[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Proposal for resolution of Issue 327: Applicability



The concerns expressed in Issue 327 appear to apply to Sections 1, 1.3, 3.1.2, 3.1.3 and 3.1.4.

Here is a proposed resolution.  Comments welcome.

Section 1

 

   This document provides guidelines for the design of RADIUS attributes

   both within the IETF as well as within other SDOs.  By articulating

   RADIUS design guidelines, it is hoped that this document will

   encourage the development and publication of high quality RADIUS

   attribute specifications.

 

   However, the advice in this document will not be helpful unless it is

   put to use.  As with "Guidelines for Authors and Reviewers of MIB

   Documents" [RFC4181], it is expected that this document will be used

   by authors to check their document against the guidelines prior to

   requesting review (such as an "Expert Review" described in

   [RFC3575]).  Similarly, it is expected that this document will used

   by reviewers (such as WG participants or the AAA Doctors [DOCTORS]),

   resulting in an improvement in the consistency of reviews.

 

   In order to meet these objectives, this document needs to cover not

   only the science of attribute design, but also the art.  Therefore,

   in addition to covering the most frequently encountered issues, this

   document attempts to provide some of the considerations motivating

   the guidelines.

 

   In order to characterize current attribute usage, both the basic and

   complex data types defined in the existing RADIUS RFCs are reviewed.

 

[BA] Proposed change is to replace Section 1 with the following:

 

   "This document provides guidelines for the design of RADIUS attributes

   used to encode service-provisioning or authentication data,

   based on the data model and attribute formats defined in RFC 2865 and

   subsequent RADIUS RFCs.   In order to characterize current attribute

   usage, both the basic and complex data types defined in the existing

   RADIUS RFCs are reviewed. Since the RADEXT WG is currently considering

   extensions to the RADIUS data model and attribute formats,

   future documents may extend the data model and guidelines provided here.

 

   In order to meet these objectives, this document needs to cover not

   only the science of attribute design, but also the art.  Therefore,

   in addition to covering the most frequently encountered issues, this

   document attempts to provide some of the considerations motivating

   the guidelines.

 

   RADIUS protocol changes, or specification of attributes

   (such as Service-Type) that can be used to, in effect, provide

   new RADIUS commands require greater expertise and deeper review,

   as do changes to the RADIUS operational model as discussed in Section

   3.3.  As a result, such changes are outside the scope of this

   document and MUST NOT be undertaken outside the IETF.  Registration policies

   for RADIUS parameters are discussed in [RFC3575] Section 2.1."

 

1.3.  Applicability

 

   As RADIUS has become more widely accepted as a management protocol,

   its usage has become more prevalent, both within the IETF as well as

   within other SDOs.  Given the expanded utilization of RADIUS, it has

   become apparent that requiring SDOs to accomplish all their RADIUS

   work within the IETF is inherently inefficient and unscalable.  By

   articulating guidelines for RADIUS attribute design, this document

   enables SDOs out of the IETF to design their own RADIUS attributes

   within the Vendor-Specific Attribute (VSA) space.

 

   It is RECOMMENDED that SDOs follow the guidelines articulated in this

   document.  Doing so will ensure the widest possible applicability and

   interoperability of the specifications, while requiring minimal

   changes to existing systems.  Specifications that do not follow the

   guidelines articulated herein are NOT RECOMMENDED.  However, we

   recognize that there are some situations where SDOs or vendors

   require the creation of specifications not following these

   guidelines.  We do not forbid these specifications, but it is

   RECOMMENDED that they are created only if they have a limited scope

   of applicability, and all attributes defined in those specifications

   are VSAs, as discussed Appendix A.5, below.

 

   It is RECOMMENDED that SDOs and vendors seek review of RADIUS

   attribute specifications from the IETF.  However, when specifications

   are SDO specific, re-use existing data types, and follow these

   guidelines, they do not require IETF review.

 

   In order to enable IETF review of such specifications, the authors

   recommend that:

 

      * SDOs make their RADIUS attribute specifications publicly

      available;

 

      * SDOs request review of RADIUS attribute specifications by

      sending email to the AAA Doctors [DOCTORS] or equivalent mailing

      list;

 

      * IETF and SDO RADIUS attribute specifications are reviewed

      according to the guidelines proposed in this document;

 

      * Reviews of specifications are posted to the RADEXT WG mailing

      list, the AAA Doctors mailing list [DOCTORS] or another IETF

      mailing list suggested by the Operations & Management Area

      Directors of the IETF.

 

   These reviews can assist with creation of specifications that meet

   the SDO requirements, and which are also compatible with the

   traditional data model and uses of RADIUS.  While these reviews

   require access to public specifications, the review process does not

   require publication of an IETF RFC.

 

   The advice in this document applies to attributes used to encode

   service-provisioning or authentication data.  RADIUS protocol

   changes, or specification of attributes (such as Service-Type) that

   can be used to, in effect, provide new RADIUS commands require

   greater expertise and deeper review, as do changes to the RADIUS

   operational model as discussed below in Section 3.3.  Such changes

   MUST NOT be undertaken outside the IETF and when handled within the

   IETF require "IETF Consensus" for adoption, as noted in [RFC3575]

   Section 2.1.

 

[BA] Proposed change is to replace Section 1.3 with the following:

 

   "   As RADIUS has become more widely accepted as a management protocol,

   its usage has become more prevalent, by vendors as well as the IETF and

  other SDOs.  By articulating RADIUS design guidelines, it is hoped that

   this document will encourage the development and publication of

   high quality RADIUS attribute specifications.  As with "Guidelines

   for Authors and Reviewers of MIB Documents" [RFC4181], it is

   expected that this document will be used by authors to check their

   document against the guidelines prior to requesting review

   (such as an "Expert Review" described in

   [RFC3575]).  Similarly, it is expected that this document will used

   by reviewers (such as WG participants or the AAA Doctors [DOCTORS]),

   resulting in an improvement in the consistency of reviews.

 

   The guidelines articulated in this document enable the design

   of attributes with the widest possible interoperability,

   while requiring minimal changes to existing systems.

   As a result, they are suitable for application to attributes

   designed for general usage.  It is RECOMMENDED

   that these guidelines be applied to review of documents

   requesting allocations within the IETF standard attribute

   space defined in [RFC2865].

 

   These guidelines may also prove useful in the design of

   attributes within the Vendor-Specific Attribute (VSA)

   space.  As noted in RFC 2865 Section 5.26, RADIUS VSAs

   were created "to allow vendors to support their own extended

   Attributes not suitable for general usage." 

   Rather than requesting allocation of standard attributes

   for those uses, [RFC2865] Section 6.2 notes that utilization

   of VSAs "should be encouraged instead of allocation of global

   attribute types, for functions specific only to one vendor's

   implementation of RADIUS, where no interoperability is deemed useful."

 

   However, experience has shown that attributes not originally

   designed for general usage can subsequently garner wide-spread

   deployment.  An example is the vendor-specific attributes

   defined in [RFC2548], which have been widely implemented

   within IEEE 802.11 Access Points.

 

   As a result, while SDOs and vendors MAY choose to create

   specifications not following these guidelines for use in

   scenarios within a limited scope of applicability,

   where general usage is possible,  adhering to the

   design guidelines is RECOMMENDED.  Discussion of

   vendor allocations is provided in Section 3.1.2;

   discussion of SDO allocations is provided in Section

   3.1.3.   SDOs and vendors desiring review of their

   specifications by the IETF are encouraged to follow

   the advice presented in Section 3.1.4."

 

3.1.2.  Vendor Allocations

 

   Vendor RADIUS Attribute specifications SHOULD allocate attributes

   from the vendor space, rather than requesting an allocation from the

   RADIUS standard attribute space.

 

   As discussed in [RFC2865] Section 5.26, the vendor space is intended

   for vendors to support their own Attributes not suitable for general

   use.  However, it is RECOMMENDED that vendors follow the guidelines

   outlined here, which are intended to enable maximum interoperability

   with minimal changes to existing systems.

 

   Following these guidelines means that RADIUS servers can be updated

   to support the vendor's equipment by editing a RADIUS dictionary.  If

   these guidelines are not followed, then the vendor's equipment can

   only be supported via code changes in RADIUS server implementations.

   Such code changes add complexity and delay to implementations.

 

[BA] Proposed change is to replace Section 3.1.2 with the following:

 

"  As noted in [RFC3575] Section 2.1, vendors are encouraged

   to utilize VSAs to define functions "specific only to one vendor's

   implementation of RADIUS, where no interoperability is deemed useful.

   For functions specific only to one vendor's implementation of RADIUS,

   the use of that should be encouraged instead of the allocation of

   global attribute types."

 

   Nevertheless, following the guidelines outlined within this document

   has several advantages, including maximum interoperability with

   minimal changes to existing systems.  Following these guidelines

   means that RADIUS servers can be updated to support the vendor's

   equipment by editing a RADIUS dictionary.  If these guidelines are

   not followed, then the vendor's equipment can only be supported

   via code changes in RADIUS server implementations.

   Such code changes add complexity and delay to implementations."

 

3.1.3.  SDO Allocations

 

   SDO RADIUS Attribute specifications SHOULD allocate attributes from

   the vendor space, rather than requesting an allocation from the

   RADIUS standard attribute space, for attributes matching any of the

   following criteria:

 

      * attributes relying on data types not defined within RADIUS

 

      * attributes intended primarily for use within an SDO

 

      * attributes intended primarily for use within a group of SDOs.

 

   Any new RADIUS attributes or values intended for interoperable use

   across a broad spectrum of the Internet Community SHOULD follow the

   normal IETF process, and SHOULD result in allocations from the RADIUS

   standard space.

 

   The recommendation for SDOs to allocate attributes from a vendor

   space rather than via the IETF process is a recognition that SDOs may

   desire to assert change control over their own RADIUS specifications.

   This change control can be obtained by requesting a PEC from the

   Internet Assigned Number Authority (IANA), for use as a Vendor-Id

   within a Vendor-Specific attribute.  Further allocation of attributes

   inside of the VSA space defined by that Vendor-Id is subject solely

   to the discretion of the SDO.  Similarly, the use of data types

   (complex or not) within that VSA space is solely under the discretion

   of the SDO.  It is RECOMMENDED that SDOs follow the guidelines

   outlined here, which are intended to enable maximum interoperability

   with minimal changes to existing systems.

 

   It should be understood that SDOs do not have to rehost VSAs into the

   standards space solely for the purpose of obtaining IETF review.

   Rehosting puts pressure on the standards space, and may be harmful to

   interoperability, since it can create two ways to provision the same

   service.  Rehosting may also require changes to the RADIUS data model

   which will affect implementations that do not intend to support the

   SDO specifications.

 

[BA] Proposed change is to replace Section 3.1.3 with the following:

 

"  Given the expanded utilization of RADIUS, it has

   become apparent that requiring SDOs to accomplish all their RADIUS

   work within the IETF is inherently inefficient and unscalable. 

 

  SDO RADIUS Attribute specifications SHOULD allocate attributes from

   the vendor space, rather than requesting an allocation from the

   RADIUS standard attribute space, for attributes matching any of the

   following criteria:

 

      * attributes relying on data types not defined within RADIUS

 

      * attributes intended primarily for use within an SDO

 

      * attributes intended primarily for use within a group of SDOs.

 

   Allocation of new RADIUS attributes or values intended for interoperable use

   across a broad spectrum of the Internet Community SHOULD follow the

   allocation process defined in [RFC3575].

 

   The recommendation for SDOs to allocate attributes from the vendor

   space rather than via the IETF process recognizes the need for SDOs

   to assert change control over their own RADIUS specifications.

   SDOs desiring to define their VSAs are encouraged to

   request allocation of a Private Enterprise Code (PEC) from the

   Internet Assigned Number Authority (IANA), for use as a Vendor-Id. 

   The SDO can then allocate attributes within the VSA space defined

   by that Vendor-Id, at their sole discretion.  Similarly, the use of

   data types (complex or otherwise) within that VSA space is solely

   under the discretion of the SDO."

 

3.1.4.  Publication of specifications

 

   SDOs are encouraged to seek early review of SSA specifications by the

   IETF.  This review may be initiated by sending mail to the AAA

   Doctors list [DOCTORS], with the understanding that this review is a

   voluntary-based service offered on best-effort basis.  Since the IETF

   is not a membership organization, in order to enable the RADIUS SSA

   specification to be reviewed, it is RECOMMENDED that it be made

   publicly available; this also encourages interoperability.  Where the

   RADIUS SSA specification is embedded within a larger document which

   cannot be made public, the RADIUS attribute and value definitions

   SHOULD be published instead as an Informational RFC, as with

   [RFC4679].  This process SHOULD be followed instead of requesting

   IANA allocations from within the standard RADIUS attribute space.

 

   Similarly, vendors are encouraged to make their specifications

   publicly available, for maximum interoperability.  However, it is not

   necessary for them to request publication of their VSA specifications

   as Informational RFCs by the IETF.

 

   All other specifications, including new authentication,

   authorization, and/or security mechanisms SHOULD be allocated via the

   standard RADIUS space, as noted in [RFC3575] Section 2.1.

 

[BA] Proposed change is to replace Section 3.1.4 with the following:

 

"  SDOs or vendors desiring review of their RADIUS specifications

   by the IETF are  encouraged to seek review as early as possible,

   so as to avoid potential delays.  Since reviews are handled by volunteers,

   responses are provided on a best-effort basis, with no service

   level guarantees.   In order to provide reviewers with access to the

   specification,  SDOs and vendors are encouraged to make them publicly available.

   Where the RADIUS specification is embedded within a larger document which

   cannot be made public, the RADIUS attribute and value definitions

   can be made available on a public web site or can be published as

   an Informational RFC, as with [RFC4679].

 

   Review can be requested by sending email to the AAA Doctors [DOCTORS]

   or equivalent mailing list.  The IETF Operations & Management

   Area Directors will then arrange for the review to be

   completed and posted to the AAA Doctors mailing list

   [DOCTORS], RADEXT WG mailing list, or other IETF mailing

   list. 

 

   The review process requires neither allocation of attributes

   within the IETF standard attribute space nor publication of an

   IETF RFC.  Requiring SDOs or vendors to rehost VSAs into the IETF standards

   attribute space solely for the purpose of obtaining review

   would put pressure on the standards space, and may be harmful to

   interoperability, since it can create two ways to provision the same

   service.  Rehosting may also require changes to the RADIUS data model

   which will affect implementations that do not intend to support the

   SDO or vendor specifications."