed" to be considered a non-ROLIE Feed, the "atom:feed" MUST NOT contain an "atom:category" element with a "scheme" attribute value of "urn:ietf:params:rolie:category:information-type".
By distinguishing between ROLIE and non-ROLIE Feeds in this way, implementations supporting ROLIE can host Feeds pertaining to security automation information alongside Feeds of other non-ROLIE information within the same AtomPub instance. This is parallel to the handling of Collections as discussed earlier in this specification (Section 5.1.2).
Field, et al. Standards Track [Page 12]
RFC 8322 ROLIE February 2018
The following Atom Feed definition represents a stricter definition of the "atom:feed" element defined in [RFC4287] when used as a ROLIE Feed. Any element not specified here inherits its definition and requirements from [RFC4287].
An "atom:feed" can contain one or more "atom:category" elements. In Atom, the naming scheme and the semantic meaning of the terms used to identify an Atom category are application defined.
The following are additional requirements on the use of the "atom:category" element when used in a ROLIE Feed:
o All member Entries in the Feed MUST represent security automation information records of the provided information type category.
o The "atom:feed" MAY include additional "atom:category" elements using a scheme other than "urn:ietf:params:rolie:category:information-type". This allows other category metadata to be included.
Link relations defined by the "atom:link" element are used to represent state transitions using a stateless approach. In Atom, a type of link relationship can be defined using the "rel" attribute.
A ROLIE Feed MUST contain one or more "atom:link" elements with rel="service" and an "href" attribute whose value is a URI that points to an Atom Service Document associated with the Feed. If a client accesses a Feed without first accessing the service's Service Document, a link with the "service" relationship provides a means to discover additional security automation information. The "service" link relationship is defined in the IANA "Link Relations" registry at <https://www.iana.org/assignments/link-relations/>.
A Feed can contain an arbitrary number of Entries. In some cases, a complete Feed may consist of a large number of Entries. Additionally, as new and updated Entries are ordered at the beginning of a Feed, a client may only be interested in retrieving the first N Entries in a Feed to process only the Entries that have changed since the last retrieval of the Feed. As a practical matter, a large set of Entries will likely need to be divided into more manageable portions, or pages. Based on Section 3 of [RFC5005], link elements SHOULD be included in all Feeds to support paging using the following link relation types:
o "first" - Indicates that the "href" attribute value of the link identifies a resource URI for the furthest preceding page of the Feed.
o "last" - Indicates that the "href" attribute value of the link identifies a resource URI for the furthest following page of the Feed.
o "previous" - Indicates that the "href" attribute value of the link identifies a resource URI for the immediately preceding page of the Feed.
o "next" - Indicates that the "href" attribute value of the link identifies a resource URI for the immediately following page of the Feed.
A reference to a historical Feed may need to be stable, and/or a Feed may need to be divided into a series of defined epochs. Implementations SHOULD support the mechanisms described in Section 4 of [RFC5005] to provide link-based state transitions for maintaining the archiving of Feeds.
A Feed MAY include additional link relationships not specified in this document. If a client encounters an unknown link relationship type, the client MUST ignore the unrecognized link and continue processing as if the unrecognized link element did not appear. The definition of new link relations that provide additional state transition extensions is discussed in Section 7.3.
The "atom:updated" element identifies the date and time that a Feed was last updated.
The "atom:updated" element MUST be populated with the current time at the instant the Feed was last updated by adding, updating, or deleting an Entry, or by changing any metadata for the Feed.
Each Entry in an Atom Feed, represented by the "atom:entry" element, describes a single referenced information record, along with descriptive information about its format, media type, and other publication metadata. The following "atom:entry" schema definition represents a stricter representation of the "atom:entry" element defined in [RFC4287] for use in a ROLIE-based Atom Feed as defined in Section 6.1.1.
The notable changes from [RFC4287] are the addition of "rolieFormat" and "rolieProperty" elements. Also, the "atomContent" element is restricted to the atomOutOfLineContent formulation and is now REQUIRED.
The following subsections contain requirements for Entries in a ROLIE Feed.
An "atom:content" element associates its containing Entry with a content resource identified by the "src" attribute.
There MUST be exactly one "atom:content" element in the Entry. The "atom:content" element MUST adhere to this definition, which is a stricter representation of the "atom:content" element defined in [RFC4287]:
atomContent = element atom:content { atomCommonAttributes, attribute type { atomMediaType }, attribute src { atomUri }, empty }
This restricts atomContent in ROLIE to the atomOutOfLineContent formulation presented in [RFC4287].
The "type" attribute MUST identify the serialization type of the content -- for example, "application/xml" or "application/json". A prefixed media type MAY be used to reflect a specific model used with a given serialization approach (e.g., "application/rdf+xml"). The "src" attribute MUST be a URI that can be dereferenced to retrieve the related content data.
Link relations can be included in an Entry to represent state transitions to and from the Entry, as well as to provide links to related information.
If there is a need to provide the same information in different data models and/or serialization formats, separate Entry instances can be included in the same Feed or a different Feed. Such an alternate content representation can be indicated using an "atom:link" having a "rel" attribute with the value "alternate".
A Feed MAY include additional link relationships not specified in this document. If a client encounters an unknown link relationship type, the client MUST ignore the unrecognized link and continue processing as if the unrecognized link element did not appear. The definition of new link relations that provide additional state transition extensions is discussed in Section 7.3.
As mentioned in Sections 1 and 4, a key goal of this specification is to allow a consumer to review a set of published security automation information resources and then identify and retrieve any resources of interest. The format of the data is a key criteria to consider when deciding what information to retrieve. For a given type of security automation information, it is expected that a number of different formats may be used to represent this information. To support this use case, both the serialization format and the specific data model expressed in that format must be known by the consumer.
In the Atom Syndication Format, a media type can be defined using the "type" attribute of the "atom:content" element of an "atom:entry". The media type can be fully descriptive of the format of the linked document, such as "application/atom+xml". In some cases, however, a format-specific media type may not be defined. An example might be when "application/xml" is used because there is no defined specific media type for the content. In such a case, the exact data model of the content cannot be known without first retrieving the content.
In cases where a specific media type does not exist, the "rolie:format" element is used to describe the data model used to express the information referenced in the "atom:content" element. The "rolie:format" element also allows a schema to be identified that can be used when parsing the content to verify or better understand the structure of the content.
When it appears, the "rolie:format" element MUST adhere to this definition:
The "rolie:format" element MUST provide a "ns" attribute that identifies the data model of the resource referenced by the "atom:content" element. For example, the namespace used may be an XML namespace URI or an identifier that represents a serialized JSON model. The URI used for the "ns" attribute MUST be absolute. The resource identified by the URI need not be resolvable.
Field, et al. Standards Track [Page 18]
RFC 8322 ROLIE February 2018
The "rolie:format" element MAY provide a "version" attribute that identifies the version of the format used for the related "atom:content" element.
The "rolie:format" element MAY provide a "schema-location" attribute, which is a URI that identifies a schema resource that can be used to validate the related "atom:content" element.
The "rolie:format" element MAY provide a "schema-type" attribute, which is a media type (as described in [RFC2045]) identifying the format of the schema resource identified by the "schema-location" attribute.
The following nominal example shows how these attributes describe the format of the content:
The previous element provides an indication that the content of the given Entry is using the Incident Object Description Exchange Format (IODEF) v2 format.
An "atom:category" element provides a way to associate a name/value pair of categorical information using the "scheme" and "term" attributes to represent the name and using the "label" attribute to represent the value. When used in this way, an "atom:category" allows a specific label to be selected from a finite set of possible label values that can be used to further classify a given Entry or Feed. Within ROLIE, there may be a need to associate additional metadata with an Entry. In such a case, the use of an "atom:category" is not practical to represent name/value data for which the allowed values are unbounded. Instead, ROLIE introduces a new "rolie:property" element that can represent non-categorical metadata as name/value pairs. Examples include content-specific identifiers, naming data, and other properties that allow for unbounded values.
There MAY be zero or more "rolie:property" elements in an "atom:entry".
Field, et al. Standards Track [Page 19]
RFC 8322 ROLIE February 2018
The element MUST adhere to this definition:
rolieProperty = element rolie:property { atomCommonAttributes, attribute name { atomUri }, attribute value { text }, empty }
The "name" attribute provides a URI that identifies the namespace and name of the property as a URI.
The "value" attribute is text that provides a value for the property identified by the "name" attribute.
For example, the nominal element <rolie:property name="urn:ietf:params:rolie:property:content-id" value="12345"/> would expose an IODEF ID value contained in a given Entry's content. The name used in the example also demonstrates the use of a registered ROLIE property extension, which is described in Section 7.4.
Implementations MAY use locally defined and namespaced elements in an Entry in order to provide additional information. Clients that do not recognize a property with an unregistered "name" attribute MUST ignore the "rolie:property" element; that is, the client MUST NOT fail parsing content that contains an unrecognized property.
If an Entry is ever shared as a standalone resource, separate from its containing Feed, then the following additional requirements apply:
o The Entry MUST have an "atom:link" element with rel="collection" and href="[URI of the containing Collection]". This allows the Feed or Feeds of which the Entry is a member to be discovered, along with the related information the Feed may contain. In the case where the Entry has multiple containing Feeds, the Entry MUST have one "atom:link" for each related Feed.
o The Entry MUST declare the information type of the content resource referenced by the Entry (see Section 7.1.2).
This specification does not require particular information types or data formats; rather, ROLIE is intended to be extended by additional specifications that define the use of new categories and link relations. The primary point of extension is through the definition of new information type category terms. Additional specifications can register new information type category terms with IANA that serve as the main characterizing feature of a ROLIE Collection/Feed or resource/Entry. These additional specifications defining new information type terms can describe additional requirements for including specific categories and link relations, as well as the use of specific data formats supporting a given information type term.
The "atom:category" element, defined in Section 4.2.2 of [RFC4287], provides a mechanism to provide additional categorization information for a content resource in ROLIE. The ability to define new categories is one of the core extension points provided by Atom. A Category Document, defined in Section 7 of [RFC5023], provides a mechanism for an Atom implementation to make discoverable the "atom:category" terms and associated allowed values.
ROLIE further defines the use of the existing Atom extension category mechanism by allowing ROLIE-specific category extensions to be registered with IANA. The "urn:ietf:params:rolie:category:information-type" category scheme, which has special meaning for implementations of ROLIE, has been assigned (see Section 8.3). This allows category scheme namespaces to be managed in a more consistent way, allowing for greater interoperability between content producers and consumers.
Any "atom:category" element whose "scheme" attribute uses an unregistered scheme MUST be considered "Private Use" as defined in [RFC8126]. Implementations encountering such a category MUST parse the content without error but MAY otherwise ignore the element.
The use of the "atom:category" element is discussed in the following subsections.
The "atom:category" element can be used for characterizing a ROLIE resource. An "atom:category" element has a "term" attribute that indicates the assigned category value and a "scheme" attribute that provides an identifier for the category type. The "scheme" provides a means to describe how a set of category terms should be used and provides a namespace that can be used to differentiate terms that are provided by multiple organizations and that have different semantic meaning.
To further differentiate category types used in ROLIE, an IANA subregistry has been established for ROLIE protocol parameters to support the registration of new category "scheme" attribute values by ROLIE extension specifications. The use of this extension point is discussed in Section 8.3, using the "name" field with a type parameter of "category" to indicate a category extension.
7.1.2. Identification of Security Automation Information Types
A ROLIE-specific extension point is provided through the "atom:category" element's "scheme" attribute value "urn:ietf:params:rolie:category:information-type". This value is a Uniform Resource Name (URN) [RFC8141] that is registered with IANA as described in Section 8.3. When used as the "scheme" attribute in this way, the "term" attribute is expected to be a registered value as defined in Section 8.4. Through this mechanism, a given security automation information type can be used to:
1. identify that an "app:collection" element in a Service Document points to an Atom Feed that contains Entries pertaining to a specific type of security automation information (see Section 5.1.2),
2. identify that an "atom:feed" element in an Atom Feed contains Entries pertaining to a specific type of security automation information (see Section 6.1.1), or
3. identify the information type of a standalone resource (see Section 6.2.5).
Field, et al. Standards Track [Page 22]
RFC 8322 ROLIE February 2018
For example, the notional security automation information type "incident" would be identified as follows:
A security automation information type represents a class of information that represents the same or similar information model [RFC3444]. Note that this document does not register any information types but offers the following as examples of potential information types:
indicator: Computing device- or network-related "observable features and phenomenon that aid in the forensic or proactive detection of malicious activity and associated metadata" (from [RFC7970]).
incident: Information pertaining to or derived from security incidents.
vulnerability reports: Information identifying and describing a vulnerability in hardware or software.
configuration checklists: Content that can be used to assess the configuration settings related to installed software.
software tags: Metadata used to identify and characterize installable software.
This is a short list to inspire new engineering of information type extensions that support the automation of security processes.
This document does not specify any information types. Instead, information types in ROLIE are expected to be registered in extension documents that describe one or more new information types. This allows the information types used by ROLIE implementations to grow over time to support new security automation use cases. These extension documents may also enhance ROLIE Service, Category, Feed, and Entry Documents by defining link relations, other categories, and Format data model extensions to address the representational needs of these specific information types. New information types are added to ROLIE through registrations to the IANA "ROLIE Information Types" registry defined in Section 8.4.
Security automation data pertaining to a given information type may be expressed using a number of supported formats. As described in Section 6.2.3, the "rolie:format" element is used to describe the specific data model used to represent the resource referenced by a given "atom:entry". The structure provided by the "rolie:format" element provides a mechanism for extension within the "atom:entry" model. ROLIE extensions MAY further restrict which data models are allowed to be used for a given information type.
By declaring the data model used for a given resource, a consumer can choose to download or ignore the resource, or look for alternate formats. This saves the consumer from downloading and parsing resources that the consumer is not interested in or resources expressed in formats that are not supported by the consumer.
This document uses several link relations defined in the IANA "Link Relation Types" registry at <https://www.iana.org/assignments/link-relations/>. Additional link relations can be registered in this registry to allow new relationships to be represented in ROLIE according to Section 4.2.7.2 of [RFC4287]. Based on the preceding reference, if the link relation is too specific or limited in its intended use, an absolute URI can be used in lieu of registering a new simple name with IANA.
As discussed previously in Section 6.2.3, many formats contain unique identifying and characterizing properties that are vital for sharing information. In order to provide a global reference for these properties, this document establishes an IANA registry that allows ROLIE extensions to register named properties using the "name" field with a type parameter of "property" to indicate a property extension; see Section 8.3. Implementations SHOULD prefer the use of registered properties over implementation-specific properties when possible.
ROLIE extensions are expected to register new properties and use existing properties to provide valuable identifying and characterizing information for a given information type and/or format.
Field, et al. Standards Track [Page 24]
RFC 8322 ROLIE February 2018
Any "rolie:property" element whose "name" attribute has "urn:ietf:params:rolie:property:local" as a prefix MUST be considered "Private Use" as defined in [RFC8126]. Implementations encountering such a property MUST parse the content without error but MAY otherwise ignore the element.
This document also registers a number of general-use properties that can be used to expose content information in any ROLIE use case. The following are descriptions of how to use these registered properties:
urn:ietf:params:rolie:property:content-author-name The "value" attribute of this property is a text representation indicating the individual or organization that authored the content referenced by the "src" attribute of the Entry's "atom:content" element. This author may differ from the "atom:author" element when the author of the content and the author of the Entry are different people or entities.
urn:ietf:params:rolie:property:content-id The "value" attribute of this property is a text representation of an identifier pertaining to or extracted from the content referenced by the "src" attribute of the Entry's "atom:content" element. For example, if the "atom:entry"'s "atom:content" element links to an IODEF document, the "content-id" value would be an identifier of that IODEF document.
urn:ietf:params:rolie:property:content-published-date The "value" attribute of this property is a text representation indicating the original publication date of the content referenced by the "src" attribute of the Entry's "atom:content" element. This date may differ from the published date of the ROLIE Entry because publication of the content and publication of the ROLIE Entry represent different events. The date MUST be formatted as specified in [RFC3339].
urn:ietf:params:rolie:property:content-updated-date The "value" attribute of this property is a text representation indicating the date that the content, referenced by the "src" attribute of the Entry's "atom:content" element, was last updated. This date may differ from the updated date of the ROLIE Entry because updates made to the content and to the ROLIE Entry are different events. The date MUST be formatted as specified in [RFC3339].
A new top-level registry has been created, titled "Resource-Oriented Lightweight Information Exchange (ROLIE) URN Parameters".
Registration in the "ROLIE URN Parameters" subregistry is via the Specification Required policy [RFC8126]. Registration requests must be sent to both the MILE Working Group mailing list (mile@ietf.org) and IANA. IANA will forward registration requests to the Designated Expert.
Each entry in this subregistry must record the following fields:
Name: A URN segment that adheres to the pattern {type}:{label}. The keywords are defined as follows:
{type}: The parameter type. The allowed values are "category" or "property". "category" denotes a category extension as discussed in Section 7.1. "property" denotes a property extension as discussed in Section 7.4.
{label}: A required US-ASCII string that conforms to the URN syntax requirements (see [RFC8141]). This string must be unique within the namespace defined by the {type} keyword. The "local" label for both the "category" and "property" types has been reserved for private use.
Extension URI: The identifier to use within ROLIE, which is the full URN using the form "urn:ietf:params:rolie:{name}", where {name} is the "name" field of this registration.
Reference: A static link to the specification and section where the definition of the parameter can be found.
Subregistry: An optional field that links to an IANA subregistry for this parameter. If the {type} is "category", the subregistry must contain a "name" field whose registered values MUST be US-ASCII. The list of names are the allowed values of the "term" attribute in the "atom:category" element (see Section 7.1.2).
Name: The full name of the security resource information type as a string from the printable ASCII character set [RFC20] with individual embedded spaces allowed. This value must be unique in the context of this table. The ABNF [RFC5234] syntax for this field is:
1*VCHAR *(SP 1*VCHAR)
Index: An IANA-assigned positive integer that identifies the registration. The first entry added to this registry uses the value 1, and this value is incremented for each subsequent entry added to the registry.
Reference: A list of one or more URIs [RFC3986] from which the registered specification can be obtained. The registered specification MUST be readily and publicly available from that URI. The URI SHOULD be a stable reference.
Allocation Policy: Specification Required, as per [RFC8126]
This document defines a resource-oriented approach for lightweight information exchange using HTTP over TLS, the Atom Syndication Format, and AtomPub. As such, implementers must understand the security considerations described in those specifications. All that follows is guidance; instructions that are more specific are out of scope for this document.
To protect the confidentiality of a given resource provided by a ROLIE implementation, requests for retrieval of the resource need to be authenticated to prevent unauthorized users from accessing the resource (see Section 5.4). It can also be useful to log and audit access to sensitive resources to verify that proper access controls remain in place over time.
Field, et al. Standards Track [Page 29]
RFC 8322 ROLIE February 2018
Access control to information published using ROLIE should use mechanisms that are appropriate to the sensitivity of the information. Primitive authentication mechanisms like HTTP Basic Authentication [RFC7617] are rarely appropriate for sensitive information. A number of authentication schemes are defined in the "HTTP Authentication Schemes" registry at <https://www.iana.org/assignments/http-authschemes/>. Of these, HTTP Origin-Bound Authentication (HOBA) [RFC7486] and SCRAM-SHA-256 [RFC7804] ("SCRAM" stands for "Salted Challenge Response Authentication Mechanism") provide improved security properties over HTTP Basic [RFC7617]and Digest [RFC7616] authentication schemes. However, sharing communities that are engaged in sensitive collaborative analysis and/or operational response for indicators and incidents targeting high-value information systems should adopt a suitably stronger user authentication solution, such as a risk-based or multi-factor approach.
Collaborating consortiums may benefit from the adoption of a federated identity solution, such as those based upon OAuth [RFC6749] with the JSON Web Token (JWT) [RFC7797], or SAML-core [SAML-core] ("SAML" stands for "Security Assertion Markup Language"), SAML-bind [SAML-bind], and SAML-prof [SAML-prof] for web-based authentication and cross-organizational single sign-on. Dependency on a trusted third-party identity provider implies that appropriate care must be exercised to sufficiently secure the identity provider. Any attacks on the federated identity system would present a risk to the consortium, as a relying party. Potential mitigations include deployment of a federation-aware identity provider that is under the control of the information-sharing consortium, with suitably stringent technical and management controls.
Authorization of resource representations is the responsibility of the source system, i.e., based on the authenticated user identity associated with an HTTP(S) request. The required authorization policies that are to be enforced must therefore be managed by the security administrators of the source system. Various authorization architectures would be suitable for this purpose, such as Role-Based Access Control (RBAC) <https://csrc.nist.gov/projects/ role-based-access-control> and/or Attribute-Based Access Control (ABAC), as embodied in the eXtensible Access Control Markup Language (XACML) [XACML]. In particular, implementers adopting XACML may benefit from the capability to represent their authorization policies in a standardized, interoperable format. Note that implementers are free to choose any suitable authorization mechanism that is capable of fulfilling the policy enforcement requirements relevant to their consortium and/or organization.
Field, et al. Standards Track [Page 30]
RFC 8322 ROLIE February 2018
Additional security requirements such as enforcing message-level security at the destination system could supplement the security enforcements performed at the source system; however, these destination-provided policy enforcements are out of scope for this specification. Implementers requiring this capability should consider leveraging, for example, the <RIDPolicy> element in the RID schema. Refer to Section 9 of [RFC6545] for more information. Additionally, the underlying serialization approach used in the representation (e.g., XML, JSON) can offer encryption and message authentication capabilities. For example, XML Digital Signatures (XMLDSIG) [RFC3275] for XML, as well as JSON Web Encryption [RFC7516] and JSON Web Signature [RFC7515] for JSON, can provide such mechanisms.
When security policies relevant to the source system are to be enforced at both the source and destination systems, implementers must take care to avoid unintended interactions of the separately enforced policies. Potential risks will include unintended denial of service and/or unintended information leakage. These problems may be mitigated by avoiding any dependence upon enforcements performed at the destination system. When distributed enforcement is unavoidable, the usage of a standard language (e.g., XACML) for the expression of authorization policies will enable the source and destination systems to better coordinate and align their respective policy expressions.
A service discovery mechanism is not explicitly specified in this document, but there are several approaches available for implementers. When selecting this mechanism, implementations need to ensure that their choice provides a means for authenticating the server. DNS SRV records [RFC2782] are a possible solution to the discovery problem described in Section 5.1.3.
The optional "author" field may provide an identification privacy issue if populated without the author's consent. This information may become public if posted to a public Feed. When aggregating or sharing Entries from other Feeds or when programmatically generating ROLIE Entries from some data source, special care should be taken to ensure that the author's personal information is not shared without the author's consent.
When using AtomPub to POST Entries to a Feed, attackers may use correlating techniques to profile the user. The request time can be compared to the generated "updated" field of the Entry in order to build out information about a given user. This correlation attempt can be mitigated by not using HTTP requests to POST Entries when profiling is a risk and instead using backend control of the Feeds.
Field, et al. Standards Track [Page 31]
RFC 8322 ROLIE February 2018
Adoption of the information-sharing approach described in this document will enable users to more easily perform correlations across separate, and potentially unrelated, cybersecurity information providers. A client may succeed in assembling a data set that would not have been permitted within the context of the authorization policies of either provider when considered individually. Thus, providers may face a risk of an attacker obtaining an access that constitutes an undetected separation of duties (SOD) violation. It is important to note that this risk is not unique to this specification, and a similar potential for abuse exists with any other cybersecurity information-sharing protocol. However, the wide availability of tools for HTTP clients and Atom Feed handling implies that the resources and technical skills required for a successful exploit may be less than it was previously. This risk can be best mitigated through appropriate vetting of the client at the time of account provisioning. In addition, any increase in the risk of this type of abuse should be offset by the corresponding increase in effectiveness that this specification affords to the defenders.
Overall, privacy concerns in ROLIE can be mitigated by following security considerations and by the careful use of the optional personally identifying elements (e.g., author) provided by Atom Syndication and ROLIE.
[RFC7525] Sheffer, Y., Holz, R., and P. Saint-Andre, "Recommendations for Secure Use of Transport Layer Security (TLS) and Datagram Transport Layer Security (DTLS)", BCP 195, RFC 7525, DOI 10.17487/RFC7525, May 2015, <https://www.rfc-editor.org/info/rfc7525>.
[W3C.REC-xml-names-20091208] Bray, T., Hollander, D., Layman, A., Tobin, R., and H. Thompson, "Namespaces in XML 1.0 (Third Edition)", World Wide Web Consortium Recommendation REC-xml-names-20091208, December 2009, <https://www.w3.org/TR/2009/ REC-xml-names-20091208>.
[REST] Fielding, R., "Architectural Styles and the Design of Network-based Software Architectures", 2000, <http://www.ics.uci.edu/~fielding/pubs/ dissertation/top.htm>.
[SAML-bind] Cantor, S., Hirsch, F., Kemp, J., Philpott, R., and E. Maler, "Bindings for the OASIS Security Assertion Markup Language (SAML) V2.0", OASIS Standard saml-bindings- 2.0-os, March 2005, <http://docs.oasis-open.org/ security/saml/v2.0/saml-bindings-2.0-os.pdf>.
Field, et al. Standards Track [Page 35]
RFC 8322 ROLIE February 2018
[SAML-core] Cantor, S., Kemp, J., Philpott, R., and E. Maler, "Assertions and Protocols for the OASIS Security Assertion Markup Language (SAML) V2.0", OASIS Standard saml-core- 2.0-os, March 2005, <http://docs.oasis-open.org/ security/saml/v2.0/saml-core-2.0-os.pdf>.
[SAML-prof] Hughes, J., Cantor, S., Hodges, J., Hirsch, F., Mishra, P., Philpott, R., and E. Maler, "Profiles for the OASIS Security Assertion Markup Language (SAML) V2.0", OASIS Standard saml-profiles-2.0-os, March 2005, <http://docs.oasis-open.org/security/saml/v2.0/ saml-profiles-2.0-os.pdf>.
[TLS-1.3] Rescorla, E., "The Transport Layer Security (TLS) Protocol Version 1.3", Work in Progress, draft-ietf-tls-tls13-23, January 2018.
[XACML] Rissanen, E., "eXtensible Access Control Markup Language (XACML) Version 3.0 Plus Errata 01", July 2017, <http://docs.oasis-open.org/xacml/3.0/ xacml-3.0-core-spec-en.pdf>.
This appendix provides a non-normative example of a client doing service discovery.
An Atom Service Document enables a client to dynamically discover what Feeds a particular publisher makes available. Thus, a provider uses an Atom Service Document to enable authorized clients to determine what specific information the provider makes available to
Field, et al. Standards Track [Page 37]
RFC 8322 ROLIE February 2018
the community. The Service Document should be made accessible from an easily found location, such as a link from the producer's home page.
A client may format an HTTP GET request to retrieve the Service Document from the specified location:
GET /rolie/servicedocument Host: www.example.org Accept: application/atomsvc+xml
Notice the use of the HTTP Accept: request header, indicating the MIME type for Atom service discovery. The response to this GET request will be an XML document that contains information on the specific Collections that are provided.
Example HTTP GET response:
HTTP/1.1 200 OK Date: Fri, 24 Aug 2016 17:09:11 GMT Content-Length: 570 Content-Type: application/atomsvc+xml;charset="utf-8"
This simple Service Document example shows that the server provides one workspace, named "Vulnerabilities". Within that workspace, the server makes one Collection available.
Field, et al. Standards Track [Page 38]
RFC 8322 ROLIE February 2018
A server may also offer a number of different Collections, each containing different types of security automation information. In the following example, a number of different Collections are provided, each with its own category and authorization scope. This categorization will help the clients to decide which Collections will meet their needs.
HTTP/1.1 200 OK Date: Fri, 24 Aug 2016 17:10:11 GMT Content-Length: 1912 Content-Type: application/atomsvc+xml;charset="utf-8"
In this example, the provider is making available a total of two Collections, organized into two different workspaces. The first workspace contains a Collection consisting of publicly available software vulnerabilities. The second workspace provides an incident Collection for use by a private sharing consortium. An appropriately authenticated and authorized client may then proceed to make HTTP requests for these Collections. The publicly provided vulnerability information may be accessible with or without authentication. However, users accessing the Collection restricted to authorized members of a private sharing consortium are expected to authenticate before access is allowed.
This appendix provides a non-normative example of a client retrieving a vulnerability Feed.
Having discovered the available Collections that share security information, a client who is a member of the general public may be interested in receiving the Collection of public vulnerabilities, expressed as Common Vulnerabilities and Exposures (CVEs). The client may retrieve the Feed for this Collection by performing an HTTP GET operation on the URL indicated by the Collection's "href" attribute.
Example HTTP GET request for a Feed:
GET /provider/public/vulns Host: www.example.org Accept: application/atom+xml
Field, et al. Standards Track [Page 40]
RFC 8322 ROLIE February 2018
The corresponding HTTP response would be an XML document containing the vulnerability Feed:
Example HTTP GET response for a Feed:
HTTP/1.1 200 OK Date: Fri, 24 Aug 2016 17:20:11 GMT Content-Length: 2882 Content-Type: application/atom+xml;charset="utf-8"
This Feed Document has two Atom Entries, one of which has been elided. The first Entry illustrates an "atom:entry" element that provides a summary of essential details about one particular vulnerability. Based upon this summary information and the provided
Field, et al. Standards Track [Page 41]
RFC 8322 ROLIE February 2018
category information, a client may choose to do an HTTP GET request on the content "src" attribute to retrieve the full details of the vulnerability.
This appendix provides a non-normative example of a client retrieving a vulnerability as an Atom Entry.
Having retrieved the Feed of interest, the client may then decide, based on the description and/or category information, that one of the Entries in the Feed is of further interest. The client may retrieve this vulnerability Entry by performing an HTTP GET operation on the URL indicated by the "src" attribute of the "atom:content" element.
Example HTTP GET request for an Entry:
GET /provider/public/vulns/123456 Host: www.example.org Accept: application/atom+xml;type=entry
The corresponding HTTP response would be an XML document containing the Atom Entry for the vulnerability record:
Example HTTP GET response for an Entry:
HTTP/1.1 200 OK Date: Fri, 24 Aug 2016 17:30:11 GMT Content-Length: 713 Content-Type: application/atom+xml;type=entry;charset="utf-8"
The example response above shows an XML document referenced by the "src" attribute of the "atom:content" element. The client may retrieve the document using this URL.
Acknowledgements
The authors gratefully acknowledge the valuable contributions of Tom Maguire, Kathleen Moriarty, and Vijayanand Bharadwaj. These individuals provided detailed review comments on earlier draft versions of this document and made many suggestions that have helped to improve this document.
The authors would also like to thank the MILE Working Group, the SACM Working Group, and countless other people from both within the IETF community and outside of it for their excellent review and effort towards constructing this document.
Authors' Addresses
John P. Field Pivotal Software, Inc. 625 Avenue of the Americas New York, New York 10011 United States of America
Phone: (646)792-5770 Email: jfield@pivotal.io
Stephen A. Banghart National Institute of Standards and Technology 100 Bureau Drive Gaithersburg, Maryland 20877 United States of America