Generic Naming Attributes for the Generic Security Services Application Programming Interface (GSS-API)
draft-williams-kitten-generic-naming-attributes-01

Abstract

This document specifies several useful generic naming attributes for use with the Generic Security Services Application Programming Interface (GSS-API) Naming Extensions specified in RFC6680.

These attributes allow applications to extract discrete components of a GSS-API “mechanism name” (MN) object: issuer (e.g., realm name, domain name, certification authority name), service and host names (for host-based service names), user names, and others.

Status of This Memo

This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.

Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at http://datatracker.ietf.org/drafts/current/.

Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."

This Internet-Draft will expire on February 14, 2014.

Copyright Notice

Copyright (c) 2013 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 (http://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 and Motivation
• 1.1. Naming Constraints
• 1.2. Conventions used in this document
• 2. Generic Attributes
• 2.1. Concrete Attributes
• 2.1.1. Issuer Name
• 2.1.2. User Name
• 2.1.3. Service Name
• 2.1.4. Host Name
• 2.1.5. Domain Name
• 2.2. Prefix Attributes
• 2.2.1. GSS_C_ATTR_GENERIC_UNCONSTRAINED
• 2.2.2. GSS_C_ATTR_GENERIC_UNCONSTRAINED_OK
• 2.2.3. GSS_C_ATTR_GENERIC_FAST
• 3. Local Name Attributes
• 3.1. GSS_C_ATTR_LOCAL_LOGIN_USER
• 4. Suggested Mechanism-Specific Name Attributes (INFORMATIONAL)
• 4.1. Suggested Kerberos-Specific Name Attributes
• 4.2. Suggested PKU2U-Specific Name Attributes
• 5. Security Considerations
• 6. IANA Considerations
• 7. References
• 7.1. Normative References
• 7.2. Informative References
• Author's Address

1. Introduction and Motivation

The Generic Security Services Application Programming Interface (GSS-API) [RFC2743] allows applications -and application protocol specifications- to use various security mechanisms in a generic way. There are some shortcomings of this API that preclude a fully-generic treatment of security mechanisms. This document builds on the naming extensions to the GSS-API [RFC6680] to correct some of those shortcomings.

In RFC6680 we introduced an interface by which to access “attributes” of names, but we did not specify any attributes. This document specifies some such attributes. Some of the new attributes are specifically intended to make it possible to use the GSS-API in a mechanism-generic way in common use cases where it is otherwise not possible to do so.

For example, some applications need to be able to observe the discrete elements of a peer principal's host-based service name, but they generally could only do so by parsing mechanism-specific display syntaxes or exported name token formats. Such applications are inherently not generic: they can only function correctly when used with security mechanism whose principal naming conventions/formats the applications understand.

More generally, we use the the extended naming interface to introduce an attribute model of principal naming.

1.1. Naming Constraints

This document also introduces a notion of naming constraints, not unlike PKIX's [RFC5280]. Naming constraints apply to “issuers” of principal names and/or their attributes. For example, to Kerberos [RFC4120] realms, to PKIX certification authorities, to identity providers (IdPs), and so on. The goal is allow specification of policies which constrain the set of principal names that a given issuer can issue credentials for.

For example, the Kerberos realm FOO.EXAMPLE would generally not be expected to issue credentials to host-based principals in domains other than “foo.example”.

For each concrete attribute specified below there are several ways to inquire a NAME's value for that attribute:

1. with naming constraint checking, providing no output if naming constraints are violated;
2. with naming constraint checking, providing an output indicator of naming constraint violations;
3. without naming constraint checking;
4. any of the above with “fast” (no slow I/O involved) naming constraint checking.

(1) is the default behavior. The others are obtained by adding an appropriate prefix to the attribute name.

Existing security mechanisms may not have any formal notion of naming constraints, but it is common to have some naming constraint conventions nonetheless. For example, Kerberos realm naming conventions are that realm names should mirror Domain Name System (DNS) [RFC1035] domain names, and that hostnames embedded in Kerberos principal names should a) be fully-qualified, b) within the domain corresponding to the DNS domain name derived from the realm's name. Or a Kerberos implementation might lookup a host's realm and check that it matches the principal's realm. Naming constraints should be formalized for all GSS-API security mechanisms.

1.2. Conventions used in this document

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119].

2. Generic Attributes

We add a number of generic name attributes, to be used via the GSS-API extended naming facility [RFC6680]. Some of these attributes can be used as prefixes of other attributes, that is, they can be used to modify the semantics of other attributes (see section 6 of RFC6680).

Note: in all cases the display form of each attribute SHALL consist of text using the character set, codeset, and encoding from the caller's locale.

2.1. Concrete Attributes

These attributes generally have a single value each. Only one of these attributes can also be used a prefix: the issuer name attribute.

2.1.1. Issuer Name

We add an attribute by which to obtain a name of an issuer of a mechanism name (MN) or of an attribute of an MN. The API name for this attribute is GSS_C_ATTR_GENERIC_ISSUERNAME, and it's actual attribute name is “urn:ietf:id:ietf-kitten-name-attrs-00-issuername”.

The display form of issuer names is mechanism-specific.

The non-display form of issuer names SHALL be the exported name token form of the issuer's name. Not all mechanisms will support issuer names as MNs, therefore implementations MAY output a null non-display value.

For example, for the Kerberos mechanism [RFC4121] an issuer name would generally (but not always!) be a Kerberos realm name, probably displayed as just the realm name. (But note that there is not yet a Kerberos realm name as MN specification. We will specify one separately.)

This attribute can be used as prefix of other attributes. When used as a prefix, this attribute indicates that the application wishes to know the name of the issuer of the prefixed attribute of the given MN.

2.1.2. User Name

We add an attribute by which to obtain the component of an MN naming a user. The API name for this attribute is GSS_C_ATTR_GENERIC_USERNAME, and it's actual attribute name is “urn:ietf:id:ietf-kitten-name-attrs-00-username”.

The display form of user names is mechanism-specific.

The non-display form of user names is mechanism-specific.

2.1.3. Service Name

We add an attribute by which to obtain the component of an MN naming a service as part of a host- or domain-based service name. The API name for this attribute is GSS_C_ATTR_GENERIC_SERVICENAME, and it's actual attribute name is “urn:ietf:id:ietf-kitten-name-attrs-00-servicename”.

The non-display form of the service name SHALL be the UTF-8 encoding of the service name.

2.1.4. Host Name

We add an attribute by which to obtain the component of an MN naming a host as part of a host- or domain-based service name. The API name for this attribute is GSS_C_ATTR_GENERIC_HOSTNAME, and it's actual attribute name is “urn:ietf:id:ietf-kitten-name-attrs-00-hostname”.

The display form of a host name MAY be stylized and SHOULD NOT be A-labels. [RFC5890].

The non-display form of host names SHOULD be a character string as described in [RFC1123], and SHOULD NOT be U-labels [RFC5890].

2.1.5. Domain Name

We add an attribute by which to obtain the component of an MN naming a domain as part of a domain-based service name. The API name for this attribute is GSS_C_ATTR_GENERIC_DOMAINNAME, and it's actual attribute name is “urn:ietf:id:ietf-kitten-name-attrs-00-domainname”.

The display form of a domain name MAY be stylized and SHOULD NOT be A-labels. [RFC5890].

The non-display form of domain names SHOULD be a character string as described in [RFC1123], and SHOULD NOT be U-labels [RFC5890].

2.2. Prefix Attributes

GSS_Get_name_attribute() using attributes described in the preceding section SHALL fail if there are any name constraints that can be applied to the issuers of those names and, in applying those constraints, it is discovered that the issuer was not permitted to issue credentials for the MN.

For example, a Kerberos realm named “FOO.EXAMPLE” might not be expected to issue credentials (tickets, keys) to host-based service names for hosts not ending in “.foo.example” or which are not “foo.example”.

Several generic attribute prefixes are described below for overriding this behavior.

2.2.1. GSS_C_ATTR_GENERIC_UNCONSTRAINED

This attribute prefix, named GSS_C_ATTR_GENERIC_UNCONSTRAINED in the API, and with an actual name of “urn:ietf:id:ietf-kitten-name-attrs-00-gen-unconstrained”, indicates that the application wants the value of the prefixed attribute without any name constraint checking.

2.2.2. GSS_C_ATTR_GENERIC_UNCONSTRAINED_OK

This attribute prefix, named GSS_C_ATTR_GENERIC_UNCONSTRAINED_OK in the API, and with an actual name of “urn:ietf:id:ietf-kitten-name-attrs-00-gen-unconstrained-ok”, indicates that the application wants the value of the prefixed attribute regardless of any applicable naming constraints, but to indicate the name constraint status via the 'authenticated' output parameter of the GSS_Get_name_attribute() interface.

2.2.3. GSS_C_ATTR_GENERIC_FAST

This attribute prefix, named GSS_C_ATTR_GENERIC_FAST in the API, and with an actual name of “urn:ietf:id:ietf-kitten-name-attrs-00-gen-fast”, indicates that the application requires that the mechanism not perform any slow operations (e.g., connecting to a directory for the purposes of name constraint validation) in obtaining the prefixed attribute of the given MN.

3. Local Name Attributes

Normally an Internet specification would not be expected to specify any local name attributes of GSS names. However, there is one common and very useful local name attribute, which we specify below. Implementations are free to use different names for this attribute or exclude it altogether -- it is a local name attribute, after all.

3.1. GSS_C_ATTR_LOCAL_LOGIN_USER

This attribute, with suggested API symbolic name GSS_C_ATTR_LOCAL_LOGIN_USER, and suggested actual name “local-login-user”, requests a local user name corresponding to the given MN, if any.

Obtaining the local user name corresponding to an MN may require complex name mapping or lookup operations that are completely implementation-defined.

4. Suggested Mechanism-Specific Name Attributes (INFORMATIONAL)

4.1. Suggested Kerberos-Specific Name Attributes

• realm (corresponding to issuer name)
• component 0 (first component of a principal name)
• component 1 (second component of a principal name)
• ..
• component 9 (10th component of a principal name)
• components (ordered set of components of a principal name)
• transit path (ordered set of realm and CA names)
• specific authorization data elements
• PKINIT client certificate
• session key enctype
• enctypes involved in transit path (this would only be available to initiators)

4.2. Suggested PKU2U-Specific Name Attributes

• issuer CA name
• trust path to a trust anchor
• certificate
• certificate subject public key
• certificate subject name
• certificate subject alternate names
• specific certificate extensions
• certificate algorithm names
• session key enctype

5. Security Considerations

[Add text regarding name constraint checking and explaining the default-to-safe design of the generic name attributes defined in section 2.]

6. IANA Considerations

[Add text regarding the registration and assignment of the name attributes described in the preceding sections. In particular we should want these attributes' names to not reflect an Internet-Draft name, but an RFC number.]

7. References

7.1. Normative References

7.2. Informative References

Author's Address