80 likes | 90 Views
Provides guidelines for designing RADIUS attributes, considering data types and backward compatibility across attribute classes. Includes milestones, issues, and recommendations for attribute design.
E N D
RADEXT WGRADIUS Attribute Guidelines draft-weber-radius-attr-guidelines-03.txt July 10, 2006 v1 IETF-66, Montréal
J F M A M J J A S O N D 2005 J F M A M J J A S O N D 2006 RADIUS Attribute Guidelines • WG Charter Item: “RADIUS design guidelines. This document will provide guidelines for design of RADIUS attributes. It will specifically consider how complex data types may be introduced in a robust manner, maintaining backwards compatibility with existing RADIUS RFCs, across all the classes of attributes: Standard, Vendor-Specific and SDO-Specific. In addition, it will review RADIUS data types and associated backwards compatibility issues.” • Milestone: Oct ’06 completionOriginally Dec ‘04 Guide-02 ExtAttr-00 Guide-01 Guide-03 Guide-00 DraftRevisions CharteredMilestones IESG Submissions WG LC1 IETF-66, Montréal LATE
RADIUS Attribute Guidelines Issues (one from the tracker, three from IETF-65 discussion): • #106: Vendor Specific Values – Closed?RFC 2882-style VSEs SHOULD NOT be used; follow 3575 instead • IETF-65: Per-attribute encryptionDon’t preclude future mechanisms (a la crypto agility); address current practice in Security Considerations • IETF-65: Independent SDO workEnable SDOs to work independently, so IETF is not blocking • IETF-65: Structured data categorizationCategorize structured data based on which entity consumes/parses the elements; consistent, explicit structure required only when this is a RADIUS entity. Current version diff: http://employees.org/~gdweber/02-03-diff.html IETF-66, Montréal
RADIUS Attribute Guidelines • Most of the content is not contentious Need consensus here IETF-66, Montréal
RADIUS Attribute Guidelines Section 7: new attributes SHOULD comply with the attribute design guidelines given in RFC 2865 unless one or more of the following applies: • The standard attribute space for new attributes has been exhausted. • The proposed maximum attribute length exceeds that available for attributes specified by RFC 2865. • The native data type of the data element is defined for Extended attribute, but not standard RADIUS, e.g. Integer64. • Structured data must be explicitly represented (see next) IETF-66, Montréal
RADIUS Attribute Guidelines Section 7: The guidance for when to use which mechanism to affect the logical grouping of data depends on the intended consumer of the attribute values: • Humans: consume attributes intended to aid in troubleshooting and operational problem determination. • Billing: accounting data are typically consumed by a business mediation layer in preparation for billing systems. • Non-RADIUS authorization applications: these entities include policy servers, EAP servers, and other consumers of data which are typically opaque to RADIUS servers and proxies. • RADIUS servers and proxies: parse and operate on individual elements within RADIUS messages. Attributes in the 4th category SHOULD explicitly represent their values using the format defined by the Extended Attribute draft. IETF-66, Montréal
RADIUS Attribute Guidelines Further recommendations: • The Vendor-Specific Enumeration (VSE) encoding mechanism as proposed by Section 2.2.1 of RFC 2882 SHOULD NOT be used. Instead, vendors should comply with the recommendations of RFC 3575. • The message lengths specified by RADIUS standards MUST NOT be exceeded. • Variable attribute content SHOULD NOT be specified. Separate attributes SHOULD be defined instead. • SDOs are RECOMMENDED to design attributes using the guidelines in this document ‘as if ’ they were destined for the standard attribute space, but of course, may continue to implement new attributes in their own attributes spaces. IETF-66, Montréal
RADIUS Attribute Guidelines Discussion IETF-66, Montréal