Application Bridging for Federated Access Beyond Web (ABFAB) Architecture
draft-lear-abfab-arch-01.txt

Abstract

Over the last decade a substantial amount of work has occurred in the space of federated authentication and authorization. Most of this effort has focused on two common use cases: network and web-based access, with few common building blocks within the architecture. This memo describes an architecture that makes use of extensions to the commonly used mechanisms for both federated and non-federated authentication and authorization, including Radius/Diameter, GSS/GS2, and SAML, to primarily address non-web based authentication, in a that will scale to large numbers of federations.

Status of this Memo

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

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

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

This Internet-Draft will expire on June 24, 2011.

Copyright Notice

Copyright (c) 2010 IETF Trust and the persons identified as the document authors. All rights reserved.

This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.

Table of Contents

1.  Introduction
    1.1.  Federation Description
    1.2.  Design Goals
    1.3.  Use of Radius
2.  Terminology
3.  Architecture
    3.1.  Federation Substrate
    3.2.  Subject To Identity Provider
    3.3.  Application to Service
    3.4.  Personalization Layer
    3.5.  Tieing Layers Together
4.  Application Security Services
    4.1.  Server (Mutual) Authentication
    4.2.  GSS-API Channel Binding
    4.3.  Host-Based Service Names
    4.4.  Per-Message Tokens
5.  Privacy Considerations
6.  Deployment Considerations
    6.1.  EAP Channel Binding
    6.2.  AAA Proxy Behavior
7.  Security Considerations
8.  IANA Considerations
9.  Acknowledgments
10.  References
    10.1.  Normative References
    10.2.  Informative References
§  Authors' Addresses

1.  Introduction

XXX This document is a first draft. Comments and contributions are requested.

The Internet makes uses of numerous authentication methods to grant access to various resources. These mechanisms have been generalized and scaled over the last decade through mechanisms such as GS2, Security Assertion Markup Language (SAML) (Cantor, S., Kemp, J., Philpott, R., and E. Maler, “Assertions and Protocol for the OASIS Security Assertion Markup Language (SAML) V2.0,” March 2005.) [OASIS.saml‑core‑2.0‑os], Radius, and Diameter. So-called "federated" access has evolved over the last decade between web servers through such standards as SAML, OpenID, and OAUTH, allowing entire domains of individuals to be authorized for resources. The key scaling points that have been addressed are the following:

• An Internet service need not copy manually authentication information from a domain to allow for authentication and authorization.
• Individual users are able to make use of a single credential to authenticate to such services.

As the number of such federated services has proliferated, however, the role of the individual has become ambiguous in certain circumstances. For example, a school might provide online access to grades to a parent who is also a teacher. She must clearly distinguish her role upon access. After all, she is probably not allowed to edit her own child's grades.

Similarly, as the number of federations proliferates, it becomes increasingly difficult to discover which identity provider a user is associated with. This is true for both the web and non-web case, but particularly acute for the latter ans many non-web authentication systems are not semantically rich enough on their own to allow for such ambiguities. For instance, in the case of an email provider, the use of SMTP and IMAP protocols does not on its own provide for a way to select a federation. However, the building blocks do exist to add this functionality.

1.1.  Federation Description

The typical setup for a three party protocol involves the following entities:

• the End Host,
• the Identity Provider, and
• the Relying Party.

These entities are illustrated graphically in Figure 1 (Three Party Authentication Framework).

                       -----
                     /-     -\
                   //         \\
                   /           \
                  |             |
 ,----------\    |               |   ,---------\
 | Identity |    |               |   | Relying |
 | Provider +----+   Federation  +---+ Party   |
 `----------'    |               |   '---------'
          <      |               |        >
           \      |             |        /
            \      \           /        /
             \     \\         //       /
             \       \-     -/        /
              \        -----         /
               \                    /
                \  +------------+  /
                 \ |            | /
                  v|  End Host  |v
                   |            |
                   +------------+

Figure 1 (Three Party Authentication Framework) also shows the logical entity 'Federation'. In a federation, policy is agreed upon by some form of administrative management, and then instantiated through an operational framework that the members use, and where compliance is measured in some fashion. Some deployments may be required to deploy message routing intermediaries, such as application layer relays or proxies, to offer the required technical functionality while in other deployments those are missing.

Often a real world entity is associated with the end host and responsible for interacting with the identity provider, even if it is only as weak as completing a web form and confirming the verification email. The outcome of this initial registration step is that credentials are made available to the identity provider and to the end host. It is important to highlight that in some scenarios there might indeed be a human behind the device denoted as end host and in other cases there is no human involved in the actual protocol execution.

To support the more generic deployment case, we assume that the identity provider and the relying party belong to different administrative domains. The nature of federation dictates that there is some form of relationship between the identity provider and the relying party. This is particularly important when the relying party wants to use information obtained from the identity provider for authorization decisions and when the identity provider does not want to release information to every relying party (or only under certain conditions). While it is possible to have a bilateral agreement between every identity provider and every relying party; on an Internet scale this setup requires the introduction of a federation concept, as the management of such pair-wise relationships would otherwise prove burdensome. While many of the non-technical aspects of such a federation, such as business practices and operational arrangements, are outside the scope of the IETF they still impact the architecture setup on how to ensure the dynamic establishment of trust.

The steps taken generally in an ABFAB federated authentication/authorization exchange are as follows (XXX not complete):

1. Principal provides NAI to Application: Somehow the client is configured with at least the realm portion of an NAI, which represents the IdP to be discovered.
2. Authentication mechanism selection: this is the step necessary to indicate that the GSS-EAP SASL/GS2 mechanism will be used for authentication/authorization.
3. Client Application provides NAI to RP: At the conclusion of mechanism selection the NAI must be provided to the RP for discovery.
4. Discovery of federated IdP: This is discussed in detail below. Either the RP is configured with authorized IdPs, or it makes use of a federation proxy.
5. Request from Relying Party to IdP: Once the RP knows who the IdP is, it or its agent will forward RADIUS request that encapsulates a GSS/EAP access request to an IdP. This may or may not contain a SAML request as a series of attributes.. At this stage, the RP will likely have no idea who the principal is. The RP claims its identity to the IdP in AAA attributes.
6. IdP informs the principal of which EAP method to use: The available and appropriate methods are discussed below in this memo.
7. A bunch of EAP messages happen between the endpoints: Messages are exchanged between the principal and the IdP until a result is determined. The number and content of those messages will depend on the EAP method. If the IdP is unable to authenticate the principal, the process concludes here. As part of this process, the principal will, under protection of EAP, assert the identity of the RP to which it intends to authenticate.
8. Successful Authentication: At the very least the EAP server / IdP has authenticated the principal, and the principal has authenticated the IdP. As a result of this step, the principal and the EAP server hold two cryptographic keys- a Master Session Key (MSK), and an Extended MSK (EMSK). If the asserted identity of the RP by the principal matches the identity the RP itself asserted, there is some confidence that the RP is now authenticated to the IdP.
9. Local IdP Policy Check: At this stage, the IdP checks local policy to determine whether the RP and principal are authorized for the assertion to be made. Additional policy checks will likely have been made earlier just through the process of discovery (see later discussion).
10. Response from the IdP to the Relying Party: Once the IdP has made a determination of whether and how to authenticate or authorize the principal to the RP, it returns either a negative answer to the RP, or it returns the identity of the principal to the RP, as well as an optional set of attributes associated with the principal. XXX XXX XXX this needs work!!!
11. Return results to principal: Once the RP has a response it must inform the client application of the result. If all has gone well, all are authenticated, and the application proceeds with appropriate authorization levels.

An example communication flow is given below:

    Relying Party    Client App       IdP

       |              (1)             | Client App gets NAI (somehow)
       |               |              |
       |<-----(2)----->|              | Mechanism Selection
       |               |              |
       |<-----(3)-----<|              | NAI transmitted to RP
       |               |              |
       |<=====(4)====================>| Discovery
       |               |              |
       |>=====(5)====================>| Access request from RP to IdP
       |               |              |
       |               |< - - (6) - -<| EAP method to Principal
       |               |              |
       |               |< - - (7) - ->| EAP Exchange to authenticate
       |               |              | Principal
       |               |              |
       |               |           (8 & 9) Local Policy Check
       |               |              |
       |<====(10)====================<| IdP Assertion to RP
       |               |              |
       |>----(11)----->|              | Results to client app.



     ----- = Between Client App and RP
     ===== = Between RP and IdP
     - - - = Between Client App and IdP

1.2.  Design Goals

Our key design goals are as follows:

• Each party of a transaction will be authenticated, and the principal will be authorized for access to a specific resource .
• Means of authentication is decoupled so as to allow for multiple authentication methods.
• Hence, the architecture requires no sharing of long term private keys.
• The system will scale to large numbers of identity providers, relying parties, and users.
• The system will be designed primarily for non-Web-based authentication.
• The system will build upon existing standards, components, and operational practices.

Designing new three party authentication and authorization protocols is hard and frought with risk of cryptographic flaws. Achieving widespead deployment is even more difficult. A lot of attention on federated access has been devoted to the Web. This document instead focuses on a non-Web-based environment and focuses on those protocols where HTTP is not used. Despite the increased excitement for layering every protocol on top of HTTP there are still a number of protocols available that do not use HTTP-based transports. Many of these protocols are lacking a native authentication and authorization framework of the style shown in Figure 1 (Three Party Authentication Framework).

1.3.  Use of Radius

Interestingly, for network access authentication the usage of the AAA framework with RADIUS [RFC2865] (Rigney, C., Willens, S., Rubens, A., and W. Simpson, “Remote Authentication Dial In User Service (RADIUS),” June 2000.) and Diameter [RFC3588] (Calhoun, P., Loughney, J., Guttman, E., Zorn, G., and J. Arkko, “Diameter Base Protocol,” September 2003.) was quite successful from a deployment point of view. To map the terminology used in Figure 1 (Three Party Authentication Framework) to the AAA framework the identity provider corresponds to the AAA server, the relying party corresponds to the AAA client, and the technical building blocks of a federation are AAA proxies, relays and redirect agents (particularly if they are operated by third parties, such as AAA brokers and clearing houses). The front-end, i.e. the end host to AAA client communication, is in case of network access authentication offered by link layer protocols that forward authentication protocol exchanges back-and-forth. An example of a large scale Radius-based federation is EDUROAM.

Is it possible to design a system that builds on top of successful protocols to offer non-Web-based protocols with a solid starting point for authentication and authorization in a distributed system?

2.  Terminology

This document uses identity management and privacy terminology from [I‑D.hansen‑privacy‑terminology] (Pfitzmann, A., Hansen, M., and H. Tschofenig, “Terminology for Talking about Privacy by Data Minimization: Anonymity, Unlinkability, Undetectability, Unobservability, Pseudonymity, and Identity Management,” August 2010.).

3.  Architecture

Section 1 (Introduction) already introduced the federated access architecture, with the illustration of the different actors that need to interact, but it did not expand on the specifics of providing support for non-Web based applications. This section details this aspect and motivates design decisions. The main theme of the work described in this document is focused on re-using existing building blocks that have been deployed already and to re-arrange them in a novel way.

Although this architecture assumes updates to both the relying party as well as to the end host for application integration, those changes are kept at a minimum. A mechanism that can demonstrate deployment benefits (based on ease of update of existing software, low implementation effort, etc.)is preferred and there may be a need to specify multiple mechanisms to support the range of different deployment scenarios.

There are a number of ways for encapsulating EAP into an application protocol. For ease of integration with a wide range of non-Web based application protocols the usage of the GSS-API was chosen. Encapsulating EAP into the GSS-API also allows EAP to be used in SASL. A description of the technical specification can be found in [I‑D.ietf‑abfab‑gss‑eap] (Hartman, S. and J. Howlett, “A GSS-API Mechanism for the Extensible Authentication Protocol,” October 2010.). Other alternatives exist as well and may be considered later, such as "TLS using EAP Authentication" [I‑D.nir‑tls‑eap] (Nir, Y., Sheffer, Y., Tschofenig, H., and P. Gutmann, “TLS using EAP Authentication,” July 2010.).

There are several architectural layers in the system; this section discusses the individual layers.

3.1.  Federation Substrate

The federation substrate is responsible for the connunication between the relying party and the identity provider. This layer is responsible for the inter-domain communication and for the technical mechanisms necessary to establish inter-domain trust.

A key design goal is the re-use of an existing infrastructure, we build upon the AAA framework as utilized by RADIUS [RFC2138] (Rigney, C., Rigney, C., Rubens, A., Simpson, W., and S. Willens, “Remote Authentication Dial In User Service (RADIUS),” April 1997.) and Diameter [RFC3588] (Calhoun, P., Loughney, J., Guttman, E., Zorn, G., and J. Arkko, “Diameter Base Protocol,” September 2003.). Since this document does not aim to re-describe the AAA framework the interested reader is referred to [RFC2904] (Vollbrecht, J., Calhoun, P., Farrell, S., Gommans, L., Gross, G., de Bruijn, B., de Laat, C., Holdrege, M., and D. Spence, “AAA Authorization Framework,” August 2000.). Building on the AAA infrastructure, and RADIUS and Diameter as protocols, modifications to that infrastructure is to be avoided. Also, modifications to AAA servers should be kept at a minimum.

One demand that the AAA substrate must make of the upper layers is that they must properly identify the end points of the communication. That is- it must be possible for the AAA server at the RP to determine where to send each radius or diameter message. Otherwise, it is the RP's responsibility to determine the identity of the principal on its own, without the assistance of an IdP. This architecture makes use of the Network Access Identifier (NAI), where the IdP is indicated in the realm component [RFC4282] (Aboba, B., Beadles, M., Arkko, J., and P. Eronen, “The Network Access Identifier,” December 2005.). The NAI is represented and consumed by the GSS-API layer as GSS_C_NT_USER_NAME as specified in [RFC2743] (Linn, J., “Generic Security Service Application Program Interface Version 2, Update 1,” January 2000.). XXX Where is EAP here?

Once an IdP has been determined by the RP, it or its proxy agent must determine whether or not the IdP itself is authorized to make assertions, as it will likely not blindly accept any old provider. Federations serve this purpose. This architecture provides for three approaches to resolve whether an IdP is authorized:

Static Configuration:

In this case, the federation provides the RP or its proxy agent with a static list of IdPs that it may trust.

Federation Dynamic Referral

In this case, the federation provides a proxy of its own that will in some way authorize the IdP to the RP, and visa versa, as not all RPs may be authorized to use all IdPs for all purposes within a federation. N.B., because the identity of the principal is likely unknown at this point, it will not be possible for a federation to authorize an IdP to an RP based on the identity of the principal.

Federation Proxy:

In this case, the authentication request is forwarded to a federation proxy, who then further forwards the request to the IdP.

In the first two cases, it is expected that RPs will be configured to consult multiple federations, as a matter of practice. The first successful query is sufficient for the RP to then contact the IdP's AAA server.

The astute reader will notice that RADIUS and Diameter have substantially similar characteristics. Why not pick one? A key difference is that today RADIUS is largely transported upon UDP, and its use is largely, though not exclusively, intra-domain. Diameter itself was designed to scale to broader uses. We leave as a deployment decision, which protocol will be appropriate.

Through the integrity protection mechanisms in the AAA framework, the relying party can establish technical trust that messages are being sent by the appropriate relying party. Any given interaction will be associated with one federation at the policy level. The legal or business relationship defines what statements the identity provider is trusted to make and how these statements are interpreted by the relying party. The AAA framework also permits the relying party or elements between the relying party and identity provider to make statements about the relying party.

The AAA framework provides transport for attributes. Statements made about the subject by the identity provider, statements made about the relying party and other information is transported as attributes.

3.2.  Subject To Identity Provider

Traditional web federation does not describe how a subject communicates with an identity provider. As a result, this communication is not standardized. There are several disadvantages to this approach. It is difficult to have subjects that are machines rather than humans that use some sort of programatic credential. In addition, use of browsers for authentication restricts the deployment of more secure forms of authentication beyond plaintext username and password known by the server. In a number of cases the authentication interface may be presented before the subject has adequately validated they are talking to the intended server. By giving control of the authentication interface to a potential attacker, then the security of the system may be reduced and phishing opportunities introduced.

As a result, it is desirable to choose some standardized approach for communication between the subject's end-host and the identity provider. There are a number of requirements this approach must meet.

Experience has taught us one key security and scalability requirement: it is important that the relying party not get in possession of the long-term secret of the entity being authenticated by the AAA server. Aside from a valuable secret being exposed, a synchronization problem can also often develop. Since there is no single authentication mechanism that will be used everywhere there is another associated requirement: The authentication framework must allow for the flexible integration of authentication mechanisms. For instance, some identity providers may require hardware tokens while others may use passwords. A service provider would want to support both sorts of federations, and others.

Fortunately, these requirements can be met by utilizing standardized and successfully deployed technology, namely by the Extensible Authentication Protocol (EAP) framework [RFC3748] (Aboba, B., Blunk, L., Vollbrecht, J., Carlson, J., and H. Levkowetz, “Extensible Authentication Protocol (EAP),” June 2004.). Figure 2 (Architecture for Federated Access of non-Web based Applications) illustrates the integration graphically.

EAP is an end-to-end framework; it provides for two-way communication between a peer (i.e,service client or principal) through the authenticator (i.e., service provider) to the back-end (i.e., identity provider). Conveniently, this is precisely the communication path that is needed for federated identity. Although EAP support is already integrated in AAA systems (see [RFC3579] (Aboba, B. and P. Calhoun, “RADIUS (Remote Authentication Dial In User Service) Support For Extensible Authentication Protocol (EAP),” September 2003.) and [RFC4072] (Eronen, P., Hiller, T., and G. Zorn, “Diameter Extensible Authentication Protocol (EAP) Application,” August 2005.)) several challenges remain: one is to carry EAP payloads from the end host to the relying party. Another is to verify statements the relying party has made to the subject, confirm these statements are consistent with statements made to the identity provider and confirm all the above are consistent with the federation and any federation-specific policy or configuration. Another challenge is choosing which identity provider to use for which service.

3.3.  Application to Service

One of the remaining layers is responsible for integration of federated authentication into the application. There are a number of approaches that applications have adopted for security. So, there may need to be multiple strategies for integration of federated authentication into applications. However, we have started with a strategy that provides integration to a large number of application protocols.

Many applications such as SSH [RFC4462] (Hutzelman, J., Salowey, J., Galbraith, J., and V. Welch, “Generic Security Service Application Program Interface (GSS-API) Authentication and Key Exchange for the Secure Shell (SSH) Protocol,” May 2006.), NFS [RFC2203] (Eisler, M., Chiu, A., and L. Ling, “RPCSEC_GSS Protocol Specification,” September 1997.), DNS [RFC3645] (Kwan, S., Garg, P., Gilroy, J., Esibov, L., Westhead, J., and R. Hall, “Generic Security Service Algorithm for Secret Key Transaction Authentication for DNS (GSS-TSIG),” October 2003.) and several non-IETF applications support the Generic Security Services Application Programming Interface [RFC2743] (Linn, J., “Generic Security Service Application Program Interface Version 2, Update 1,” January 2000.). Many applications such as IMAP, SMTP, XMPP and LDAP support e Simple Authentication and Security Layer (SASL) [RFC4422] (Melnikov, A. and K. Zeilenga, “Simple Authentication and Security Layer (SASL),” June 2006.) framework. These two approaches work together nicely: by creating a GSS-API mechanism, SASL integration is also addressed [RFC5801] (Josefsson, S. and N. Williams, “Using Generic Security Service Application Program Interface (GSS-API) Mechanisms in Simple Authentication and Security Layer (SASL): The GS2 Mechanism Family,” July 2010.). In effect, using a GSS-API mechanism with SASL simply requires placing some headers on the front of the mechanism and constraining certain GSS-API options.

GSS-API is specified in terms of an abstract set of operations which can be mapped into a programming language to form an API. When people are first introduced to GSS-API, they focus on it as an API. However, from the prospective of authentication for non-web applications, GSS-API should be thought of as a protocol not an API. It consists of some abstract operations such as the initial context exchange, which includes two sub-operations (gss_init_sec_context and gss_accept_sec_context). An application defines which abstract operations it is going to use and where messages produced by these operations fit into the application architecture. A GSS-API mechanism will define what actual protocol messages result from that abstract message for a given abstract operation. So, since this work is focusing on a particular GSS-API mechanism, we generally focus on protocol elements rather than the API view of GSS-API.

The API view has significant value. Since the abstract operations are well defined, the set of information that a mechanism gets from the application is well defined. Also, the set of assumptions the application is permitted to make is generally well defined. As a result, an application protocol that supports GSS-API or SASL is very likely to be usable with a new approach to authentication including this one with no required modifications. In some cases, support for a new authentication mechanism has been added using plugin interfaces to applications without the application being modified at all. Even when modifications are required, they can often be limited to supporting a new naming and authorization model. For example, this work focuses on privacy; an application that assumes it will always obtain an identifier for the principal will need to be modified to support anonymity, unlinkability or pseudonymity.

So, we use GSS-API and SASL because a number of the application protocols we wish to federate support these strategies for security integration. What does this mean from a protocol standpoint and how does this relate to other layers? This means we need to design a concrete GSS-API mechanism. We have chosen to use a GSS-API mechanism that encapsulates EAP authentication. So, GSS-API (and SASL) encapsulate EAP between the end-host and the service. The AAA framework encapsulates EAP between the relying party and the identity provider. The GSS-API mechanism includes rules about how principals and services are named as well as per-message security and other facilities required by the applications we wish to support.

3.4.  Personalization Layer

The AAA framework provides a way to transport statements from the identity provider to the relying party. However, we also need to say more about the content of these statements. In simple cases, attributes particular to the AAA protocol can be defined. However in more complicated situations it is strongly desirable to re-use an existing protocol for asking questions and receiving information about subjects. SAML is used for this.

SAML usage may be as simple as the identity provider including a SAML Response message in the AAA response. Alternatively the relying party may generate a SAML request.

3.5.  Tieing Layers Together

                                 +--------------+
                                 |  AAA Server  |
                                 |  (Identity   |
                                 |  Provider)   |
                                 +-^----------^-+
                                   * EAP      | RADIUS/
                                   *          | Diameter
                                 --v----------v--
                              ///                \\\
                            //                      \\   ***
                           |        Federation        |  back-
                           |                          |  end
                            \\                      //   ***
                              \\\                ///
                                 --^----------^--
                                   * EAP      | RADIUS/
                 Application       *          | Diameter
+-------------+  Data            +-v----------v--+
|             |<---------------->|               |
| Client      |  EAP/EAP Method  | Server Side   |
| Application |<****************>| Application   |
| @ End Host  |  GSS-API         |(Relying Party)|
|             |<---------------->|               |
|             |  Application     |               |
|             |  Protocol        |               |
|             |<================>|               |
+-------------+                  +---------------+
               *** front-end ***

Legend:

 <****>: End-to-end exchange
 <---->: Hop-by-hop exchange
 <====>: Protocol through which GSS-API/GS2 exchanges are tunnelled

4.  Application Security Services

One of the key goals is to integrate federated authentication into existing application protocols and where possible, existing implementations of these protocols. Another goal is to perform this integration while meeting the best security practices of the technologies used to perform the integration. This section describes security services and properties required by the EAP GSS-API mechanism in order to meet these goals. This information could be viewed as specific to that mechanism. However, other future application integration strategies are very likely to need similar services. So, it is likely that these services will be expanded across application integration strategies if new application integration strategies are adopted.

4.1.  Server (Mutual) Authentication

GSS-API provides an optional security service called mutual authentication. This service means that in addition to the initiator providing (potentially anonymous or pseudonymous) identity to the acceptor, the acceptor confirms its identity to the initiator. Especially for the ABFAB context, this service is confusingly named. We still say that mutual authentication is provided when the identity of an acceptor is strongly authenticated to an anonymous initiator.

RFC 2743 does not explicitly talk about what mutual authentication means. Within the GSS-API community successful mutual authentication has come to mean:

• If a target name is supplied by the initiator, then the initiator trusts that the supplied target name describes the acceptor. This implies both that appropriate cryptographic exchanges took place for the initiator to make such a trust decision, and that after evaluating the results of these exchanges, the initiator's policy trusts that the target name is accurate.
• The initiator trusts that its idea of the acceptor name correctly names the entity it is communicating with.
• Both the initiator and acceptor have the same key material for per-message keys and both parties have confirmed they actually have the key material. In EAP terms, there is a protected indication of success.

Mutual authentication is an important defense against certain aspects of phishing. Intuitively, users would like to assume that if some party asks for their credentials as part of authentication, successfully gaining access to the resource means that they are talking to the expected party. Without mutual authentication, the acceptor could "grant access" regardless of what credentials are supplied. Mutual authentication better matches this user intuition.

The GSS-EAP mechanism MUST implement mutual authentication. That is, an initiator needs to be able to request mutual authentication. When mutual authentication is requested, only EAP methods capabale of providing the necessary service can be used, and appropriate steps need to be taken to provide mutual authentication. A broader set of EAP methods could be supported when a particular application does not request mutual authentication. It is an open question whether the mechanism will permit this.

4.2.  GSS-API Channel Binding

[RFC5056] (Williams, N., “On the Use of Channel Bindings to Secure Channels,” November 2007.) defines a concept of channel binding to prevent man-in-the-middle attacks. It is common to provide SASL and GSS-API with another layer to provide transport security; Transport Layer Security (TLS) is the most common such layer. TLS provides its own server authentication. However there are a variety of situations where this authentication is not checked for policy or usability reasons. Even when it is checked, if the trust infrastructure behind the TLS authentication is different from the trust infrastructure behind the GSS-API mutual authentication. If the endpoints of the GSS-API authentication are different than the endpoints of the lower layer, this is a strong indication of a problem such as a man-in-the-middle attack. Channel binding provides a facility to determine whether these endpoints are the same.

The GSS-EAP mechanism needs to support channel binding. When an application provides channel binding data, the mechanism needs to confirm this is the same on both sides consistent with the GSS-API specification. XXXThere is an open question here as to the details; today RFC 5554 governs. We could use that and the current draft assumes we will. However in Beijing we became aware of some changes to these details that would make life much better for GSS authentication of HTTP. We should resolve this with kitten and replace this note with a reference to the spec we're actually following.

Typically when considering channel binding, people think of channel binding in combination with mutual authentication. This is sufficiently common that without additional qualification channel binding should be assumed to imply mutual authentication. Without mutual authentication, only one party knows that the endpoints are correct. That's sometimes useful. Consider for example a user who wishes to access a protected resource from a shared whiteboard in a conference room. The whiteboard is the initiator; it does not need to actually authenticate that it is talking to the correct resource because the user will be able to recognize whether the displayed content is correct. If channel binding were used without mutual authentication, it would in effect be a request to only disclose the resource in the context of a particular channel. Such an authentication would be similar in concept to a holder-of-key SAML assertion. However, also note that while it is not happening in the protocol, mutual authentication is happening in the overall system: the user is able to visually authenticate the content. This is consistent with all uses of channel binding without protocol level mutual authentication found so far.

RFC 5056 channel binding (also called GSS-API channel binding when GSS-API is involved) is not the same thing as EAP channel binding. EAP channel binding is also used in the ABFAB context in order to implement acceptor naming and mutual authentication. Details are discussed in the mechanisms specification [I‑D.ietf‑abfab‑gss‑eap] (Hartman, S. and J. Howlett, “A GSS-API Mechanism for the Extensible Authentication Protocol,” October 2010.).

4.3.  Host-Based Service Names

IETF security mechanisms typically take the name of a service entered by a user and make some trust decision about whether the remote party in an interaction is the intended party. GSS-API has a relatively flexible naming architecture. However most of the IETF applications that use GSS-API, including SSH, NFS, IMAP, LDAP and XMPP, have chosen to use host-based service names when they use GSS-API. In this model, the initiator names an acceptor based on a service such as "imap" or "host" (for login services such as SSH) and a host name.

Using host-based service names leads to a challenging trust delegation problem. Who is allowed to decide whether a particular hostname maps to an entity. The public-key infrastructure (PKI) used by the web has chosen to have a number of trust anchors (root certificate authorities) each of wich can map any name to a public key. A number of GSS-API mechanisms suchs as Kerberos [RFC1964] (Linn, J., “The Kerberos Version 5 GSS-API Mechanism,” June 1996.) split the problem into two parts. A new concept called a realm is introduced. Then the mechanism decides what realm is responsible for a given name. That realm is responsible for deciding if the acceptor entity is allowed to claim the name. ABFAB needs to adopt this approach.

Host-based service names do not work ideally when different instances of a service are running on different ports. Also, these do not work ideally when SRV record or other insecure referrals are used.

The GSS-EAP mechanism needs to support host-based service names in order to work with existing IETF protocols.

4.4.  Per-Message Tokens

GSs-API provides per-message security services that can provide confidentiality and integrity. Some IETF protocols such as NFS and SSH take advantage of these services. As a result GSS-EAP needs to support these services. As with mutual authentication, per-message services will limit the set of EAP methods that are available. Any method that produces a Master Session Key (MSK) should be able to support per-message security services.

GSS-API provides a pseudo-random function. While the pseudo-random function does not involve sending data over the wire, it provides an algorithm that both the initiator and acceptor can run in order to arrive at the same key value. This is useful for designs where a successful authentication is used to key some other function. This is similar in concept to the TLS extractor. No current IETF protocols require this. However GSS-EAP supports this service because it is valuable for the future and easy to do given per-message services. Non-IETF protocols are expected to take advantage of this in the near future.

5.  Privacy Considerations

Sharing identity information may lead to privacy violations. A future verison of this document will provide a discussion of privacy considerations in a federated access environment.

6.  Deployment Considerations

6.1.  EAP Channel Binding

Discuss the implications of needing EAP channel binding.

6.2.  AAA Proxy Behavior

Discuss deployment implications of our proxy requirements.

7.  Security Considerations

This entire document is about security. A future version of the document will highlight some important security concepts.

8.  IANA Considerations

This document does not require actions by IANA.

9.  Acknowledgments

We would like to thank Mayutan Arumaithurai and Klaas Wierenga for their feedback. Additionally, we would like to thank Eve Maler, Nicolas Williams, Bob Morgan, Scott Cantor, Jim Fenton, and Luke Howard for their feedback on the federation terminology question.

Furthermore, we would like to thank Klaas Wierenga for his review of the pre-00 draft version.

10.  References

10.1. Normative References

10.2. Informative References

Authors' Addresses