idnits 2.17.00 (12 Aug 2021) /tmp/idnits49718/draft-ietf-abfab-arch-07.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** The document seems to lack a both a reference to RFC 2119 and the recommended RFC 2119 boilerplate, even if it appears to use RFC 2119 keywords. RFC 2119 keyword, line 501: '...method to be used. The IdP SHOULD NOT...' RFC 2119 keyword, line 503: '... this case the IdP MUST accept a realm...' RFC 2119 keyword, line 1324: '...A infrastructure MAY hide the initiato...' RFC 2119 keyword, line 1811: '... by an IdP, care MUST be taken in the ...' Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (July 30, 2013) is 3216 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- -- Looks like a reference, but probably isn't: '1' on line 142 -- Looks like a reference, but probably isn't: '2' on line 749 == Unused Reference: 'RFC2903' is defined on line 1878, but no explicit reference was found in the text == Unused Reference: 'I-D.nir-tls-eap' is defined on line 1882, but no explicit reference was found in the text == Unused Reference: 'RFC4017' is defined on line 1905, but no explicit reference was found in the text == Unused Reference: 'RFC5106' is defined on line 1909, but no explicit reference was found in the text == Unused Reference: 'RFC2138' is defined on line 1925, but no explicit reference was found in the text == Unused Reference: 'RFC2904' is defined on line 1965, but no explicit reference was found in the text ** Obsolete normative reference: RFC 3588 (Obsoleted by RFC 6733) == Outdated reference: draft-ietf-abfab-gss-eap has been published as RFC 7055 == Outdated reference: draft-ietf-abfab-aaa-saml has been published as RFC 7833 == Outdated reference: draft-ietf-radext-nai has been published as RFC 7542 == Outdated reference: draft-ietf-oauth-v2 has been published as RFC 6749 == Outdated reference: draft-iab-privacy-considerations has been published as RFC 6973 == Outdated reference: A later version (-06) exists of draft-perez-radext-radius-fragmentation-05 -- Obsolete informational reference (is this intentional?): RFC 2138 (Obsoleted by RFC 2865) -- Obsolete informational reference (is this intentional?): RFC 5849 (Obsoleted by RFC 6749) == Outdated reference: draft-ietf-emu-crypto-bind has been published as RFC 7029 -- No information found for draft-ietf-emu-eap-tunnel-method - is the name correct? == Outdated reference: draft-ietf-radext-dtls has been published as RFC 7360 == Outdated reference: draft-ietf-radext-dynamic-discovery has been published as RFC 7585 Summary: 2 errors (**), 0 flaws (~~), 16 warnings (==), 6 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 ABFAB J. Howlett 3 Internet-Draft JANET(UK) 4 Intended status: Informational S. Hartman 5 Expires: January 31, 2014 Painless Security 6 H. Tschofenig 7 Nokia Siemens Networks 8 E. Lear 9 Cisco Systems GmbH 10 J. Schaad 11 Soaring Hawk Consulting 12 July 30, 2013 14 Application Bridging for Federated Access Beyond Web (ABFAB) 15 Architecture 16 draft-ietf-abfab-arch-07.txt 18 Abstract 20 Over the last decade a substantial amount of work has occurred in the 21 space of federated access management. Most of this effort has 22 focused on two use cases: network access and web-based access. 23 However, the solutions to these use cases that have been proposed and 24 deployed tend to have few common building blocks in common. 26 This memo describes an architecture that makes use of extensions to 27 the commonly used security mechanisms for both federated and non- 28 federated access management, including the Remote Authentication Dial 29 In User Service (RADIUS) and the Diameter protocol, the Generic 30 Security Service (GSS), the Extensible Authentication Protocol (EAP) 31 and the Security Assertion Markup Language (SAML). The architecture 32 addresses the problem of federated access management to primarily 33 non-web-based services, in a manner that will scale to large numbers 34 of identity providers, relying parties, and federations. 36 Status of This Memo 38 This Internet-Draft is submitted in full conformance with the 39 provisions of BCP 78 and BCP 79. 41 Internet-Drafts are working documents of the Internet Engineering 42 Task Force (IETF). Note that other groups may also distribute 43 working documents as Internet-Drafts. The list of current Internet- 44 Drafts is at http://datatracker.ietf.org/drafts/current/. 46 Internet-Drafts are draft documents valid for a maximum of six months 47 and may be updated, replaced, or obsoleted by other documents at any 48 time. It is inappropriate to use Internet-Drafts as reference 49 material or to cite them other than as "work in progress." 51 This Internet-Draft will expire on January 31, 2014. 53 Copyright Notice 55 Copyright (c) 2013 IETF Trust and the persons identified as the 56 document authors. All rights reserved. 58 This document is subject to BCP 78 and the IETF Trust's Legal 59 Provisions Relating to IETF Documents 60 (http://trustee.ietf.org/license-info) in effect on the date of 61 publication of this document. Please review these documents 62 carefully, as they describe your rights and restrictions with respect 63 to this document. Code Components extracted from this document must 64 include Simplified BSD License text as described in Section 4.e of 65 the Trust Legal Provisions and are provided without warranty as 66 described in the Simplified BSD License. 68 Table of Contents 70 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 71 1.1. Terminology . . . . . . . . . . . . . . . . . . . . . . . 5 72 1.1.1. Channel Binding . . . . . . . . . . . . . . . . . . . 6 73 1.2. An Overview of Federation . . . . . . . . . . . . . . . . 7 74 1.3. Challenges for Contemporary Federation . . . . . . . . . 10 75 1.4. An Overview of ABFAB-based Federation . . . . . . . . . . 11 76 1.5. Design Goals . . . . . . . . . . . . . . . . . . . . . . 14 77 2. Architecture . . . . . . . . . . . . . . . . . . . . . . . . 14 78 2.1. Relying Party to Identity Provider . . . . . . . . . . . 16 79 2.1.1. AAA, RADIUS and Diameter . . . . . . . . . . . . . . 17 80 2.1.2. Discovery and Rules Determination . . . . . . . . . . 18 81 2.1.3. Routing and Technical Trust . . . . . . . . . . . . . 19 82 2.1.4. AAA Security . . . . . . . . . . . . . . . . . . . . 20 83 2.1.5. SAML Assertions . . . . . . . . . . . . . . . . . . . 21 84 2.2. Client To Identity Provider . . . . . . . . . . . . . . . 23 85 2.2.1. Extensible Authentication Protocol (EAP) . . . . . . 23 86 2.2.2. EAP Channel Binding . . . . . . . . . . . . . . . . . 25 87 2.3. Client to Relying Party . . . . . . . . . . . . . . . . . 25 88 2.3.1. GSS-API . . . . . . . . . . . . . . . . . . . . . . . 26 89 2.3.2. Protocol Transport . . . . . . . . . . . . . . . . . 27 90 2.3.3. Reauthentication . . . . . . . . . . . . . . . . . . 27 91 3. Application Security Services . . . . . . . . . . . . . . . . 28 92 3.1. Authentication . . . . . . . . . . . . . . . . . . . . . 28 93 3.2. GSS-API Channel Binding . . . . . . . . . . . . . . . . . 29 94 3.3. Host-Based Service Names . . . . . . . . . . . . . . . . 30 95 3.4. Additional GSS-API Services . . . . . . . . . . . . . . . 32 96 4. Privacy Considerations . . . . . . . . . . . . . . . . . . . 32 97 4.1. Entities and their roles . . . . . . . . . . . . . . . . 33 98 4.2. Privacy Aspects of ABFAB Communication Flows . . . . . . 34 99 4.2.1. Client to RP . . . . . . . . . . . . . . . . . . . . 34 100 4.2.2. Client to IdP (via Federation Substrate) . . . . . . 35 101 4.2.3. IdP to RP (via Federation Substrate) . . . . . . . . 36 102 4.3. Relationship between User and Entities . . . . . . . . . 37 103 4.4. Accounting Information . . . . . . . . . . . . . . . . . 37 104 4.5. Collection and retention of data and identifiers . . . . 37 105 4.6. User Participation . . . . . . . . . . . . . . . . . . . 38 106 5. Security Considerations . . . . . . . . . . . . . . . . . . . 38 107 6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 39 108 7. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 39 109 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 40 110 8.1. Normative References . . . . . . . . . . . . . . . . . . 40 111 8.2. Informative References . . . . . . . . . . . . . . . . . 41 112 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 43 114 1. Introduction 116 The Internet uses numerous security mechanisms to manage access to 117 various resources. These mechanisms have been generalized and scaled 118 over the last decade through mechanisms such as Simple Authentication 119 and Security Layer (SASL) with the Generic Security Server 120 Application Program Interface (GSS-API) (known as the GS2 family) 121 [RFC5801], Security Assertion Markup Language (SAML) 122 [OASIS.saml-core-2.0-os], and the Authentication, Authorization, and 123 Accounting (AAA) architecture as embodied in RADIUS [RFC2865] and 124 Diameter [RFC3588]. 126 A Relying Party (RP) is the entity that manages access to some 127 resource. The entity that is requesting access to that resource is 128 often described as the Client. Many security mechanisms are 129 manifested as an exchange of information between these entities. The 130 RP is therefore able to decide whether the Client is authorized, or 131 not. 133 Some security mechanisms allow the RP to delegate aspects of the 134 access management decision to an entity called the Identity Provider 135 (IdP). This delegation requires technical signaling, trust and a 136 common understanding of semantics between the RP and IdP. These 137 aspects are generally managed within a relationship known as a 138 'federation'. This style of access management is accordingly 139 described as 'federated access management'. 141 Federated access management has evolved over the last decade through 142 specifications like SAML [OASIS.saml-core-2.0-os], OpenID [1], OAuth 143 [RFC5849], [I-D.ietf-oauth-v2] and WS-Trust [WS-TRUST]. The benefits 144 of federated access management include: 146 Single or Simplified sign-on: 148 An Internet service can delegate access management, and the 149 associated responsibilities such as identity management and 150 credentialing, to an organization that already has a long-term 151 relationship with the Client. This is often attractive as Relying 152 Parties frequently do not want these responsibilities. The Client 153 also requires fewer credentials, which is also desirable. 155 Data Minimization and User Participation: 157 Often a Relying Party does not need to know the identity of a 158 Client to reach an access management decision. It is frequently 159 only necessary for the Relying Party know specific attributes 160 about the client, for example, that the client is affiliated with 161 a particular organization or has a certain role or entitlement. 162 Sometimes the RP only needs to know a pseudonym of the client. 164 Prior to the release of attributes to the RP from the IdP, the IdP 165 will check configuration and policy to determine if the attributes 166 are to be released. There is currently no direct client 167 participation in this decision. 169 Provisioning: 171 Sometimes a Relying Party needs, or would like, to know more about 172 a client than an affiliation or a pseudonym. For example, a 173 Relying Party may want the Client's email address or name. Some 174 federated access management technologies provide the ability for 175 the IdP to supply this information, either on request by the RP or 176 unsolicited. 178 This memo describes the Application Bridging for Federated Access 179 Beyond the Web (ABFAB) architecture. This architecture makes use of 180 extensions to the commonly used security mechanisms for both 181 federated and non-federated access management, including the RADIUS 182 and the Diameter protocols, the Generic Security Service (GSS), the 183 Extensible Authentication Protocol (EAP) and SAML. The architecture 184 addresses the problem of federated access management primarily for 185 non-web-based services. It does so in a manner that will scale to 186 large numbers of identity providers, relying parties, and 187 federations. 189 1.1. Terminology 191 This document uses identity management and privacy terminology from 192 [I-D.iab-privacy-considerations]. In particular, this document uses 193 the terms identity provider, relying party, identifier, pseudonymity, 194 unlinkability, and anonymity. 196 In this architecture the IdP consists of the following components: an 197 EAP server, a RADIUS or a Diameter server, and optionally a SAML 198 Assertion service. 200 This document uses the term Network Access Identifier (NAI), as 201 defined in [I-D.ietf-radext-nai]. An NAI consists of a realm 202 identifier, which is associated with an IdP and a username which is 203 associated with a specific client of the IdP. 205 One of the problems people will find with reading this document is 206 that the terminology sometimes appears to be inconsistent. This is 207 due the fact that the terms used by the different standards we are 208 referencing are not consistent. In general the document uses either 209 a the ABFAB term or the term associated with the standard under 210 discussion as appropriate. For reference we include this table which 211 maps the different terms into a single table. 213 +--------------+--------------+-----------------+-------------------+ 214 | Protocol | Client | Relying Party | Identity Provider | 215 +--------------+--------------+-----------------+-------------------+ 216 | ABFAB | Client | Relying Party | Identity Provider | 217 | | | (RP) | (IdP) | 218 | | | | | 219 | | Initiator | Acceptor | | 220 | | | | | 221 | | | Server | | 222 | | | | | 223 | SAML | Subject | Service | Issuer | 224 | | | Provider | | 225 | | | | | 226 | GSS-API | Initiator | Acceptor | | 227 | | | | | 228 | EAP | EAP peer | | EAP server | 229 | | | | | 230 | AAA | | AAA Client | AAA server | 231 | | | | | 232 | RADIUS | user | NAS | RADIUS server | 233 | | | | | 234 | | | RADIUS client | | 235 +--------------+--------------+-----------------+-------------------+ 236 Note that in some cases a cell has been left empty; in these cases 237 there is no name that represents the entity. 239 1.1.1. Channel Binding 241 This document uses the term channel binding with two different 242 meanings. 244 EAP channel binding is used to provide GSS-API naming semantics. 245 Channel binding sends a set of attributes from the peer to the EAP 246 server either as part of the EAP conversation or as part of a secure 247 association protocol. In addition, attributes are sent in the 248 backend protocol from the authenticator to the EAP server. The EAP 249 server confirms the consistency of these attributes and provides the 250 confirmation back to the peer. In this document, channel binding 251 without qualification refers to EAP channel binding. 253 GSS-API channel binding provides protection against man-in-the-middle 254 attacks when GSS-API is used for authentication inside of some 255 tunnel; it is similar to a facility called cryptographic binding in 256 EAP. The binding works by each side deriving a cryptographic value 257 from the tunnel itself and then using that cryptographic value to 258 prove to the other side that it knows the value. 260 See [RFC5056] for a discussion of the differences between these two 261 facilities. However, the difference can be summarized as GSS-API 262 channel binding says that there is nobody between the client and the 263 authenticator while EAP channel binding allows the client to have 264 knowledge about attributes of the authenticator (such as it's name). 266 Typically when considering channel binding, people think of channel 267 binding in combination with mutual authentication. This is 268 sufficiently common that without additional qualification channel 269 binding should be assumed to imply mutual authentication. Without 270 mutual authentication, only one party knows that the endpoints are 271 correct. That's sometimes useful. Consider for example a user who 272 wishes to access a protected resource from a shared whiteboard in a 273 conference room. The whiteboard is the initiator; it does not need 274 to actually authenticate that it is talking to the correct resource 275 because the user will be able to recognize whether the displayed 276 content is correct. If channel binding is used without mutual 277 authentication, it is effectively a request to disclose the resource 278 in the context of a particular channel. Such an authentication would 279 be similar in concept to a holder-of-key SAML assertion. However, 280 also note that while it is not happening in the protocol, mutual 281 authentication is happening in the overall system: the user is able 282 to visually authenticate the content. This is consistent with all 283 uses of channel binding without protocol level mutual authentication 284 found so far. 286 1.2. An Overview of Federation 288 In the previous section we introduced the following entities: 290 o the Client, 292 o the Identity Provider, and 294 o the Relying Party. 296 The final entity that needs to be introduced is the Individual. An 297 Individual is a human being that is using the Client. In any given 298 situation, an Individual may or may not exist. Clients can act 299 either as front ends for Individuals or they may be independent 300 entities that are setup and allowed to run autonomously. An example 301 of such an entity can be found in the trust routing protocol where 302 the routers use ABFAB to authenticate to each other. 304 These entities and their relationships are illustrated graphically in 305 Figure 1. 307 ,----------\ ,---------\ 308 | Identity | Federation | Relying | 309 | Provider + <-------------------> + Party | 310 `----------' '---------' 311 < 312 \ 313 \ Authentication 314 \ 315 \ 316 \ 317 \ 318 \ +---------+ 319 \ | | O 320 v| Client | \|/ Individual 321 | | | 322 +---------+ / \ 324 Figure 1: Entities and their Relationships 326 The relationships between the entities in Figure 1 are: 328 Federation 330 The Identity Provider and the Relying Parties are part of a 331 Federation. The relationship may be direct (they have an explicit 332 trust relationship) or transitive (the trust relationship is 333 mediated by one or more entities). The federation relationship is 334 governed by a federation agreement. Within a single federation, 335 there may be multiple Identity Providers as well as multiple 336 Relying Parties. A federation is governed by a federation 337 agreement. 339 Authentication 341 There is a direct relationship between the Client and the Identity 342 Provider by which the entities trust and can securely authenticate 343 each other. 345 A federation agreement typically encompasses operational 346 specifications and legal rules: 348 Operational Specifications: 350 These include the technical specifications (e.g. protocols used to 351 communicate between the three parties), process standards, 352 policies, identity proofing, credential and authentication 353 algorithm requirements, performance requirements, assessment and 354 audit criteria, etc. The goal of operational specifications is to 355 provide enough definition that the system works and 356 interoperability is possible. 358 Legal Rules: 360 The legal rules take the legal framework into consideration and 361 provide contractual obligations for each entity. The rules define 362 the responsibilities of each party and provide further 363 clarification of the operational specifications. These legal 364 rules regulate the operational specifications, make operational 365 specifications legally binding to the participants, define and 366 govern the rights and responsibilities of the participants. The 367 legal rules may, for example, describe liability for losses, 368 termination rights, enforcement mechanisms, measures of damage, 369 dispute resolution, warranties, etc. 371 The Operational Specifications can demand the usage of a 372 sophisticated technical infrastructure, including requirements on the 373 message routing intermediaries, to offer the required technical 374 functionality. In other environments, the Operational Specifications 375 require fewer technical components in order to meet the required 376 technical functionality. 378 The Legal Rules include many non-technical aspects of federation, 379 such as business practices and legal arrangements, which are outside 380 the scope of the IETF. The Legal Rules can still have an impact on 381 the architectural setup or on how to ensure the dynamic establishment 382 of trust. 384 While a federation agreement is often discussed within the context of 385 formal relationships, such as between an enterprise and an employee 386 or a government and a citizen, a federation agreement does not have 387 to require any particular level of formality. For an IdP and a 388 Client, it is sufficient for a relationship to be established by 389 something as simple as using a web form and confirmation email. For 390 an IdP and an RP, it is sufficient for the IdP to publish contact 391 information along with a public key and for the RP to use that data. 392 Within the framework of ABFAB, it will generally be required that a 393 mechanism exists for the IdP to be able to trust the identity of the 394 RP, if this is not present then the IdP cannot provide the assurances 395 to the client that the identity of the RP has been established. 397 The nature of federation dictates that there is some form of 398 relationship between the identity provider and the relying party. 399 This is particularly important when the relying party wants to use 400 information obtained from the identity provider for access management 401 decisions and when the identity provider does not want to release 402 information to every relying party (or only under certain 403 conditions). 405 While it is possible to have a bilateral agreement between every IdP 406 and every RP; on an Internet scale this setup requires the 407 introduction of the multi-lateral federation concept, as the 408 management of such pair-wise relationships would otherwise prove 409 burdensome. 411 The IdP will typically have a long-term relationship with the Client. 412 This relationship typically involves the IdP positively identifying 413 and credentialing the Client (for example, at time of employment 414 within an organization). When dealing with individuals, this process 415 is called identity proofing [NIST-SP.800-63]. The relationship will 416 often be instantiated within an agreement between the IdP and the 417 Client (for example, within an employment contract or terms of use 418 that stipulates the appropriate use of credentials and so forth). 420 The nature and quality of the relationship between the Client and the 421 IdP is an important contributor to the level of trust that an RP may 422 attribute to an assertion describing a Client made by an IdP. This 423 is sometimes described as the Level of Assurance [NIST-SP.800-63]. 425 Federation does not require an a priori relationship or a long-term 426 relationship between the RP and the Client; it is this property of 427 federation that yields many of its benefits. However, federation 428 does not preclude the possibility of a pre-existing relationship 429 between the RP and the Client, nor that they may use the introduction 430 to create a new long-term relationship independent of the federation. 432 Finally, it is important to reiterate that in some scenarios there 433 might indeed be an Individual behind the Client and in other cases 434 the Client may be autonomous. 436 1.3. Challenges for Contemporary Federation 438 As the number of federated services has proliferated, the role of the 439 individual can become ambiguous in certain circumstances. For 440 example, a school might provide online access for a student's grades 441 to their parents for review, and to the student's teacher for 442 modification. A teacher who is also a parent must clearly 443 distinguish her role upon access. 445 Similarly, as the number of federations proliferates, it becomes 446 increasingly difficult to discover which identity provider(s) a user 447 is associated with. This is true for both the web and non-web case, 448 but is particularly acute for the latter as many non-web 449 authentication systems are not semantically rich enough on their own 450 to allow for such ambiguities. For instance, in the case of an email 451 provider, the use of SMTP and IMAP protocols do not have the ability 452 for the server to get additional information, beyond the clients NAI, 453 in order to provide additional input to decide between multiple 454 federations it may be associated with. However, the building blocks 455 do exist to add this functionality. 457 1.4. An Overview of ABFAB-based Federation 459 The previous section described the general model of federation, and 460 the application of access management within the federation. This 461 section provides a brief overview of ABFAB in the context of this 462 model. 464 In this example, a client is attempting to connect to a server in 465 order to either get access to some data or perform some type of 466 transaction. In order for the client to mutually authenticate with 467 the server, the following steps are taken in an ABFAB federated 468 architecture: 470 1. Client Configuration: The Client Application is configured with 471 an NAI assigned by the IdP. It is also configured with any 472 keys, certificates, passwords or other secret and public 473 information needed to run the EAP protocols between it and the 474 IdP. 476 2. Authentication mechanism selection: The GSS-EAP GSS-API 477 mechanism is selected for authentication/authorization. 479 3. Client provides an NAI to RP: The client application sets up a 480 transport to the RP and begins the GSS-EAP authentication. In 481 response, the RP sends an EAP request message (nested in the 482 GSS-EAP protocol) asking for the Client's name. The Client 483 sends an EAP response with an NAI name form that, at a minimum, 484 contains the realm portion of its full NAI. 486 4. Discovery of federated IdP: The RP uses pre-configured 487 information or a federation proxy to determine what IdP to use 488 based on policy and the realm portion of the provided Client 489 NAI. This is discussed in detail below (Section 2.1.2). 491 5. Request from Relying Party to IdP: Once the RP knows who the IdP 492 is, it (or its agent) will send a RADIUS/Diameter request to the 493 IdP. The RADIUS/Diameter access request encapsulates the EAP 494 response. At this stage, the RP will likely have no idea who 495 the client is. The RP sends its identity to the IdP in AAA 496 attributes, and it may send a SAML Attribute Requests in a AAA 497 attribute. The AAA network checks that the identity claimed by 498 the RP is valid. 500 6. IdP begins EAP with the client: The IdP sends an EAP message to 501 the client with an EAP method to be used. The IdP SHOULD NOT 502 re-request the clients name in this message, but clients need to 503 be able to handle it. In this case the IdP MUST accept a realm 504 only in order to protect the client's name from the RP. The 505 available and appropriate methods are discussed below in this 506 memo (Section 2.2.1). 508 7. The EAP protocol is run: A bunch of EAP messages are passed 509 between the client (EAP peer) and the IdP (EAP server), until 510 the result of the authentication protocol is determined. The 511 number and content of those messages depends on the EAP method 512 selected. If the IdP is unable to authenticate the client, the 513 IdP sends a EAP failure message to the RP. As part of the EAP 514 protocol, the client sends a channel bindings EAP message to the 515 IdP (Section 2.2.2). In the channel binding message the client 516 identifies, among other things, the RP to which it is attempting 517 to authenticate. The IdP checks the channel binding data from 518 the client with that provided by the RP via the AAA protocol. 519 If the bindings do not match the IdP sends an EAP failure 520 message to the RP. 522 8. Successful EAP Authentication: At this point, the IdP (EAP 523 server) and client (EAP peer) have mutually authenticated each 524 other. As a result, the client and the IdP hold two 525 cryptographic keys: a Master Session Key (MSK), and an Extended 526 MSK (EMSK). At this point the client has a level of assurance 527 about the identity of the RP based on the name checking the IdP 528 has done using the RP naming information from the AAA framework 529 and from the client (by the channel binding data). 531 9. Local IdP Policy Check: At this stage, the IdP checks local 532 policy to determine whether the RP and client are authorized for 533 a given transaction/service, and if so, what if any, attributes 534 will be released to the RP. If the IdP gets a policy failure, 535 it sends an EAP failure message to the RP.[[Should this be an 536 EAP failure to the client as well?]] (The RP will have done its 537 policy checks during the discovery process.) 539 10. IdP provide the RP with the MSK: The IdP sends a positive result 540 EAP to the RP, along with an optional set of AAA attributes 541 associated with the client (usually as one or more SAML 542 assertions). In addition, the EAP MSK is returned to the RP. 544 11. RP Processes Results: When the RP receives the result from the 545 IdP, it should have enough information to either grant or refuse 546 a resource access request. It may have information that 547 associates the client with specific authorization identities. 548 If additional attributes are needed from the IdP the RP may make 549 a new SAML Request to the IdP. It will apply these results in 550 an application-specific way. 552 12. RP returns results to client: Once the RP has a response it must 553 inform the client application of the result. If all has gone 554 well, all are authenticated, and the application proceeds with 555 appropriate authorization levels. The client can now complete 556 the authentication of the RP by the use of the EAP MSK value. 558 An example communication flow is given below: 560 Relying Client Identity 561 Party App Provider 563 | (1) | Client Configuration 564 | | | 565 |<-----(2)----->| | Mechanism Selection 566 | | | 567 |<-----(3)-----<| | NAI transmitted to RP 568 | | | 569 |<=====(4)====================>| Discovery 570 | | | 571 |>=====(5)====================>| Access request from RP to IdP 572 | | | 573 | |< - - (6) - -<| EAP method to Client 574 | | | 575 | |< - - (7) - ->| EAP Exchange to authenticate 576 | | | Client 577 | | | 578 | | (8 & 9) Local Policy Check 579 | | | 580 |<====(10)====================<| IdP Assertion to RP 581 | | | 582 (11) | | RP processes results 583 | | | 584 |>----(12)----->| | Results to client app. 586 ----- = Between Client App and RP 587 ===== = Between RP and IdP 588 - - - = Between Client App and IdP 590 1.5. Design Goals 592 Our key design goals are as follows: 594 o Each party of a transaction will be authenticated, although 595 perhaps not identified, and the client will be authorized for 596 access to a specific resource. 598 o Means of authentication is decoupled so as to allow for multiple 599 authentication methods. 601 o The architecture requires no sharing of long term private keys 602 between clients and servers. 604 o The system will scale to large numbers of identity providers, 605 relying parties, and users. 607 o The system will be designed primarily for non-Web-based 608 authentication. 610 o The system will build upon existing standards, components, and 611 operational practices. 613 Designing new three party authentication and authorization protocols 614 is hard and fraught with risk of cryptographic flaws. Achieving 615 widespread deployment is even more difficult. A lot of attention on 616 federated access has been devoted to the Web. This document instead 617 focuses on a non-Web-based environment and focuses on those protocols 618 where HTTP is not used. Despite the increased excitement for 619 layering every protocol on top of HTTP there are still a number of 620 protocols available that do not use HTTP-based transports. Many of 621 these protocols are lacking a native authentication and authorization 622 framework of the style shown in Figure 1. 624 2. Architecture 626 We have already introduced the federated access architecture, with 627 the illustration of the different actors that need to interact, but 628 did not expand on the specifics of providing support for non-Web 629 based applications. This section details this aspect and motivates 630 design decisions. The main theme of the work described in this 631 document is focused on re-using existing building blocks that have 632 been deployed already and to re-arrange them in a novel way. 634 Although this architecture assumes updates to the relying party, the 635 client application, and the Identity Provider, those changes are kept 636 at a minimum. A mechanism that can demonstrate deployment benefits 637 (based on ease of update of existing software, low implementation 638 effort, etc.) is preferred and there may be a need to specify 639 multiple mechanisms to support the range of different deployment 640 scenarios. 642 There are a number of ways for encapsulating EAP into an application 643 protocol. For ease of integration with a wide range of non-Web based 644 application protocols the usage of the GSS-API was chosen. A 645 description of the technical specification can be found in 646 [I-D.ietf-abfab-gss-eap]. 648 The architecture consists of several building blocks, which is shown 649 graphically in Figure 2. In the following sections, we discuss the 650 data flow between each of the entities, the protocols used for that 651 data flow and some of the trade-offs made in choosing the protocols. 653 +--------------+ 654 | Identity | 655 | Provider | 656 | (IdP) | 657 +-^----------^-+ 658 * EAP o RADIUS/ 659 * o Diameter 660 --v----------v-- 661 /// \\\ 662 // \\ 663 | Federation | 664 | Substrate | 665 \\ // 666 \\\ /// 667 --^----------^-- 668 * EAP o RADIUS/ 669 * o Diameter 670 +-------------+ +-v----------v--+ 671 | | | | 672 | Client | EAP/EAP Method | Relying Party | 673 | Application |<****************>| (RP) | 674 | | GSS-API | | 675 | |<---------------->| | 676 | | Application | | 677 | | Protocol | | 678 | |<================>| | 679 +-------------+ +---------------+ 681 Legend: 683 <****>: Client-to-IdP Exchange 684 <---->: Client-to-RP Exchange 685 : RP-to-IdP Exchange 686 <====>: Protocol through which GSS-API/GS2 exchanges are tunneled 688 Figure 2: ABFAB Protocol Instantiation 690 2.1. Relying Party to Identity Provider 692 Communications between the Relying Party and the Identity Provider is 693 done by the federation substrate. This communication channel is 694 responsible for: 696 o Establishing the trust relationship between the RP and the IdP. 698 o Determining the rules governing the relationship. 700 o Conveying authentication packets from the client to the IdP and 701 back. 703 o Providing the means of establishing a trust relationship between 704 the RP and the client. 706 o Providing a means for the RP to obtain attributes about the client 707 from the IdP. 709 The ABFAB working group has chosen the AAA framework for the messages 710 transported between the RP and IdP. The AAA framework supports the 711 requirements stated above as follows: 713 o The AAA backbone supplies the trust relationship between the RP 714 and the IdP. 716 o The agreements governing a specific AAA backbone contains the 717 rules governing the relationships within the AAA federation. 719 o A method exists for carrying EAP packets within RADIUS [RFC3579] 720 and Diameter [RFC4072]. 722 o The use of EAP channel binding [RFC6677] along with the core ABFAB 723 protocol provide the pieces necessary to establish the identities 724 of the RP and the client, while EAP provides the cryptographic 725 methods for the RP and the client to validate they are talking to 726 each other. 728 o A method exists for carrying SAML packets within RADIUS 729 [I-D.ietf-abfab-aaa-saml] and Diameter (work in progress) which 730 allows the RP to query attributes about the client from the IdP. 732 Future protocols that support the same framework but do different 733 routing may be used in the future. One such effort is to setup a 734 framework that creates a trusted point-to-point channel on the fly. 736 2.1.1. AAA, RADIUS and Diameter 738 Interestingly, for network access authentication the usage of the AAA 739 framework with RADIUS [RFC2865] and Diameter [RFC3588] was quite 740 successful from a deployment point of view. To map to the 741 terminology used in Figure 1 to the AAA framework the IdP corresponds 742 to the AAA server, the RP corresponds to the AAA client, and the 743 technical building blocks of a federation are AAA proxies, relays and 744 redirect agents (particularly if they are operated by third parties, 745 such as AAA brokers and clearing houses). The front-end, i.e. the 746 end host to AAA client communication, is in case of network access 747 authentication offered by link layer protocols that forward 748 authentication protocol exchanges back-and-forth. An example of a 749 large scale RADIUS-based federation is EDUROAM [2]. 751 By using the AAA framework, ABFAB gets a lot of mileage as many of 752 the federation agreements already exist and merely need to be 753 expanded to cover the ABFAB additions. The AAA framework has already 754 addressed some of the problems outlined above. For example, 756 o It already has a method for routing requests based on a domain. 758 o It already has an extensible architecture allowing for new 759 attributes to be defined and transported. 761 o Pre-existing relationships can be re-used. 763 The astute reader will notice that RADIUS and Diameter have 764 substantially similar characteristics. Why not pick one? RADIUS and 765 Diameter are deployed in different environments. RADIUS can often be 766 found in enterprise and university networks, and is also in use by 767 fixed network operators. Diameter, on the other hand, is deployed by 768 mobile operators. Another key difference is that today RADIUS is 769 largely transported upon UDP. We leave as a deployment decision, 770 which protocol will be appropriate. The protocol defines all the 771 necessary new AAA attributes as RADIUS attributes. A future document 772 would define the same AAA attributes for a Diameter environment. We 773 also note that there exist proxies which convert from RADIUS to 774 Diameter and back. This makes it possible for both to be deployed in 775 a single federation substrate. 777 Through the integrity protection mechanisms in the AAA framework, the 778 identity provider can establish technical trust that messages are 779 being sent by the appropriate relying party. Any given interaction 780 will be associated with one federation at the policy level. The 781 legal or business relationship defines what statements the identity 782 provider is trusted to make and how these statements are interpreted 783 by the relying party. The AAA framework also permits the relying 784 party or elements between the relying party and identity provider to 785 make statements about the relying party. 787 The AAA framework provides transport for attributes. Statements made 788 about the client by the identity provider, statements made about the 789 relying party and other information are transported as attributes. 791 One demand that the AAA substrate makes of the upper layers is that 792 they must properly identify the end points of the communication. It 793 must be possible for the AAA client at the RP to determine where to 794 send each RADIUS or Diameter message. Without this requirement, it 795 would be the RP's responsibility to determine the identity of the 796 client on its own, without the assistance of an IdP. This 797 architecture makes use of the Network Access Identifier (NAI), where 798 the IdP is indicated by the realm component [I-D.ietf-radext-nai]. 799 The NAI is represented and consumed by the GSS-API layer as 800 GSS_C_NT_USER_NAME as specified in [RFC2743]. The GSS-API EAP 801 mechanism includes the NAI in the EAP Response/Identity message. 803 2.1.2. Discovery and Rules Determination 805 While we are using the AAA protocols to communicate with the IdP, the 806 RP may have multiple federation substrates to select from. The RP 807 has a number of criteria that it will use in selecting which of the 808 different federations to use: 810 o The federation selected must be able to communicate with the IdP. 812 o The federation selected must match the business rules and 813 technical policies required for the RP security requirements. 815 The RP needs to discover which federation will be used to contact the 816 IdP. The first selection criteria used during discovery is going to 817 be the name of the IdP to be contacted. The second selection 818 criteria used during discovery is going to be the set of business 819 rules and technical policies governing the relationship; this is 820 called rules determination. The RP also needs to establish technical 821 trust in the communications with the IdP. 823 Rules determination covers a broad range of decisions about the 824 exchange. One of these is whether the given RP is permitted to talk 825 to the IdP using a given federation at all, so rules determination 826 encompasses the basic authorization decision. Other factors are 827 included, such as what policies govern release of information about 828 the client to the RP and what policies govern the RP's use of this 829 information. While rules determination is ultimately a business 830 function, it has significant impact on the technical exchanges. The 831 protocols need to communicate the result of authorization. When 832 multiple sets of rules are possible, the protocol must disambiguate 833 which set of rules are in play. Some rules have technical 834 enforcement mechanisms; for example in some federations 835 intermediaries validate information that is being communicated within 836 the federation. 838 At the time of writing no protocol mechanism has been specified to 839 allow a AAA client to determine whether a AAA proxy will indeed be 840 able to route AAA requests to a specific IdP. The AAA routing is 841 impacted by business rules and technical policies that may be quite 842 complex and at the present time, the route selection is based on 843 manual configuration. 845 2.1.3. Routing and Technical Trust 847 Several approaches to having messages routed through the federation 848 substrate are possible. These routing methods can most easily be 849 classified based on the mechanism for technical trust that is used. 850 The choice of technical trust mechanism constrains how rules 851 determination is implemented. Regardless of what deployment strategy 852 is chosen, it is important that the technical trust mechanism be able 853 to validate theg identities of both parties to the exchange. The 854 trust mechanism must to ensure that the entity acting as IdP for a 855 given NAI is permitted to be the IdP for that realm, and that any 856 service name claimed by the RP is permitted to be claimed by that 857 entity. Here are the categories of technical trust determination: 859 AAA Proxy: 860 The simplest model is that an RP is an AAA client and can send the 861 request directly to an AAA proxy. The hop-by-hop integrity 862 protection of the AAA fabric provides technical trust. An RP can 863 submit a request directly to a federation. Alternatively, a 864 federation disambiguation fabric can be used. Such a fabric takes 865 information about what federations the RP is part of and what 866 federations the IdP is part of and routes a message to the 867 appropriate federation. The routing of messages across the fabric 868 plus attributes added to requests and responses provides rules 869 determination. For example, when a disambiguation fabric routes a 870 message to a given federation, that federation's rules are chosen. 871 Name validation is enforced as messages travel across the fabric. 872 The entities near the RP confirm its identity and validate names 873 it claims. The fabric routes the message towards the appropriate 874 IdP, validating the IdP's name in the process. The routing can be 875 statically configured. Alternatively a routing protocol could be 876 developed to exchange reachability information about given a IdP 877 and to apply policy across the AAA fabric. Such a routing 878 protocol could flood naming constraints to the appropriate points 879 in the fabric. 881 Trust Broker: 882 Instead of routing messages through AAA proxies, some trust broker 883 could establish keys between entities near the RP and entities 884 near the IdP. The advantage of this approach is efficiency of 885 message handling. Fewer entities are needed to be involved for 886 each message. Security may be improved by sending individual 887 messages over fewer hops. Rules determination involves decisions 888 made by trust brokers about what keys to grant. Also, associated 889 with each credential is context about rules and about other 890 aspects of technical trust including names that may be claimed. A 891 routing protocol similar to the one for AAA proxies is likely to 892 be useful to trust brokers in flooding rules and naming 893 constraints. 895 Global Credential: 896 A global credential such as a public key and certificate in a 897 public key infrastructure can be used to establish technical 898 trust. A directory or distributed database such as the Domain 899 Name System is used by the RP to discover the endpoint to contact 900 for a given NAI. Either the database or certificates can provide 901 a place to store information about rules determination and naming 902 constraints. Provided that no intermediates are required (or 903 appear to be required) and that the RP and IdP are sufficient to 904 enforce and determine rules, rules determination is reasonably 905 simple. However applying certain rules is likely to be quite 906 complex. For example if multiple sets of rules are possible 907 between an IdP and RP, confirming the correct set is used may be 908 difficult. This is particularly true if intermediates are 909 involved in making the decision. Also, to the extent that 910 directory information needs to be trusted, rules determination may 911 be more complex. 913 Real-world deployments are likely to be mixtures of these basic 914 approaches. For example, it will be quite common for an RP to route 915 traffic to a AAA proxy within an organization. That proxy could then 916 use any of the three methods to get closer to the IdP. It is also 917 likely that rather than being directly reachable, the IdP may have a 918 proxy on the edge of its organization. Federations will likely 919 provide a traditional AAA proxy interface even if they also provide 920 another mechanism for increased efficiency or security. 922 2.1.4. AAA Security 923 For the AAA framework there are two different places where security 924 needs to be examined. The first is the security that is in place for 925 the links in the AAA backbone being used. The second is the nodes 926 that the backbone consists of. 928 The default link security for RADIUS is showing its age as it uses 929 MD5 and a shared secret to both obfuscate passwords and to provide 930 integrity on the RADIUS messages. While some EAP methods have 931 designed in the ability to protect the client authentication 932 credentials, the MSK returned from the IDP to the RP is protected 933 only by the RADIUS security. In many environments this is considered 934 to be insufficient, especially as not all attributes are obfuscated 935 and can thus leak information to a passive eavesdropper. The use of 936 RADIUS with TLS [RFC6614] and/or DTLS [I-D.ietf-radext-dtls] 937 addresses these attacks. The same level of security is included in 938 the base Diameter specifications. 940 2.1.5. SAML Assertions 942 For the traditional use of AAA frameworks, network access, the only 943 requirement that was necessary to grant access was an affirmative 944 response from the IdP. In the ABFAB world, the RP may need to get 945 additional information about the client before granting access. 946 ABFAB therefore has a requirement that it can transport an arbitrary 947 set of attributes about the client from the IdP to the RP. 949 Security Assertions Markup Language (SAML) [OASIS.saml-core-2.0-os] 950 was designed in order to carry an extensible set of attributes about 951 a subject. Since SAML is extensible in the attribute space, ABFAB 952 has no immediate needs to update the core SAML specifications for our 953 work. It will be necessary to update IdPs that need to return SAML 954 assertions to RPs and for both the IdP and the RP to implement a new 955 SAML profile designed to carry SAML assertions in AAA. The new 956 profile can be found in RFCXXXX [I-D.ietf-abfab-aaa-saml]. As SAML 957 statements will frequently be large, RADIUS servers and clients that 958 deal with SAML statements will need to implement RFC XXXX 959 [I-D.perez-radext-radius-fragmentation] 961 There are several issues that need to be highlighted: 963 o The security of SAML assertions. 965 o Namespaces and mapping of SAML attributes. 967 o Subject naming of entities. 969 o Making multiple queries about the subject(s). 971 o Level of Assurance for authentication. 973 SAML assertions have an optional signature that can be used to 974 protect and provide origination of the assertion. These signatures 975 are normally based on asymmetric key operations and require that the 976 verifier be able to check not only the cryptographic operation, but 977 also the binding of the originators name and the public key. In a 978 federated environment it will not always be possible for the RP to 979 validate the binding, for this reason the technical trust established 980 in the federation is used as an alternate method of validating the 981 origination and integrity of the SAML Assertion. 983 Attributes placed in SAML assertions can have different namespaces 984 assigned to the same name. In many, but not all, cases the 985 federation agreements will determine what attributes can be used in a 986 SAML statement. This means that the RP needs to map from the 987 federation names, types and semantics into the ones that the policies 988 of the RP are written in. In other cases the federation substrate 989 may modify the SAML assertions in transit to do the necessary 990 namespace, naming and semantic mappings as the assertion crosses the 991 different boundaries in the federation. If the proxies are modifying 992 the SAML Assertion, then they will obviously remove any signatures as 993 they would no longer validate. In this case the technical trust is 994 the required mechanism for validating the integrity of the assertion. 995 Finally, the attributes may still be in the namespace of the 996 originating IdP. When this occurs the RP will need to get the 997 required mapping operations from the federation agreements and do the 998 appropriate mappings itself. 1000 The RADIUS SAML RFC [I-D.ietf-abfab-aaa-saml] has define a new SAML 1001 name format that corresponds to the NAI name form defined by RFC XXXX 1002 [I-D.ietf-radext-nai]. This allows for easy name matching in many 1003 cases as the name form in the SAML statement and the name form used 1004 in RADIUS or Diameter will be the same. In addition to the NAI name 1005 form, the document also defines a pair of implicit name forms 1006 corresponding to the Client and the Client's machine. These implicit 1007 name forms are based on the Identity-Type enumeration defined in TEAP 1008 [I-D.ietf-emu-eap-tunnel-method]. If the name form returned in a 1009 SAML statement is not based on the NAI, then it is a requirement on 1010 the EAP server that it validate that the subject of the SAML 1011 assertion, if any, is equivalent to the subject identified by the NAI 1012 used in the RADIUS or Diameter session. 1014 RADIUS has the ability to deal with multiple SAML queries for those 1015 EAP Servers which follow RFC 5080 [RFC5080]. In this case a State 1016 attribute will always be returned with the Access-Accept. The EAP 1017 client can then send a new Access-Request with the State attribute 1018 and the new SAML Request Multiple SAML queries can then be done by 1019 making a new Access-Request using the State attribute returned in the 1020 last Access-Accept to link together the different RADIUS sessions. 1022 Some RPs need to ensure that specific criteria are met during the 1023 authentication process. This need is met by using Levels of 1024 Assurance. The way a Level of Assurance is communicated to the RP 1025 from the EAP server is by the use of a SAML Authentication Request 1026 using the Authentication Profile from RFC XXX 1027 [I-D.ietf-abfab-aaa-saml] When crossing boundaries between different 1028 federations, either the policy specified will need to be shared 1029 between the two federations, the policy will need to be mapped by the 1030 proxy server on the boundary or the proxy server on the boundary will 1031 need to supply information the EAP server so that it can do the 1032 required mapping. If this mapping is not done, then the EAP server 1033 will not be able to enforce the desired Level of Assurance as it will 1034 not understand the policy requirements. 1036 2.2. Client To Identity Provider 1038 Looking at the communications between the client and the IdP, the 1039 following items need to be dealt with: 1041 o The client and the IdP need to mutually authenticate each other. 1043 o The client and the IdP need to mutually agree on the identity of 1044 the RP. 1046 ABFAB selected EAP for the purposes of mutual authentication and 1047 assisted in creating some new EAP channel binding documents for 1048 dealing with determining the identity of the RP. A framework for the 1049 channel binding mechanism has been defined in RFC 6677 [RFC6677] that 1050 allows the IdP to check the identity of the RP provided by the AAA 1051 framework with that provided by the client. 1053 2.2.1. Extensible Authentication Protocol (EAP) 1055 Traditional web federation does not describe how a client interacts 1056 with an identity provider for authentication. As a result, this 1057 communication is not standardized. There are several disadvantages 1058 to this approach. Since the communication is not standardized, it is 1059 difficult for machines to correctly enter their credentials with 1060 different authentications, where Individuals can correctly identify 1061 the entire mechanism on the fly. The use of browsers for 1062 authentication restricts the deployment of more secure forms of 1063 authentication beyond plaintext username and password known by the 1064 server. In a number of cases the authentication interface may be 1065 presented before the client has adequately validated they are talking 1066 to the intended server. By giving control of the authentication 1067 interface to a potential attacker, the security of the system may be 1068 reduced and phishing opportunities introduced. 1070 As a result, it is desirable to choose some standardized approach for 1071 communication between the client's end-host and the identity 1072 provider. There are a number of requirements this approach must 1073 meet. 1075 Experience has taught us one key security and scalability 1076 requirement: it is important that the relying party not get 1077 possession of the long-term secret of the client. Aside from a 1078 valuable secret being exposed, a synchronization problem can develop 1079 when the client changes keys with the IdP. 1081 Since there is no single authentication mechanism that will be used 1082 everywhere there is another associated requirement: The 1083 authentication framework must allow for the flexible integration of 1084 authentication mechanisms. For instance, some IdPs require hardware 1085 tokens while others use passwords. A service provider wants to 1086 provide support for both authentication methods, and other methods 1087 from IdPs not yet seen. 1089 These requirements can be met by utilizing standardized and 1090 successfully deployed technology, namely by the Extensible 1091 Authentication Protocol (EAP) framework [RFC3748]. Figure 2 1092 illustrates the integration graphically. 1094 EAP is an end-to-end framework; it provides for two-way communication 1095 between a peer (i.e. client or individual) through the authenticator 1096 (i.e., relying party) to the back-end (i.e., identity provider). 1097 Conveniently, this is precisely the communication path that is needed 1098 for federated identity. Although EAP support is already integrated 1099 in AAA systems (see [RFC3579] and [RFC4072]) several challenges 1100 remain: 1102 o The first is how to carry EAP payloads from the end host to the 1103 relying party. 1105 o Another is to verify statements the relying party has made to the 1106 client, confirm these statements are consistent with statements 1107 made to the identity provider and confirm all the above are 1108 consistent with the federation and any federation-specific policy 1109 or configuration. 1111 o Another challenge is choosing which identity provider to use for 1112 which service. 1114 The EAP method used for ABFAB needs to meet the following 1115 requirements: 1117 o It needs to provide mutual authentication of the client and IdP. 1119 o It needs to support channel binding. 1121 As of this writing, the only EAP method that meets these criteria is 1122 TEAP [I-D.ietf-emu-eap-tunnel-method] either alone (if client 1123 certificates are used) or with an inner EAP method that does mutual 1124 authentication. 1126 2.2.2. EAP Channel Binding 1128 EAP channel binding is easily confused with a facility in GSS-API 1129 also called channel binding. GSS-API channel binding provides 1130 protection against man-in-the-middle attacks when GSS-API is used as 1131 authentication inside some tunnel; it is similar to a facility called 1132 cryptographic binding in EAP. See [RFC5056] for a discussion of the 1133 differences between these two facilities and Section 6.1 for how GSS- 1134 API channel binding is handled in this mechanism. 1136 The client knows, in theory, the name of the RP that it attempted to 1137 connect to, however in the event that an attacker has intercepted the 1138 protocol, the client and the IdP need to be able to detect this 1139 situation. A general overview of the problem along with a 1140 recommended way to deal with the channel binding issues can be found 1141 in RFC 6677 [RFC6677]. 1143 Since that document was published, a number of possible attacks were 1144 found and methods to address these attacks have been outlined in 1145 [I-D.ietf-emu-crypto-bind]. 1147 2.3. Client to Relying Party 1149 The final set of interactions between parties to consider are those 1150 between the client and the RP. In some ways this is the most complex 1151 set since at least part of it is outside the scope of the ABFAB work. 1152 The interactions between these parties include: 1154 o Running the protocol that implements the service that is provided 1155 by the RP and desired by the client. 1157 o Authenticating the client to the RP and the RP to the client. 1159 o Providing the necessary security services to the service protocol 1160 that it needs beyond authentication. 1162 o Deal with client re-authentication where desired. 1164 2.3.1. GSS-API 1166 One of the remaining layers is responsible for integration of 1167 federated authentication into the application. There are a number of 1168 approaches that applications have adopted for security. So, there 1169 may need to be multiple strategies for integration of federated 1170 authentication into applications. However, we have started with a 1171 strategy that provides integration to a large number of application 1172 protocols. 1174 Many applications such as SSH [RFC4462], NFS [RFC2203], DNS [RFC3645] 1175 and several non-IETF applications support the Generic Security 1176 Services Application Programming Interface [RFC2743]. Many 1177 applications such as IMAP, SMTP, XMPP and LDAP support the Simple 1178 Authentication and Security Layer (SASL) [RFC4422] framework. These 1179 two approaches work together nicely: by creating a GSS-API mechanism, 1180 SASL integration is also addressed. In effect, using a GSS-API 1181 mechanism with SASL simply requires placing some headers on the front 1182 of the mechanism and constraining certain GSS-API options. 1184 GSS-API is specified in terms of an abstract set of operations which 1185 can be mapped into a programming language to form an API. When 1186 people are first introduced to GSS-API, they focus on it as an API. 1187 However, from the prospective of authentication for non-web 1188 applications, GSS-API should be thought of as a protocol as well as 1189 an API. When looked at as a protocol, it consists of abstract 1190 operations such as the initial context exchange, which includes two 1191 sub-operations (gss_init_sec_context and gss_accept_sec_context). An 1192 application defines which abstract operations it is going to use and 1193 where messages produced by these operations fit into the application 1194 architecture. A GSS-API mechanism will define what actual protocol 1195 messages result from that abstract message for a given abstract 1196 operation. So, since this work is focusing on a particular GSS-API 1197 mechanism, we generally focus on protocol elements rather than the 1198 API view of GSS-API. 1200 The API view of GSS-API does have significant value as well, since 1201 the abstract operations are well defined, the set of information that 1202 a mechanism gets from the application is well defined. Also, the set 1203 of assumptions the application is permitted to make is generally well 1204 defined. As a result, an application protocol that supports GSS-API 1205 or SASL is very likely to be usable with a new approach to 1206 authentication including this one with no required modifications. In 1207 some cases, support for a new authentication mechanism has been added 1208 using plugin interfaces to applications without the application being 1209 modified at all. Even when modifications are required, they can 1210 often be limited to supporting a new naming and authorization model. 1211 For example, this work focuses on privacy; an application that 1212 assumes it will always obtain an identifier for the client will need 1213 to be modified to support anonymity, unlinkability or pseudonymity. 1215 So, we use GSS-API and SASL because a number of the application 1216 protocols we wish to federate support these strategies for security 1217 integration. What does this mean from a protocol standpoint and how 1218 does this relate to other layers? This means we need to design a 1219 concrete GSS-API mechanism. We have chosen to use a GSS-API 1220 mechanism that encapsulates EAP authentication. So, GSS-API (and 1221 SASL) encapsulate EAP between the end-host and the service. The AAA 1222 framework encapsulates EAP between the relying party and the identity 1223 provider. The GSS-API mechanism includes rules about how initiators 1224 and services are named as well as per-message security and other 1225 facilities required by the applications we wish to support. 1227 2.3.2. Protocol Transport 1229 The transport of data between the client and the relying party is not 1230 provided by GSS-API. GSS-API creates and consumes messages, but it 1231 does not provide the transport itself, instead the protocol using 1232 GSS-API needs to provide the transport. In many cases HTTP or HTTPS 1233 is used for this transport, but other transports are perfectly 1234 acceptable. The core GSS-API document [RFC2743] provides some 1235 details on what requirements exist. 1237 In addition we highlight the following: 1239 o The transport does not need to provide either privacy or 1240 integrity. After GSS-EAP has finished negotiation, GSS-API can be 1241 used to provide both services. If the negotiation process itself 1242 needs protection from eavesdroppers then the transport would need 1243 to provide the necessary services. 1245 o The transport needs to provide reliable transport of the messages. 1247 o The transport needs to ensure that tokens are delivered in order 1248 during the negotiation process. 1250 o GSS-API messages need to be delivered atomically. If the 1251 transport breaks up a message it must also reassemble the message 1252 before delivery. 1254 2.3.3. Reauthentication 1256 TBD. 1258 3. Application Security Services 1260 One of the key goals is to integrate federated authentication into 1261 existing application protocols and where possible, existing 1262 implementations of these protocols. Another goal is to perform this 1263 integration while meeting the best security practices of the 1264 technologies used to perform the integration. This section describes 1265 security services and properties required by the EAP GSS-API 1266 mechanism in order to meet these goals. This information could be 1267 viewed as specific to that mechanism. However, other future 1268 application integration strategies are very likely to need similar 1269 services. So, it is likely that these services will be expanded 1270 across application integration strategies if new application 1271 integration strategies are adopted. 1273 3.1. Authentication 1275 GSS-API provides an optional security service called mutual 1276 authentication. This service means that in addition to the initiator 1277 providing (potentially anonymous or pseudonymous) identity to the 1278 acceptor, the acceptor confirms its identity to the initiator. 1279 Especially for the ABFAB context, this service is confusingly named. 1280 We still say that mutual authentication is provided when the identity 1281 of an acceptor is strongly authenticated to an anonymous initiator. 1283 RFC 2743, unfortunately, does not explicitly talk about what mutual 1284 authentication means. Within this document we therefore define it 1285 as: 1287 o If a target name is configured for the initiator, then the 1288 initiator trusts that the supplied target name describes the 1289 acceptor. This implies both that appropriate cryptographic 1290 exchanges took place for the initiator to make such a trust 1291 decision, and that after evaluating the results of these 1292 exchanges, the initiator's policy trusts that the target name is 1293 accurate. 1295 o If no target name is configured for the initiator, then the 1296 initiator trusts that the acceptor name, supplied by the acceptor, 1297 correctly names the entity it is communicating with. 1299 o Both the initiator and acceptor have the same key material for 1300 per-message keys and both parties have confirmed they actually 1301 have the key material. In EAP terms, there is a protected 1302 indication of success. 1304 Mutual authentication is an important defense against certain aspects 1305 of phishing. Intuitively, clients would like to assume that if some 1306 party asks for their credentials as part of authentication, 1307 successfully gaining access to the resource means that they are 1308 talking to the expected party. Without mutual authentication, the 1309 server could "grant access" regardless of what credentials are 1310 supplied. Mutual authentication better matches this user intuition. 1312 It is important, therefore, that the GSS-EAP mechanism implement 1313 mutual authentication. That is, an initiator needs to be able to 1314 request mutual authentication. When mutual authentication is 1315 requested, only EAP methods capable of providing the necessary 1316 service can be used, and appropriate steps need to be taken to 1317 provide mutual authentication. While a broader set of EAP methods 1318 could be supported by not requiring mutual authentication, it was 1319 decided that the client needs to always have the ability to request 1320 it. In some cases the IdP and the RP will not support mutual 1321 authentication, however the client will always be able to detect this 1322 and make an appropriate security decision. 1324 The AAA infrastructure MAY hide the initiator's identity from the 1325 GSS-API acceptor, providing anonymity between the initiator and the 1326 acceptor. At this time, whether the identity is disclosed is 1327 determined by EAP server policy rather than by an indication from the 1328 initiator. Also, initiators are unlikely to be able to determine 1329 whether anonymous communication will be provided. For this reason, 1330 initiators are unlikely to set the anonymous return flag from 1331 GSS_Init_Sec_context. 1333 3.2. GSS-API Channel Binding 1335 [RFC5056] defines a concept of channel binding which is used prevent 1336 man-in-the-middle attacks. The channel binding works by taking a 1337 cryptographic value from the transport security and checks that both 1338 sides of the GSS-API conversation know this value. Transport Layer 1339 Security (TLS) is the most common transport security layer used for 1340 this purpose. 1342 It needs to be stressed that RFC 5056 channel binding (also called 1343 GSS-API channel binding when GSS-API is involved) is not the same 1344 thing as EAP channel binding. GSS-API channel binding is used for 1345 detecting Man-In-The-Middle attacks. EAP channel binding is used for 1346 mutual authentication and acceptor naming checks. Details are 1347 discussed in the mechanisms specification [I-D.ietf-abfab-gss-eap]. 1348 A fuller description of the differences between the facilities can be 1349 found in RFC 5056 [RFC5056]. 1351 The use of TLS can provide both encryption and integrity on the 1352 channel. It is common to provide SASL and GSS-API with these other 1353 security services. 1355 One of the benefits that the use of TLS provides, is that client has 1356 the ability to validate the name of the server. However this 1357 validation is predicated on a couple of things. The TLS sessions 1358 needs to be using certificates and not be an anonymous session. The 1359 client and the TLS need to share a common trust point for the 1360 certificate used in validating the server. TLS provides its own 1361 server authentication. However there are a variety of situations 1362 where this authentication is not checked for policy or usability 1363 reasons. Even when it is checked, if the trust infrastructure behind 1364 the TLS authentication is different from the trust infrastructure 1365 behind the GSS-API mutual authentication then confirming the end- 1366 points using both trust infrastructures is likely to enhance 1367 security. If the endpoints of the GSS-API authentication are 1368 different than the endpoints of the lower layer, this is a strong 1369 indication of a problem such as a man-in-the-middle attack. Channel 1370 binding provides a facility to determine whether these endpoints are 1371 the same. 1373 The GSS-EAP mechanism needs to support channel binding. When an 1374 application provides channel binding data, the mechanism needs to 1375 confirm this is the same on both sides consistent with the GSS-API 1376 specification. 1378 3.3. Host-Based Service Names 1380 IETF security mechanisms typically take a host name and perhaps a 1381 service, entered by a user, and make some trust decision about 1382 whether the remote party in the interaction is the intended party. 1383 This decision can be made by the use of certificates, pre-configured 1384 key information or a previous leap of trust. GSS-API has defined a 1385 relatively flexible name convention, however most of the IETF 1386 applications that use GSS-API (including SSH, NFS, IMAP, LDAP and 1387 XMPP) have chosen to use a more restricted naming convention based on 1388 the host name. The GSS-EAP mechanism needs to support host-based 1389 service names in order to work with existing IETF protocols. 1391 The use of host-based service names leads to a challenging trust 1392 delegation problem. Who is allowed to decide whether a particular 1393 host name maps to a specific entity. Possible solutions to this 1394 problem have been looked at. 1396 o The public-key infrastructure (PKI) used by the web has chosen to 1397 have a number of trust anchors (root certificate authorities) each 1398 of which can map any host name to a public key. 1400 o A number of GSS-API mechanisms, such as Kerberos [RFC1964], have 1401 split the problem into two parts. A new concept called a realm is 1402 introduced, the realm is responsible for host mapping within that 1403 realm. The mechanism then decides what realm is responsible for a 1404 given name. This is the approach adopted by ABFAB. 1406 GSS-EAP defines a host naming convention that takes into account the 1407 host name, the realm, the service and the service parameters. An 1408 example of GSS-API service name is "xmpp/foo@example.com". This 1409 identifies the XMPP service on the host foo in the realm example.com. 1410 Any of the components, except for the service name may be omitted 1411 from a name. When omitted, then a local default would be used for 1412 that component of the name. 1414 While there is no requirement that realm names map to Fully Qualified 1415 Domain Names (FQDN) within DNS, in practice this is normally true. 1416 Doing so allows for the realm portion of service names and the 1417 portion of NAIs to be the same. It also allows for the use of DNS in 1418 locating the host of a service while establishing the transport 1419 channel between the client and the relying party. 1421 It is the responsibility of the application to determine the server 1422 that it is going to communicate with, GSS-API has the ability to help 1423 confirm that the server is the desired server but not to determine 1424 the name of the server to use. It is also the responsibility of the 1425 application to determine how much of the information identifying the 1426 service needs to be validated by the ABFAB system. The information 1427 that needs to be validated is used to build up the service name 1428 passed into the GSS-EAP mechanism. What information is to be 1429 validated will depend on both what information was provided by the 1430 client, and what information is considered significant. If the 1431 client only cares about getting a specific service, then the host and 1432 realm that provides the service does not need to be validated. 1434 In many cases applications may retrieve information about providers 1435 of services from DNS. When Service Records (SRV) and Naming 1436 Authority Pointer (NAPTR) records are used to help find a host that 1437 provides a service, the security requirements on the referrals is 1438 going to interact with the information used in the service name. If 1439 a host name is returned from the DNS referrals, and the host name is 1440 to be validated by GS-EAP, then it makes sense that the referrals 1441 themselves should be secure. On the other hand, if the host name 1442 returned is not validated, i.e. only the service is passed in, then 1443 it is less important that the host name be obtained in a secure 1444 manner. 1446 Another issue that needs to be addressed for host-based service names 1447 is that they do not work ideally when different instances of a 1448 service are running on different ports. If the services are 1449 equivalent, then it does not matter. However if there are 1450 substantial differences in the quality of the service that 1451 information needs to be part of the validation process. If one has 1452 just a host name and not a port in the information being validated, 1453 then this is not going to be a successful strategy. 1455 3.4. Additional GSS-API Services 1457 GSS-API provides per-message security services that can provide 1458 confidentiality and/or integrity. Some IETF protocols such as NFS 1459 and SSH take advantage of these services. As a result GSS-EAP needs 1460 to support these services. As with mutual authentication, per- 1461 message services will limit the set of EAP methods that can be used 1462 to those that generate a Master Session Key (MSK). Any EAP method 1463 that produces an MSK is able to support per-message security services 1464 described in [RFC2743]. 1466 GSS-API provides a pseudo-random function. This function generates a 1467 pseudo-random sequence using the shared private key as the seed for 1468 the bytes generated. This provides an algorithm that both the 1469 initiator and acceptor can run in order to arrive at the same key 1470 value. The use of this feature allows for an application to generate 1471 keys or other shared secrets for use in other places in the protocol. 1472 In this regards, it is similar in concept to the TLS extractor (RFC 1473 5705 [RFC5705].). While no current IETF protocols require this, non- 1474 IETF protocols are expected to take advantage of this in the near 1475 future. Additionally, a number of protocols have found the TLS 1476 extractor to be useful in this regards so it is highly probably that 1477 IETF protocols may also start using this feature. 1479 4. Privacy Considerations 1481 ABFAB, as an architecture designed to enable federated authentication 1482 and allow for the secure transmission of identity information between 1483 entities, obviously requires careful consideration around privacy and 1484 the potential for privacy violations. 1486 This section examines the privacy related information presented in 1487 this document, summarising the entities that are involved in ABFAB 1488 communications and what exposure they have to identity information. 1489 In discussing these privacy considerations in this section, we use 1490 terminology and ideas from [I-D.iab-privacy-considerations]. 1492 Note that the ABFAB architecture uses at its core several existing 1493 technologies and protocols; detailed privacy discussion around these 1494 is not examined. This section instead focuses on privacy 1495 considerations specifically related to overall architecture and usage 1496 of ABFAB. 1498 +--------+ +---------------+ +--------------+ 1499 | Client | <---> | RP | <---> | AAA Client | 1500 +--------+ +---------------+ +--------------+ 1501 ^ 1502 | 1503 v 1504 +---------------+ +--------------+ 1505 | SAML Server | | AAA Proxy(s) | 1506 +---------------+ +--------------+ 1507 ^ ^ 1508 | | 1509 v v 1510 +------------+ +---------------+ +--------------+ 1511 | EAP Server | <---> | IdP | <---> | AAA Server | 1512 +------------+ +---------------+ +--------------+ 1514 Figure 3: Entities and Data Flow 1516 4.1. Entities and their roles 1518 Categorizing the ABFAB entities shown in the Figure 3 according to 1519 the taxonomy of terms from [I-D.iab-privacy-considerations] the 1520 entities shown in Figure 3 is somewhat complicated as during the 1521 various phases of ABFAB communications the roles of each entity 1522 changes. The three main phases of relevance are the Client to RP 1523 communication phase, the Client to IdP (via the Federation Substrate) 1524 phase, and the IdP to RP (via the Federation Substrate) phase. 1526 In the Client to RP communication phase, we have: 1528 Initiator: Client. 1530 Observers: Client, RP. 1532 Recipient: RP. 1534 In the Client to IdP (via the Federation Substrate) communication 1535 phase, we have: 1537 Initiator: Client. 1539 Observers: Client, RP, AAA Client, AAA Proxy(s), AAA Server, IdP. 1541 Recipient: IdP 1543 In the IdP to Relying party (via the Federation Substrate) 1544 communication phase, we have: 1546 Initiator: RP. 1548 Observers: IdP, AAA Server, AAA Proxy(s), AAA Client, RP. 1550 Recipient: IdP 1552 Eavesdroppers and Attackers can reside on any communication link 1553 between entities in Figure 3. 1555 The Federation Substrate consists of all of the AAA entities. In 1556 some cases the AAA Proxies entities may not exist as the AAA Client 1557 can talk directly to the AAA Server. Specifications such as the 1558 Trust Router Protocol and RADIUS dynamic discovery 1559 [I-D.ietf-radext-dynamic-discovery] can be used to shorten the path 1560 between the AAA client and the AAA server (and thus stop these AAA 1561 Proxies from being Observers), however even in these circumstances 1562 there may be AAA Proxies in the path. 1564 In Figure 3 the IdP has been divided into multiple logical pieces, in 1565 actual implementations these pieces will frequently be tightly 1566 coupled. The links between these pieces provide the greatest 1567 opportunity for attackers and eavesdroppers to acquire information, 1568 however, as they are all under the control of a single entity they 1569 are also the easiest to have tightly secured. 1571 4.2. Privacy Aspects of ABFAB Communication Flows 1573 In the ABFAB architecture, there are a few different types of data 1574 and identifiers in use. The best way to understand them, and the 1575 potential privacy impacts of them, is to look at each phase of 1576 communication in ABFAB. 1578 4.2.1. Client to RP 1580 The flow of data between the client and the RP is divided into two 1581 parts. The first part consists of all of the data exchanged as part 1582 of the ABFAB authentication process. The second part consists of all 1583 of the data exchanged after the authentication process has been 1584 finished. 1586 During the initial communications phase, the client sends an NAI (see 1587 [I-D.ietf-radext-nai]) to the RP. Many EAP methods (but not all) 1588 allow for the client to disclose an NAI to RP the in a form that 1589 includes only a realm component during this communications phase. 1590 This is the minimum amount of identity information necessary for 1591 ABFAB to work - it indicates an IdP that the principal has a 1592 relationship with. EAP methods that do not allow this will 1593 necessarily also reveal an identifier for the principal in the IdP 1594 realm (e.g. a username). 1596 The data shared during the initial communication phase may be 1597 protected by a channel protocol such as TLS. This will prevent the 1598 leak of information to passive eavesdroppers, however an active 1599 attacker may still be able to setup as a man-in-the-middle. The 1600 client may not be able to validate the certificates (if any) provided 1601 by the service, defering the check of the identity of the RP until 1602 the completion of the ABFAB authentication protocol (i.e., using EAP 1603 channel binding). 1605 The data exchanged after the authentication process can have privacy 1606 and authentication using the GSS-API services. If the overall 1607 application protocol allows for the process of re-authentication, 1608 then the same privacy impliciations as discussed in previous 1609 paragraphs apply. 1611 4.2.2. Client to IdP (via Federation Substrate) 1613 This phase sees a secure TLS tunnel initiated between the Client and 1614 the IdP via the RP and federation substrate. The process is 1615 initiated by the RP using the realm information given to it by the 1616 client. Once set up, the tunnel is used to send credentials to IdP 1617 to authenticate. 1619 Various operational information is transported between RP and IdP, 1620 over the AAA infrastructure, for example using RADIUS headers. As no 1621 end-to-end security is provided by AAA, all AAA entities on the path 1622 between the RP and IdP have the ability to eavesdrop on this 1623 information unless additional security measures are taken (such as 1624 the use of TLS for RADIUS [I-D.ietf-radext-dtls]). Some of this 1625 information may form identifiers or explicit identity information: 1627 o The Relying Party knows the IP address of the Client. It is 1628 possible that the Relying Party could choose to expose this IP 1629 address by including it in a RADIUS header such as Calling Station 1630 ID. This is a privacy consideration to take into account of the 1631 application protocol. 1633 o The EAP MSK is transported between the IdP and the RP over the AAA 1634 infrastructure, for example through RADIUS headers. This is a 1635 particularly important privacy consideration, as any AAA Proxy 1636 that has access to the EAP MSK is able to decrypt and eavesdrop on 1637 any traffic encrypted using that EAP MSK (i.e. all communications 1638 between the Client and IdP). 1640 o Related to the above, the AAA server has access to the material 1641 necessary to derive the session key, thus the AAA server can 1642 observe any traffic encrypted between the Client and RP. This 1643 "feature" was" chosen as a simplification and to make performance 1644 faster; if it was decided that this trade-off was not desireable 1645 for privacy and security reasons, then extensions to ABFAB that 1646 make use of techniques such as Diffie-Helman key exchange would 1647 mitigate against this. 1649 The choice of EAP method used has other potential privacy 1650 implications. For example, if the EAP method in use does not support 1651 trust anchors to enable mutual authentication, then there are no 1652 guarantees that the IdP is who it claims to be, and thus the full NAI 1653 including a username and a realm might be sent to any entity 1654 masquerading as a particular IdP. 1656 Note that ABFAB has not specified any AAA accounting requirements. 1657 Implementations that use the accounting portion of AAA should 1658 consider privacy appropriately when designing this aspect. 1660 4.2.3. IdP to RP (via Federation Substrate) 1662 In this phase, the IdP communicates with the RP informing it as to 1663 the success or failure of authentication of the user, and optionally, 1664 the sending of identity information about the principal. 1666 As in the previous flow (Client to IdP), various operation 1667 information is transported between IdP and RP over the AAA 1668 infrastructure, and the same privacy considerations apply. However, 1669 in this flow, explicit identity information about the authenticated 1670 principal can be sent from the IdP to the RP. This information can 1671 be sent through RADIUS headers, or using SAML 1672 [I-D.ietf-abfab-aaa-saml]. This can include protocol specific 1673 identitifiers, such as SAML NameIDs, as well as arbitrary attribute 1674 information about the principal. What information will be released 1675 is controlled by policy on the Identity Provider. As before, when 1676 sending this through RADIUS headers, all AAA entities on the path 1677 between the RP and IdP have the ability to eavesdrop unless 1678 additional security measures are taken (such as the use of TLS for 1679 RADIUS [I-D.ietf-radext-dtls]). When sending this using SAML, as 1680 specified in [I-D.ietf-abfab-aaa-saml], confidentiality of the 1681 information should however be guaranteed as [I-D.ietf-abfab-aaa-saml] 1682 requires the use of TLS for RADIUS. 1684 4.3. Relationship between User and Entities 1686 o Between User and IdP - the IdP is an entity the user will have a 1687 direct relationship with, created when the organisation that 1688 operates the entity provisioned and exchanged the user's 1689 credentials. Privacy and data protection guarantees may form a 1690 part of this relationship. 1692 o Between User and RP - the RP is an entity the user may or may not 1693 have a direct relationship with, depending on the service in 1694 question. Some services may only be offered to those users where 1695 such a direct relationship exists (for particularly sensitive 1696 services, for example), while some may not require this and would 1697 instead be satisfied with basic federation trust guarantees 1698 between themselves and the IdP). This may well include the option 1699 that the user stays anonymous with respect to the RP (though 1700 obviously never to the IdP). If attempting to preserve privacy 1701 through the mitigation of data minimisation, then the only 1702 attribute information about individuals exposed to the RP should 1703 be that which is strictly necessary for the operation of the 1704 service. 1706 o Between User and Federation substrate - the user is highly likely 1707 to have no knowledge of, or relationship with, any entities 1708 involved with the federation substrate (not that the IdP and/or RP 1709 may, however). Knowledge of attribute information about 1710 individuals for these entities is not necessary, and thus such 1711 information should be protected in such a way as to prevent access 1712 to this information from being possible. 1714 4.4. Accounting Information 1716 Alongside the core authentication and authorization that occurs in 1717 AAA communications, accounting information about resource consumption 1718 may be delivered as part of the accounting exchange during the 1719 lifetime of the granted application session. 1721 4.5. Collection and retention of data and identifiers 1723 In cases where Relying Parties do not require to identify a 1724 particular individual when an individual wishes to make use of their 1725 service, the ABFAB architecture enable anonymous or pseudonymous 1726 access. Thus data and identifiers other than pseudonyms and 1727 unlinkable attribute information need not be stored and retained. 1729 However, in cases where Relying Parties require the ability to 1730 identify a particular individual (e.g. so they can link this identity 1731 information to a particular account in their service, or where 1732 identity information is required for audit purposes), the service 1733 will need to collect and store such information, and to retain it for 1734 as long as they require. Deprovisioning of such accounts and 1735 information is out of scope for ABFAB, but obviously for privacy 1736 protection any identifiers collected should be deleted when they are 1737 no longer needed. 1739 4.6. User Participation 1741 In the ABFAB architecture, by its very nature users are active 1742 participants in the sharing of their identifiers as they initiate the 1743 communications exchange every time they wish to access a server. 1744 They are, however, not involved in control of the set of information 1745 related to them that transmitted from the IdP to RP for authorisation 1746 purposes; rather, this is under the control of policy on the IdP. 1747 Due to the nature of the AAA communication flows, with the current 1748 ABFAB architecture there is no place for a process of gaining user 1749 consent for the information to be released from IdP to RP. 1751 5. Security Considerations 1753 This document describes the architecture for Application Bridging for 1754 Federated Access Beyond Web (ABFAB) and security is therefore the 1755 main focus. This section highlights the main communication channels 1756 and their security properties: 1758 Client-to-RP Channel: 1760 The channel binding material is provided by any certificates and 1761 the final message (i.e., a cryptographic token for the channel). 1762 Authentication may be provided by the RP to the client but a 1763 deployment without authentication at the TLS layer is possible as 1764 well. In addition, there is a channel between the GSS requestor 1765 and the GSS acceptor, but the keying material is provided by a 1766 "third party" to both entities. The client can derive keying 1767 material locally, but the RP gets the material from the IdP. In 1768 the absence of a transport that provides encryption and/or 1769 integrity, the channel between the client and the RP has no 1770 ability to have any cryptographic protection until the EAP 1771 authentication has been completed and the MSK is transferred from 1772 the IdP to the RP. 1774 RP-to-IdP Channel: 1776 The security of this communication channel is mainly provided by 1777 the functionality offered via RADIUS and Diameter. At the time of 1778 writing there are no end-to-end security mechanisms standardized 1779 and thereby the architecture has to rely on hop-by-hop security 1780 with trusted AAA entities or, as an alternative but possible 1781 deployment variant, direct communication between the AAA client to 1782 the AAA server. Note that the authorization result the IdP 1783 provides to the RP in the form of a SAML assertion may, however, 1784 be protected such that the SAML related components are secured 1785 end-to-end. 1787 The MSK is transported from the IdP to the RP over this channel. 1788 As no end-to-end security is provided by AAA, all AAA entities on 1789 the path between the RP and IdP have the ability to eavesdrop if 1790 no additional security measures are taken. One such measure is to 1791 use a transport between the client and the IdP that provides 1792 confidentiality. 1794 Client-to-IdP Channel: 1796 This communication interaction is accomplished with the help of 1797 EAP and EAP methods. The offered security protection will depend 1798 on the EAP method that is chosen but a minimum requirement is to 1799 offer mutual authentication, and key derivation. The IdP is 1800 responsible during this process to determine that the RP that is 1801 communication to the client over the RP-to-IdP channel is the same 1802 one talking to the IdP. This is accomplished via the EAP channel 1803 binding. 1805 Partial list of issues to be addressed in this section: Privacy, 1806 SAML, Trust Anchors, EAP Algorithm Selection, Diameter/RADIUS/AAA 1807 Issues, Naming of Entities, Protection of passwords, Channel Binding, 1808 End-point-connections (TLS), Proxy problems 1810 When a pseudonym is generated as a unique long term identifier for a 1811 client by an IdP, care MUST be taken in the algorithm that it cannot 1812 easily be reverse engineered by the service provider. If it can be 1813 reversed then the service provider can consult an oracle to determine 1814 if a given unique long term identifier is associated with a different 1815 known identifier. 1817 6. IANA Considerations 1819 This document does not require actions by IANA. 1821 7. Acknowledgments 1822 We would like to thank Mayutan Arumaithurai, Klaas Wierenga and Rhys 1823 Smith for their feedback. Additionally, we would like to thank Eve 1824 Maler, Nicolas Williams, Bob Morgan, Scott Cantor, Jim Fenton, Paul 1825 Leach, and Luke Howard for their feedback on the federation 1826 terminology question. 1828 Furthermore, we would like to thank Klaas Wierenga for his review of 1829 the pre-00 draft version. 1831 8. References 1833 8.1. Normative References 1835 [RFC2743] Linn, J., "Generic Security Service Application Program 1836 Interface Version 2, Update 1", RFC 2743, January 2000. 1838 [RFC2865] Rigney, C., Willens, S., Rubens, A., and W. Simpson, 1839 "Remote Authentication Dial In User Service (RADIUS)", RFC 1840 2865, June 2000. 1842 [RFC3588] Calhoun, P., Loughney, J., Guttman, E., Zorn, G., and J. 1843 Arkko, "Diameter Base Protocol", RFC 3588, September 2003. 1845 [RFC3748] Aboba, B., Blunk, L., Vollbrecht, J., Carlson, J., and H. 1846 Levkowetz, "Extensible Authentication Protocol (EAP)", RFC 1847 3748, June 2004. 1849 [RFC3579] Aboba, B. and P. Calhoun, "RADIUS (Remote Authentication 1850 Dial In User Service) Support For Extensible 1851 Authentication Protocol (EAP)", RFC 3579, September 2003. 1853 [RFC4072] Eronen, P., Hiller, T., and G. Zorn, "Diameter Extensible 1854 Authentication Protocol (EAP) Application", RFC 4072, 1855 August 2005. 1857 [I-D.ietf-abfab-gss-eap] 1858 Hartman, S. and J. Howlett, "A GSS-API Mechanism for the 1859 Extensible Authentication Protocol", draft-ietf-abfab-gss- 1860 eap-09 (work in progress), August 2012. 1862 [I-D.ietf-abfab-aaa-saml] 1863 Howlett, J. and S. Hartman, "A RADIUS Attribute, Binding, 1864 Profiles, Name Identifier Format, and Confirmation Methods 1865 for SAML", draft-ietf-abfab-aaa-saml-05 (work in 1866 progress), February 2013. 1868 [I-D.ietf-radext-nai] 1869 DeKok, A., "The Network Access Identifier", draft-ietf- 1870 radext-nai-02 (work in progress), January 2013. 1872 [RFC6677] Hartman, S., Clancy, T., and K. Hoeper, "Channel-Binding 1873 Support for Extensible Authentication Protocol (EAP) 1874 Methods", RFC 6677, July 2012. 1876 8.2. Informative References 1878 [RFC2903] de Laat, C., Gross, G., Gommans, L., Vollbrecht, J., and 1879 D. Spence, "Generic AAA Architecture", RFC 2903, August 1880 2000. 1882 [I-D.nir-tls-eap] 1883 Nir, Y., Sheffer, Y., Tschofenig, H., and P. Gutmann, "A 1884 Flexible Authentication Framework for the Transport Layer 1885 Security (TLS) Protocol using the Extensible 1886 Authentication Protocol (EAP)", draft-nir-tls-eap-13 (work 1887 in progress), December 2011. 1889 [I-D.ietf-oauth-v2] 1890 Hardt, D., "The OAuth 2.0 Authorization Framework", draft- 1891 ietf-oauth-v2-31 (work in progress), August 2012. 1893 [I-D.iab-privacy-considerations] 1894 Cooper, A., Tschofenig, H., Aboba, B., Peterson, J., 1895 Morris, J., Hansen, M., and R. Smith, "Privacy 1896 Considerations for Internet Protocols", draft-iab-privacy- 1897 considerations-03 (work in progress), July 2012. 1899 [I-D.perez-radext-radius-fragmentation] 1900 Perez-Mendez, A., Lopez, R., Pereniguez-Garcia, F., Lopez- 1901 Millan, G., Lopez, D., and A. DeKok, "Support of 1902 fragmentation of RADIUS packets", draft-perez-radext- 1903 radius-fragmentation-05 (work in progress), February 2013. 1905 [RFC4017] Stanley, D., Walker, J., and B. Aboba, "Extensible 1906 Authentication Protocol (EAP) Method Requirements for 1907 Wireless LANs", RFC 4017, March 2005. 1909 [RFC5106] Tschofenig, H., Kroeselberg, D., Pashalidis, A., Ohba, Y., 1910 and F. Bersani, "The Extensible Authentication Protocol- 1911 Internet Key Exchange Protocol version 2 (EAP-IKEv2) 1912 Method", RFC 5106, February 2008. 1914 [RFC1964] Linn, J., "The Kerberos Version 5 GSS-API Mechanism", RFC 1915 1964, June 1996. 1917 [RFC2203] Eisler, M., Chiu, A., and L. Ling, "RPCSEC_GSS Protocol 1918 Specification", RFC 2203, September 1997. 1920 [RFC3645] Kwan, S., Garg, P., Gilroy, J., Esibov, L., Westhead, J., 1921 and R. Hall, "Generic Security Service Algorithm for 1922 Secret Key Transaction Authentication for DNS (GSS-TSIG)", 1923 RFC 3645, October 2003. 1925 [RFC2138] Rigney, C., Rigney, C., Rubens, A., Simpson, W., and S. 1926 Willens, "Remote Authentication Dial In User Service 1927 (RADIUS)", RFC 2138, April 1997. 1929 [RFC4462] Hutzelman, J., Salowey, J., Galbraith, J., and V. Welch, 1930 "Generic Security Service Application Program Interface 1931 (GSS-API) Authentication and Key Exchange for the Secure 1932 Shell (SSH) Protocol", RFC 4462, May 2006. 1934 [RFC4422] Melnikov, A. and K. Zeilenga, "Simple Authentication and 1935 Security Layer (SASL)", RFC 4422, June 2006. 1937 [RFC5056] Williams, N., "On the Use of Channel Bindings to Secure 1938 Channels", RFC 5056, November 2007. 1940 [RFC5080] Nelson, D. and A. DeKok, "Common Remote Authentication 1941 Dial In User Service (RADIUS) Implementation Issues and 1942 Suggested Fixes", RFC 5080, December 2007. 1944 [RFC5705] Rescorla, E., "Keying Material Exporters for Transport 1945 Layer Security (TLS)", RFC 5705, March 2010. 1947 [RFC5801] Josefsson, S. and N. Williams, "Using Generic Security 1948 Service Application Program Interface (GSS-API) Mechanisms 1949 in Simple Authentication and Security Layer (SASL): The 1950 GS2 Mechanism Family", RFC 5801, July 2010. 1952 [RFC5849] Hammer-Lahav, E., "The OAuth 1.0 Protocol", RFC 5849, 1953 April 2010. 1955 [RFC6614] Winter, S., McCauley, M., Venaas, S., and K. Wierenga, 1956 "Transport Layer Security (TLS) Encryption for RADIUS", 1957 RFC 6614, May 2012. 1959 [OASIS.saml-core-2.0-os] 1960 Cantor, S., Kemp, J., Philpott, R., and E. Maler, 1961 "Assertions and Protocol for the OASIS Security Assertion 1962 Markup Language (SAML) V2.0", OASIS Standard saml- 1963 core-2.0-os, March 2005. 1965 [RFC2904] Vollbrecht, J., Calhoun, P., Farrell, S., Gommans, L., 1966 Gross, G., de Bruijn, B., de Laat, C., Holdrege, M., and 1967 D. Spence, "AAA Authorization Framework", RFC 2904, August 1968 2000. 1970 [I-D.ietf-emu-crypto-bind] 1971 Hartman, S., Wasserman, M., and D. Zhang, "EAP Mutual 1972 Cryptographic Binding", draft-ietf-emu-crypto-bind-03 1973 (work in progress), March 2013. 1975 [I-D.ietf-emu-eap-tunnel-method] 1976 Zhou, H., Cam-Winget, N., Salowey, J., and S. Hanna, 1977 "Tunnel EAP Method (TEAP) Version 1", draft-ietf-emu-eap- 1978 tunnel-method-05 (work in progress), February 2013. 1980 [I-D.ietf-radext-dtls] 1981 DeKok, A., "DTLS as a Transport Layer for RADIUS", draft- 1982 ietf-radext-dtls-03 (work in progress), January 2013. 1984 [I-D.ietf-radext-dynamic-discovery] 1985 Winter, S. and M. McCauley, "NAI-based Dynamic Peer 1986 Discovery for RADIUS/TLS and RADIUS/DTLS", draft-ietf- 1987 radext-dynamic-discovery-06 (work in progress), February 1988 2013. 1990 [WS-TRUST] 1991 Lawrence, K., Kaler, C., Nadalin, A., Goodner, M., Gudgin, 1992 M., Barbir, A., and H. Granqvist, "WS-Trust 1.4", OASIS 1993 Standard ws-trust-200902, February 2009, . 1996 [NIST-SP.800-63] 1997 Burr, W., Dodson, D., and W. Polk, "Electronic 1998 Authentication Guideline", NIST Special Publication 1999 800-63, April 2006. 2001 Authors' Addresses 2003 Josh Howlett 2004 JANET(UK) 2005 Lumen House, Library Avenue, Harwell 2006 Oxford OX11 0SG 2007 UK 2009 Phone: +44 1235 822363 2010 Email: Josh.Howlett@ja.net 2011 Sam Hartman 2012 Painless Security 2014 Email: hartmans-ietf@mit.edu 2016 Hannes Tschofenig 2017 Nokia Siemens Networks 2018 Linnoitustie 6 2019 Espoo 02600 2020 Finland 2022 Phone: +358 (50) 4871445 2023 Email: Hannes.Tschofenig@gmx.net 2024 URI: http://www.tschofenig.priv.at 2026 Eliot Lear 2027 Cisco Systems GmbH 2028 Richtistrasse 7 2029 Wallisellen, ZH CH-8304 2030 Switzerland 2032 Phone: +41 44 878 9200 2033 Email: lear@cisco.com 2035 Jim Schaad 2036 Soaring Hawk Consulting 2038 Email: ietf@augustcellars.com