AAA Working Group Pat R. Calhoun Internet-Draft Sun Microsystems, Inc. Category: Standards Track Haseeb Akhtar Nortel Networks Jari Arkko Oy LM Ericsson Ab Erik Guttman Sun Microsystems, Inc. Allan C. Rubens Tut Systems, Inc. Glen Zorn Cisco Systems, Inc. June 2001 Diameter Base Protocol Status of this Memo This document is an Internet-Draft and is in full conformance with all provisions of Section 10 of RFC2026. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet-Drafts. 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." The list of current Internet-Drafts can be accessed at: http://www.ietf.org/ietf/1id-abstracts.txt The list of Internet-Draft Shadow Directories can be accessed at: http://www.ietf.org/shadow.html. Distribution of this memo is unlimited. Copyright (C) The Internet Society 2001. All Rights Reserved. Calhoun et al. expires December 2001 [Page 1] Internet-Draft June 2001 Abstract The Diameter base protocol is intended to provide a AAA framework for Mobile-IP, NASREQ and ROAMOPS. This draft specifies the message format, transport, error reporting and security services to be used by all Diameter applications and MUST be supported by all Diameter implementations. Table of Contents 1.0 Introduction 1.1 Diameter Protocol 1.2 Requirements language 1.3 Terminology 2.0 Protocol Overview 2.1 Transport 2.1.1 SCTP Guidelines 2.2 Securing Diameter Messages 2.3 Diameter Protocol Extensibility 2.3.1 Defining new AVP Values 2.3.2 Creating new AVPs 2.3.3 Creating a new Diameter Applications 2.3.4 Application authentication procedures 2.4 Diameter Applications 2.5 Role of Diameter Agents 2.5.1 Relay Agents 2.5.2 Proxy Agents 2.5.3 Redirector Agents 2.5.4 Translation Agents 2.6 Diameter Server Discovery 2.7 Diameter Identity Encoding 3.0 Diameter Header 3.1 Command Code Definitions 3.2 Command Code ABNF specification 3.3 Diameter Command Naming Conventions 4.0 Diameter AVPs 4.1 AVP Header 4.2 Optional Header Elements 4.3 AVP Data Formats 4.4 Grouped AVP Values 4.4.1 Example AVP with a Grouped Data type 4.5 Diameter Base Protocol AVPs 5.0 Diameter message processing 5.1 Processing Local Messages 5.2 Message Forwarding 5.2.1 Peer Table 5.3 Message Routing Calhoun et al. expires December 2001 [Page 2] Internet-Draft June 2001 5.3.1 Realm-Based Routing Table 5.3.2 Redirecting requests 5.3.3 Relaying and Proxying Requests 5.3.4 Relaying and Proxying Answers 5.3.5 Hiding Network Topology 5.4 Origin-Host AVP 5.5 Origin-Realm AVP 5.6 Destination-Host AVP 5.7 Destination-Realm AVP 5.8 Routing AVPs 5.8.1 Route-Record AVP 5.8.2 Proxy-Info AVP 5.8.3 Proxy-Host AVP 5.8.4 Proxy-State AVP 5.9 Redirect-Host AVP 6.0 Capabilities Negotiation 6.1 Application Identifiers 6.2 Capabilities-Exchange-Request 6.3 Capabilities-Exchange-Answer 6.4 Vendor-Id AVP 6.5 Firmware-Revision AVP 6.6 Auth-Application-Id AVP 6.7 Host-IP-Address AVP 6.8 Supported-Vendor-Id AVP 6.9 Product-Name AVP 6.10 Acct-Application-Id AVP 6.11 Vendor-Specific-Application-Id AVP 7.0 Transport Failure Detection 7.1 Device-Watchdog-Request 7.2 Device-Watchdog-Answer 7.3 Failover/Failback Procedures 8.0 Peer State Machine 8.1 States 8.2 Events 8.3 Actions 8.4 The Election Process 9.0 Error Handling 9.1 Result-Code AVP 9.1.1 Informational 9.1.2 Success 9.1.3 Protocol Errors 9.1.4 Transient Failures 9.1.5 Permanent Failures 9.2 Message-Reject-Answer 9.3 Error-Message AVP 9.4 Error-Reporting-Host AVP 9.5 Failed-AVP AVP 10.0 "User" Sessions Calhoun et al. expires December 2001 [Page 3] Internet-Draft June 2001 10.1 Authorization Session State Machine 10.2 Accounting Session State Machine 10.3 Session-Id AVP 10.4 Authorization-Lifetime AVP 10.5 Session-Timeout AVP 10.6 User-Name AVP 10.7 Session Termination 10.7.1 Session-Termination-Request 10.7.2 Session-Termination-Answer 10.8 Aborting a Session 10.8.1 Abort-Session-Request 10.8.2 Abort-Session-Answer 10.9 Termination-Cause AVP 10.10 Inferring Session Termination from Origin-State-Id 10.11 Origin-State-Id AVP 11.0 Accounting 11.1 Server Directed Model 11.2 Protocol Messages 11.3 Application document requirements 11.4 Fault Resilience 11.5 Accounting Records 12.0 Accounting Command-Codes 12.1 Accounting-Request 12.2 Accounting-Answer 12.3 Accounting-Poll-Ind 13.0 Accounting AVPs 13.1 Accounting-Record-Type AVP 13.2 Accounting-Interim-Interval AVP 13.3 Accounting-Record-Number AVP 13.4 Accounting-Session-Id AVP 13.5 Accounting-Multi-Session-Id AVP 14.0 AVP Occurrence Table 14.1 Base Protocol Command AVP Table 14.2 Accounting AVP Table 15.0 IANA Considerations 15.1 AVP Header 15.1.1 AVP Code 15.1.2 AVP Flags 15.2 Diameter Header 15.2.1 Command Codes 15.2.2 Message Flags 15.3 Application Identifier Values 15.4 Result-Code AVP Values 15.5 Accounting-Record-Type AVP Values 15.6 Termination-Cause AVP Values 15.7 Diameter TCP/SCTP Port Numbers 16.0 Diameter protocol related configurable parameters 17.0 Security Considerations Calhoun et al. expires December 2001 [Page 4] Internet-Draft June 2001 18.0 References 19.0 Acknowledgements 20.0 Authors' Addresses 21.0 Full Copyright Statement 22.0 Expiration Date Appendix A. Diameter Service Template Calhoun et al. expires December 2001 [Page 5] Internet-Draft June 2001 1.0 Introduction Historically, the RADIUS protocol has been used to provide AAA services for dial-up PPP [42] and terminal server access. Over time, routers and network access servers (NAS) have increased in complexity and density, making the RADIUS protocol increasingly unsuitable for use in such networks. The Roaming Operations Working Group (ROAMOPS) has published a set of specifications [20, 43, 44] that define how a PPP user can gain access to the Internet without having to dial into his/her home service provider's modem pool. This is achieved by allowing service providers to cross-authenticate their users. Effectively, a user can dial into any service provider's point of presence (POP) that has a roaming agreement with his/her home Internet service provider (ISP), the benefit being that the user does not have to incur a long distance charge while traveling, which can sometimes be quite expensive. Given the number of ISPs today, ROAMOPS realized that requiring each ISP to set up roaming agreements with all other ISPs did not scale. Therefore, the working group defined a "broker", which acts as an intermediate server, whose sole purpose is to set up these roaming agreements. A collection of ISPs and a broker is called a "roaming consortium". There are many such brokers in existence today; many also provide settlement services for member ISPs. The Mobile-IP Working Group has recently changed its focus to inter administrative domain mobility, which is a requirement for cellular carriers wishing to deploy IETF-based mobility protocols. The current cellular carriers requirements [22, 23] are very similar to the ROAMOPS model, with the exception that the access protocol is Mobile-IP [45] instead of PPP. The Diameter protocol was not designed from the ground up. Instead, the basic RADIUS model was retained while fixing the flaws in the RADIUS protocol itself. Diameter does not share a common protocol data unit (PDU) with RADIUS, but does borrow sufficiently from the protocol to ease migration. The basic concept behind Diameter is to provide a base protocol that can be extended in order to provide AAA services to new access technologies. Currently, the protocol only concerns itself with Internet access, both in the traditional PPP sense as well as taking into account the ROAMOPS model, and Mobile-IP. Although Diameter could be used to solve a wider set of AAA problems, we are currently limiting the scope of the protocol in order to Calhoun et al. expires December 2001 [Page 6] Internet-Draft June 2001 ensure that the effort remains focused on satisfying the requirements of network access. Note that a truly generic AAA protocol used by many applications might provide functionality not provided by Diameter. Therefore, it is imperative that the designers of new applications understand their requirements before using Diameter. 1.1 Diameter Protocol The Diameter protocol allows peers to exchange a variety of messages. The base protocol provides the following facilities: - Delivery of AVPs (attribute value pairs) - Capabilities negotiation, as required in [20] - Error notification - Extensibility, through addition of new commands and AVPs, as required in [21] All data delivered by the protocol is in the form of an AVP. Some of these AVP values are used by the Diameter protocol itself, while others deliver data associated with particular applications which employ Diameter. AVPs may be added arbitrarily to Diameter messages, so long as the required AVPs are included and AVPs which are explicitly excluded are not included. AVPs are used by base Diameter protocol to support the following required features: - Transporting of user authentication information, for the purposes of enabling the Diameter server to authenticate the user. - Transporting of service specific authorization information, between client and servers, allowing the peers to decide whether a user's access request should be granted. - Exchanging resource usage information, which MAY be used for accounting purposes, capacity planning, etc. - Relaying, proxying and re-directing of Diameter messages through a server hierarchy. The Diameter base protocol provides the minimum requirements needed for an AAA transport protocol, as required by NASREQ [21], Mobile IP [22, 23], and ROAMOPS [20]. The base protocol is not intended to be used by itself, and must be used with a Diameter application, such as Mobile IP [10]. The Diameter protocol was heavily inspired and builds upon the tradition of the RADIUS [1] protocol. See section 2.4. for more information on Diameter applications. Any node can initiate a request. In that sense, Diameter is a peer to peer protocol. In this document, a Diameter client is the device that normally initiates a request for authentication and/or authorization Calhoun et al. expires December 2001 [Page 7] Internet-Draft June 2001 of a user. A Diameter server is the device that either forwards the request to another Diameter server (known as a proxy), or one that performs the actual authentication and/or authorization of the user based on some profile. Given that the server MAY send unsolicited messages to clients, it is possible for the server to initiate such messages. An example of an unsolicited message would be for a request that the client issue an accounting update. 1.2 Requirements language In this document, the key words "MAY", "MUST", "MUST NOT", "optional", "recommended", "SHOULD", and "SHOULD NOT", are to be interpreted as described in [13]. 1.3 Terminology Accounting The act of collecting information on resource usage for the purpose of trend analysis, auditing, billing, or cost allocation. Accounting record A session record represents a summary of the resource consumption of a user over the entire session. Accounting gateways creating the session record may do so by processing interim accounting events or accounting events from several Authentication The act of verifying the identity of an entity (subject). Authorization The act of determining whether a requesting entity (subject) will be allowed access to a resource (object). AVP The Diameter protocol consists of a header followed by one or more Attribute-Value-Pair (AVP). The AVP includes a header and is used to encapsulation authentication, authorization or accounting information. Broker A broker is a business term commonly used in AAA infrastructures. A broker is either a relay, proxy or redirect server, and MAY be operated by roaming consortiums. Diameter Agent Calhoun et al. expires December 2001 [Page 8] Internet-Draft June 2001 A Diameter Agent is a host that is providing either server, relay, proxy or redirector services. Diameter Client A Diameter Client is a device at the edge of the network that performs access control. An example of a Diameter client is a Network Access Server (NAS) or a Foreign Agent (FA). Diameter Node A Diameter node is a host that implements the Diameter protocol, and acts either as a Client, or as a Proxy, Redirector, Server or Translation agent. Diameter Server A Diameter Server is one that handles authentication, authorization and accounting requests for a particular realm. By its very nature, a Diameter Server MUST support Diameter applications in addition to the base protocol. Downstream Server Diameter Proxy servers identify a downstream server as one that is providing routing services towards the Diameter client. Home Domain A Home Domain is the administrative domain with whom the user maintains an account relationship. Home Server See Diameter Server. Interim accounting An interim accounting message provides a snapshot of usage during a user's session. It is typically implemented in order to provide for partial accounting of a user's session in the event of a device reboot or other network problem that prevents the reception of a session summary message or session record. Local Domain A local domain is the administrative domain providing services to a user. An administrative domain MAY act as a local domain for certain users, while being a home domain for others. Network Access Identifier The Network Access Identifier, or NAI [3], is used in the Diameter protocol to extract a user's identity and realm. The identity is used to identify the user during authentication and/or authorization, while the realm is used for message routing purposes. Calhoun et al. expires December 2001 [Page 9] Internet-Draft June 2001 Proxy In addition to forwarding requests and responses, proxies enforce policies relating to resource usage and provisioning. This is typically accomplished by tracking the state of NAS devices. While proxies typically do not respond to client Requests prior to receiving a Response from the server, they may originate Reject messages in cases where policies are violated. As a result, proxies need to understand the semantics of the messages passing through them, and may not support all Diameter applications. Realm The string in the NAI that immediately follows the '@' character. NAI realm names are required to be unique, and are piggybacked on the administration of the DNS namespace. Diameter makes use of the realm, also loosely referred to as domain, to determine whether messages can be satisfied locally, or whether they must be proxied. Real-time Accounting Real-time accounting involves the processing of information on resource usage within a defined time window. Time constraints are typically imposed in order to limit financial risk. Relay Relays forward requests and responses based on routing-related AVPs and domain forwarding table entries. Since relays do not enforce policies, they do not examine or alter non-routing AVPs. As a result, relays never originate messages, do not need to understand the semantics of messages or non-routing AVPs, and are capable of handling any Diameter applications or message type. Since relays make decisions based on information in routing AVPs and domain forwarding tables they do not keep state on NAS resource usage or conversations in progress. Redirector Rather than forwarding requests and responses between clients and servers, Re-directs refer clients to servers and allow them to communicate directly. Since Re-directs do not sit in the forwarding path, they do not alter any AVPs transitting between client and server. Re-direct proxies do not originate messages and are capable of handling any message type, although they may be configured only to re-direct messages of certain types, while acting as Routing or Policy proxies for other types. As with Routing proxies, re-directs do not keep state with respect to conversations or NAS resources. Roaming Relationships Calhoun et al. expires December 2001 [Page 10] Internet-Draft June 2001 Roaming relationships include relationships between companies and ISPs, relationships among peer ISPs within a roaming association, and relationships between an ISP and a roaming consortia. Together, the set of relationships forming a path between a local ISP's authentication proxy and the home authentication server is known as the roaming relationship path. Session The Diameter protocol is session based. When an authorization request is initially transmitted, it includes a session identifier that is used for the duration of the session. The Session- Identifier AVP contains the identifier and must be globally unique. devices serving the same user. Upstream Server Diameter Proxy servers identify an upstream server as one that is providing routing services towards the home server for a particular message. 2.0 Protocol Overview The base Diameter protocol is never used on its own. It is always extended for a particular application. Three Diameter applications are defined by companion documents: NASREQ [7], Mobile IP [10], End-to-End Security [11]. These options are introduced in this document but specified elsewhere. Additional Diameter applications MAY be defined in the future (see Section 15.3). Diameter Clients MUST support the base protocol, which includes accounting. In addition, they MUST fully support each Diameter application which is needed to implement the client's service, e.g. NASREQ and/or Mobile IP. A Diameter Client which does not support both NASREQ and Mobile IP, MUST be referred to as "Diameter X Client" where X is the application which it supports, and not a "Diameter Client." Diameter Servers must support the base protocol, which includes accounting. In addition, they MUST fully support each Diameter application which is needed to implement the intended service, e.g. NASREQ and/or Mobile IP. A Diameter Server which does not support both NASREQ and Mobile IP, MUST be referred to as "Diameter X Server" where X is the application which it supports, and not a "Diameter Server." Diameter Relays and Redirectors are, by definition, protocol transparent, and MUST transparently support the Diameter base protocol, which includes accounting, and all Diameter applications. Calhoun et al. expires December 2001 [Page 11] Internet-Draft June 2001 Diameter Proxies MUST suppport the base protocol, which includes accounting. in addition, they MUST fully support each Diameter application which is needed to implement proxied services, e.g. NASREQ and/or Mobile IP. A Diameter Proxy which does not support also both NASREQ and Mobile IP, MUST be referred to as "Diameter X Proxy" where X is the application which it supports, and not a "Diameter Proxy." The base Diameter protocol concerns itself with capabilities negotiation, and how messages are sent and how peers may eventually be abandoned. The base protocol also defines certain rules which apply to all exchanges of messages between Diameter peers. Communication between Diameter peers begins with one peer sending a message to another Diameter peer. The set of AVPs included in the message is determined by a particular Diameter application. One AVP that is included to reference a user's session is the Session-Id. The initial request for authentication and/or authorization of a user would include the Session-Id. The Session-Id is then used in all subsequent messages to identify the user's session (see section 10.0 for more information). The communicating party may accept the request, or reject it by returning an answer message with Result-Code AVP set to indicate an error occurred. The specific behavior of the diameter server or client receiving a request depends on the Diameter application employed. Session state (associated with a Session-Id) MUST be freed upon receipt of the Session-Termination-Request, Session-Termination- Answer, expiration of authorized service time in the Session-Timeout AVP, and according to rules established in a particular Diameter application. The Diameter base protocol provides the Authorization-Lifetime AVP, which MAY be used by applications to specify the duration of a specific authorized session. 2.1 Transport The base Diameter protocol is run on port TBD of both TCP [27] and SCTP [26] transport protocols (for interoperability test purposes port 1812 will be used until IANA assigns a port to the protocol). When used with TLS [38], The Diameter protocol is run on port TBD of both TCP and SCTP. Diameter clients MUST support either TCP or SCTP, while agents and servers MUST support both. Future versions of this specification MAY Calhoun et al. expires December 2001 [Page 12] Internet-Draft June 2001 mandate that clients support SCTP. A Diameter node MAY initiate connections from any source port, but MUST be prepared to receive connections on port TBD. Note that the source and destination addresses used in request and replies MAY any of a peer's valid IP addresses. A given Diameter process SHOULD use the same port number to send all messages to aid in identifying which process sent a given message. More than one Diameter process MAY exist within a single host, so the sender's port number is needed to discriminate them. When no transport connection exists with a peer, an attempt to connect SHOULD be periodically attempted. This behavior is handled via the Tc timer, whose recommended value is 30 seconds. 2.1.1 SCTP Guidelines The following are guidelines for Diameter implementations that support SCTP: 1. For interoperability: All Diameter nodes MUST be prepared to receive Diameter messages on any SCTP stream in the association. 2. To prevent blocking: All Diameter nodes SHOULD utilize all SCTP streams available to the association to prevent head-of-the- line blocking. 2.2 Securing Diameter Messages Diameter clients, such as Network Access Servers (NASes) and Foreign Agents MUST support IP Security [37], and MAY support TLS [38]. Diameter servers MUST support TLS, but the administrator MAY opt to configure IPSec instead of using TLS. Operating the Diameter protocol without any security mechanism is not recommended. 2.3 Diameter Protocol Extensibility There are various ways the Diameter protocol can be extended. This section is intended to assist protocol designers in selecting the best method of using the Diameter protocol. 2.3.1 Defining new AVP Values Calhoun et al. expires December 2001 [Page 13] Internet-Draft June 2001 Defining a new AVP value is the best approach when a new application needs to make use of an existing Diameter application, but requires that an existing AVP communicate different service-specific information (e.g. NAS-Port-Type set to avian carriers). When an existing AVP can be used to communicate the new information, this approach is preferred over creating new AVPs. In order to allocate a new AVP value, a request MUST be sent to IANA, with a detailed explanation of the value. Furthermore, if the command code on which the AVP value is to be used would require a different set of mandatory AVPs be present, the list of AVPs must accompany the request. 2.3.2 Creating new AVPs New AVPs may be created when a new application requiring Diameter support can make use of an existing Diameter application, but requires new AVPs to communicate service-specific information. Prior to defining the AVP, the AVP type MUST be one of the types listed in section 4.3. In the event that a logical grouping of AVPs is necessary, and multiple "groups" are possible in a given command, it is highly recommended that a Grouped AVP be used (see Section 4.4). In order to create a new AVP, a request MUST be sent to IANA, with a detailed explanation of the AVP, its type and possible values. Furthermore, the request MUST include the commands that would make use of the AVP. Note that new AVPS to be used with an existing application MUST NOT be defined to have the 'M'andatory bit set. 2.3.3 Creating new Diameter Applications Should a new application require Diameter support, but it cannot fit within an existing application without requiring major changes to the specification, it may be desirable to create a new Diameter application. Major changes to an application include: - Requiring a whole different set of mandatory AVPs to a command - Requiring a command that has a different number of round trips to satisfy a request (e.g. application foo has a command that requires one round trip, but new application bar has a command that requires two round trips to complete). - The method used to authenticate the user is drastically Calhoun et al. expires December 2001 [Page 14] Internet-Draft June 2001 different from any existing application, and the authentication information cannot be carried within the AVPs defined in the application. Note that the creation of a new application should be viewed as a last resort. New Diameter applications MUST define at least one Command Code, the expected AVPs in an ABNF [31] grammar (see section 3.2), and MAY also define new AVPs. If the Diameter application has any accounting requirements, it MUST also specify the AVPs that are to be present in the Diameter Accounting messages (see section 11.3). When possible, a new Diameter application SHOULD attempt to re-use any existing Diameter AVP, in order to reduce the possibility of having multiple AVPs that carry similar information. Every Diameter application specification MUST have an IANA assigned Application Identifier (see section 2.4). 2.3.4 Application authentication procedures When possible, applications SHOULD be designed such that new authentication methods MAY be added without requiring changes to the application. This MAY require that new AVP values be assigned to represent the new authentication transform, or any other scheme that produces similar results. When possible, authentication frameworks, such as Extensible Authentication Protocol [25], SHOULD be used. 2.4 Diameter Application Compliance Application Identifiers are advertised during the capabilities exchange phase (see section 6.0). For a given application, there are two different ways of advertising support. First, advertising support of the application via the Auth-Application-Id implies that the sender supports all authentication and authorization command codes, and the AVPs specified in the associated ABNFs, described in the specification. Second, advertising support of the application via the Acct-Application-Id implies that the sender supports the Accounting command codes defined in this specification, as well as the accounting AVPs defined in the application's specification. An implementation MAY add arbitrary AVPs to any command defined in an application, including vendor-specific AVPs. However, implementations that add such AVPs with the Mandatory 'M' bit set are not compliant, and are at fault if the peer rejects the request. If the sender of Calhoun et al. expires December 2001 [Page 15] Internet-Draft June 2001 such a message wishes to provide service, it MUST resend the message with the offending AVPs removed. 2.5 Role of Diameter Agents In addition to client and servers, the Diameter protocol introduces relays, redirectors, proxies and translation gateways, each of which is defined in Section 1.3. These Diameter agents are useful for several reasons: - They can distribute administration of systems to a configurable grouping, including the maintenance of security associations. - They can be used for concentration of requests from an number of co-located or distributed NAS equipment sets to a set of like user groups. - They can do value-added processing to the requests or responses. - They can used for load balancing. - A complex network will have multiple authentication sources, they can sort requests and forward towards the correct target. The Diameter protocol requires that agents maintain transaction state, which is used for failover purposes. Transaction state implies that upon forwarding a request, it's Hop-by-Hop identifier is saved, the field is replaced with a locally unique identifier, which is restored to its original value when the corresponding answer is received. The request's state is released upon receipt of the answer. A stateless agent is one that only maintains transaction state. The Proxy-Info AVP allows stateless agent to add local state to a Diameter request, with the guarantee that the same state will be present in the answer. However, the protocol's failover procedures requires that agents maintain a copy of pending requests. A stateful agent is one that maintains session state information, by keeping track of all authorized active sessions. Each authorized session is bound to a particular service, and its state is considered active either until it is notified otherwise, or by expiration. Each authorized session has a expiration, which is communicated by Diameter servers via the Authorized-Lifetime AVP. Maintaining session state MAY be useful in certain applications, such as: - Protocol translation (e.g. RADIUS <-> Diameter) - Limiting resources authorized to a particular user - Per user or transaction auditing A Diameter agent MAY act in a stateful manner for some requests, while be stateless for others. A Diameter implementation MAY act as Calhoun et al. expires December 2001 [Page 16] Internet-Draft June 2001 one type of agent for some requests, and as another type of agent for others. 2.5.1 Relay Agents Relay Agents are Diameter agents that accept requests and routes the message to another Diameter agent based on information found in the message (e.g. Destination-Realm). This routing decision is performed using a list of supported domains, and known peers. This is known as the Diameter Routing Table, as is defined further in section x.x. Relays MAY be used to aggregate requests from multiple Network Access Servers (NASes) within a common geographical area (POP). The use of Relays is advantageous since it eliminates the need for NASes to be configured with the necessary security information it would otherwise require to communicate with Diameter servers in other realms. Likewise, this reduces the configuration load on Diameter servers that would otherwise be necessary when NASes are added, changed or deleted. Relays modify Diameter messages by inserting, and removing, routing information, but do not modify any other portion of a message. Further, Relays inherent simplicity implies that they are stateless, and therefore SHOULD NOT maintain session state, but MUST maintain transaction state. +------+ ---------> +------+ ---------> +------+ | | 1. Request | | 2. Request | | | NAS | | DRL | | HMS | | | 4. Answer | | 3. Answer | | +------+ <--------- +------+ <--------- +------+ mno.net mno.net abc.com Figure 1: Relaying of Diameter messages The example provided in Figure 1 depicts a request issued from NAS, which is an access device, for the user bob@abc.com. Prior to issuing the request, NAS performs a Diameter route lookup, using "abc.com" as the key, and determines that the message is to be relayed to DRL, which is a Diameter Relay. DRL performs the same route lookup as NAS, and relays the message to HMS, which is abc.com's Home Diameter Server. HMS identifies that the request can be locally supported (via the realm), processes the authentication and/or authorization request, and replies with an answer, which is routed back to NAS using Diameter routing AVPs. Since Relays do not perform any application level processing, they provide relaying services for all Diameter applications, and Calhoun et al. expires December 2001 [Page 17] Internet-Draft June 2001 therefore MUST advertise the Relay Application Identifier. 2.5.2 Proxy Agents Similarly to Relays, Proxy agents route Diameter messages using the Diameter Routing Table. However, they differ since they modify messages to implement policy enforcement. This requires that proxies maintain the state of their downstream peers (e.g. access devices) to enforce resource usage, provide admission control, and provisioning. It is important to note that although proxies MAY provide a value-add function for NASes, they do not allow access devices to use the Diameter End-to-End Security application, since modifying messages breaks end-to-end authentication. Proxies MAY be used in call control centers or access ISPs that provide outsourced connections, they can monitor the number and types of ports in use, and make allocation and admission decisions according to their configuration. Proxies that wish to limit resources MUST be stateful, and all Proxies MUST maintain transaction state. Proxy agents MUST NOT allow end-to-end security to be established between two peers if it expects to modify ANY non-routing AVP in messages exchanged between the peers. See [11] for more information. Since enforcing policies requires an understanding of the service being provided, Proxies MUST only advertise the Diameter applications they support. 2.5.3 Redirector Agents Redirector agents provide Realm to Server address resolution, and use the Diameter routing table to determine where a given request should be forwarded to. When a request is received by a Diameter redirector, a special answer is created, which includes the identity of the Diameter server(s) the originator of the request should contact directly. Redirectors are useful in scenarios where the Diameter routing configuration needs to be centralized. An example is a redirector that provides services to all members of a consortium, but does not wish to be burdened with relaying all messages between domains. This scenario is advantageous since it does not require that the consortium provide routing updates to its members when changes are Calhoun et al. expires December 2001 [Page 18] Internet-Draft June 2001 made to a member's infrastructure. Since redirectors do not relay messages, and only return an answer with the information necessary for Diameter agents to communicate directly, they do not modify messages, and therefore MUST NOT maintain session state. Further, since redirectors never relay requests, they are not required to maintain transaction state. +------+ | | | DRD | | | +------+ ^ | 2. Request | | 3. Redirection | | Notification | v +------+ ---------> +------+ ---------> +------+ | | 1. Request | | 4. Request | | | NAS | | DRL | | HMS | | | 6. Answer | | 5. Answer | | +------+ <--------- +------+ <--------- +------+ mno.net mno.net abc.com Figure 2: Redirecting a Diameter Message The example provided in Figure 2 depicts a request issued from the access device, NAS, for the user bob@abc.com. The message is forwarded by the NAS to its relay, DRL, which does not have a routing entry in its Diameter Routing Table for abc.com. DRL has a default route configured to DRD, which is a redirector that returns a redirect notification to DLR, as well as HMS' contact information. Upon receipt of the redirect notification, DRL establishes a transport connection with HMS, if one doesn't already exist, and forwards the request to it. Since Redirectors do not perform any application level processing, they provide relaying services for all Diameter applications, and therefore MUST advertise the Relay Application Identifier. 2.5.4 Translation Agents A Translation Agent is a device that provides translation between two protocols (e.g. RADIUS<->Diameter, TACACS+<->Diameter). Translation agents are likely to be used as aggregation servers to communicate with a Diameter infrastructure, while allowing for the embedded systems to be migrated at a slower pace. Calhoun et al. expires December 2001 [Page 19] Internet-Draft June 2001 Given that the Diameter protocol introduces the concept of long-lived authorized sessions, translation agents MUST be stateful and MUST maintain transaction state. Translation of messages can only occur if the agent recognizes the application of a particular request, and therefore MUST only advertise their locally supported applications. +------+ ---------> +------+ ---------> +------+ | | RADIUS Request | | Diameter Request | | | NAS | | TLA | | HMS | | | RADIUS Answer | | Diameter Answer | | +------+ <--------- +------+ <--------- +------+ mno.net mno.net abc.com Figure 3: Translation of RADIUS to Diameter 2.6 Diameter Agent Discovery Allowing for dynamic Diameter agent discovery will make it possible for simpler and more robust deployment of AAA services. In order to promote interoperable implementations of Diameter agent discovery, the following mechanisms are described. These are based on existing IETF standards. There are two cases where Diameter agent discovery may be performed. The first is when a Diameter client needs to discover a first-hop Diameter agent. The second case is when a Diameter agent needs to discover another agent - for further handling of a Diameter operation. In both cases, the following 'search order' is recommended: 1. The Diameter implementation consults its list of static (manual) configured Diameter agent locations. These will be used if they exist and respond. 2. The Diameter implementation uses SLPv2 [28] to discover Diameter services. The Diameter service template [32] is included in Appendix A. It is recommended that SLPv2 security be deployed (this requires distributing keys to SLPv2 agents.) This is discussed further in Appendix A. SLPv2 will allow Diameter implementations to discover the location of Diameter agents in the local site, as well as their characteristics. Diameter agents with specific capabilities (say support for the Mobile IP application) can be requested, and only those will be discovered. Calhoun et al. expires December 2001 [Page 20] Internet-Draft June 2001 3. The Diameter implementation uses DNS to request the SRV RR [33] for the '_diameter._sctp' and/or '_diameter._tcp' server in a particular domain. The Diameter implementation has to know in advance which domain to look for an Diameter agent in. This could be deduced, for example, from the 'realm' in a NAI that an Diameter implementation needed to perform an Diameter operation on. Diameter allows AAA peers to protect the integrity and privacy of communication as well as to perform end-point authentication. Still, it is prudent to employ DNS Security as a precaution when using DNS SRV RRs to look up the location of a Diameter agent. [34, 35, 36] 2.7 Diameter Identity Encoding Several Diameter AVPs are used to include a node's identity, such as the Destination-Host, Origin-Host, Route-Record, etc. The contents of such AVPs follow the Uniform Resource Identifiers (URI) syntax [29] rules specified below: Diameter-Identity = [protocol] fqdn [ port ] [ transport ] protocol-name = ( "diameter" | "radius" | "tacacs+" ) protocol = protocol-name "://" ; If absent, the default is "diameter://" fqdn = Fully Qualified Host Name port = ":" 1*DIGIT ; If absent, the default Diameter port (TBD) is assumed. transport = ";transport=" ( "tcp" | "sctp" | "udp") ; If absent, the default SCTP [26] protocol is assumed. ; UDP is ONLY used when the protocol is set to RADIUS The following are examples of valid Diameter host identities: host.abc.com:6666;transport=tcp diameter://host.abc.com diameter://host.abc.com:6666 diameter://host.abc.com;transport=tcp diameter://host.abc.com:6666;transport=tcp radius://host.abc.com:1813;transport=udp Calhoun et al. expires December 2001 [Page 21] Internet-Draft June 2001 3.0 Diameter Header A summary of the Diameter header format is shown below. The fields are transmitted in network byte order. 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Ver | Message Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |R r r r r r r r| Command-Code | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Vendor-ID | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Hop-by-Hop Identifier | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | End-to-End Identifier | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | AVPs ... +-+-+-+-+-+-+-+-+-+-+-+-+- Version This Version field MUST be set to 1 to indicate Diameter Version 1. Message Length The Message Length field is two octets and indicates the length of the Diameter message including the header fields. Command Flags The Command Flags field is eight bits. The following bits are assigned: R(equest) - If set, the message is a request. If cleared, the message is an answer. r(eserved) - this flag bit is reserved for future use, and MUST be set to zero. Command-Code The Command-Code field is three octets, and is used in order to communicate the command associated with the message. The 24-bit address space is managed by IANA (see section 15.2). Vendor-ID In the event that the Command-Code field contains a vendor specific command, the four octet Vendor-ID field contains the IANA assigned "SMI Network Management Private Enterprise Codes" [2] value. If the Command-Code field contains an IETF standard Calhoun et al. expires December 2001 [Page 22] Internet-Draft June 2001 Command, the Vendor-ID field MUST be set to zero (0). Any vendor wishing to implement a vendor-specific Diameter command MUST use their own Vendor-ID along with their privately managed Command- Code address space, guaranteeing that they will not collide with any other vendor's vendor-specific command, nor with future IETF applications. Hop-by-Hop Identifier The Hop-by-Hop Identifier field is four octets, and aids in matching requests and replies. The sender MUST ensure that the Hop-by-Hop identifier in a request is locally unique (to the sender) at any given time, and MAY attempt to ensure that the number is unique across reboots. The sender of an Answer message MUST ensure that the Hop-by-Hop Identifier field contains the same value that was found in the corresponding request. The Hop-by-Hop identifier is normally a monotonically increasing number, whose start value was randomly generated. An answer message that is received with an unknown Hop-by-Hop Identifier MUST be discarded. End-to-End Identifier Unlike the Hop-by-Hop Identifier, the End-to-End Identifier is used to detect duplicate messages, and relay agents MUST NOT modify this field. The sender of a request or answer message MUST insert a locally unique value in this field. The combination of the Origin-Host AVP and this field is used to detect duplicates. An Answer message which is received with a previously seen End- to-End Identifier, and is to be locally consumed (meaning that the Destination-Host AVP contains the local node's identity) SHOULD be silently discarded. AVPs AVPs are a method of encapsulating information relevant to the Diameter message. See section 4. for more information on AVPs. 3.1 Command Codes Each command Request/Answer pair is assigned a command code, and the sub-type (e.g. request or answer) is identified via the 'R' bit in the Command Flags field of the Diameter header. Every Diameter message MUST contain a command code in its header's Command-Code field, which is used to determine the action that is to be taken for a particular message. The following Command Codes are defined in the Diameter base protocol: Calhoun et al. expires December 2001 [Page 23] Internet-Draft June 2001 Command-Name Abbrev. Code Reference -------------------------------------------------------- Abort-Session-Request ASR 274 10.8.1 Abort-Session-Answer ASA 274 10.8.2 Accounting-Answer ACA 271 12.2 Accounting-Poll-Ind API 273 12.3 Accounting-Request ACR 271 12.1 Capabilities-Exchange- CER 257 6.2 Request Capabilities-Exchange- CEA 257 6.3 Answer Message-Reject-Answer MRA 282 9.2 Device-Watchdog-Answer DWA 280 7.2 Device-Watchdog-Request DWR 280 7.1 Session-Termination- STR 275 10.7.1 Request Session-Termination- STA 275 10.7.2 3.2 Command Code ABNF specification Every Command Code defined MUST include a corresponding ABNF specification, which is used to define the AVPs that MUST, MAY and MUST NOT be present. The following format is used in the definition: command-def = command-name "::=" diameter-message diameter-name = ALPHA *(ALPHA / DIGIT / "-") command-name = diameter-name ; The command-name has to be Command name, ; defined in the base or extended Diameter ; specifications. diameter-message = header [ *fixed] [ *required] [ *optional] [ *fixed] header = "" command-id = 1*DIGIT ; The Command Code assigned to the command r-bit = ", REQUEST" ; If present, the 'R' bit in the Command ; Flags is set, indicating that the message ; is a request, as opposed to an answer. fixed = [qual] "<" avp-spec ">" Calhoun et al. expires December 2001 [Page 24] Internet-Draft June 2001 required = [qual] "{" avp-spec "}" optional = [qual] "[" avp-name "]" ; The avp-name in the 'optional' rule cannot ; evaluate to any AVP Name which is included ; in a fixed or required rule. qual = [min] "*" [max] ; See ABNF conventions, RFC 2234 section 6.6. ; The absence of any qualifiers implies that ; one and only one such AVP MUST be present. ; ; NOTE: "[" and "]" have a different meaning ; than in ABNF (see the optional rule, above). ; These braces cannot be used to express ; optional fixed rules (such as an optional ; ICV at the end.) To do this, the convention ; is '0*1fixed'. min = 1*DIGIT ; The minimum number of times the element may ; be present. max = 1*DIGIT ; The maximum number of times the element may ; be present. avp-spec = diameter-name ; The avp-spec has to be an AVP Name, defined ; in the base or extended Diameter ; specifications. avp-name = avp-spec | "AVP" ; The string "AVP" stands for *any* arbitrary ; AVP Name, which does not conflict with the ; required or fixed position AVPs defined in ; the command code definition. The following is a definition of a fictitious command code: Example-Request ::= < Diameter-Header: 9999999, REQUEST > { User-Name } * { Origin-Host } * [ AVP ] 3.3 Diameter Command Naming Conventions Calhoun et al. expires December 2001 [Page 25] Internet-Draft June 2001 The following conventions are required for the naming of Diameter messages. Diameter commands typically start with an object name, and end with either the Request or Answer verb. The Request/Answer message pair is used when a Diameter node requests that some action be performed by a peer (e.g. authorize a user, terminate a session). The corresponding answer MUST contain either a positive or negative result code, informing the requester whether the request was successful or not. Other information MAY also be returned in the Answer message. Request and Answer messages share the same command code, and the R(equest) bit in the Diameter header is used to identify whether a message is the request or answer. 4.0 Diameter AVPs Diameter AVPs carry specific authentication, accounting and authorization information, security information as well as configuration details for the request and reply. Some AVPs MAY be listed more than once. The effect of such an AVP is specific, and is specified in each case by the AVP description. Each AVP of type OctetString MUST be padded to align on a 32 bit boundary, while other AVP types align naturally. NULL bytes are added to the end of the AVP Data field till a word boundary is reached. The length of the padding is not reflected in the AVP Length field. 4.1 AVP Header The fields in the AVP header MUST be sent in network byte order. The format of the header is: 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | AVP Code | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |V M P r r r r r| AVP Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Vendor-ID (opt) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Data ... +-+-+-+-+-+-+-+-+ Calhoun et al. expires December 2001 [Page 26] Internet-Draft June 2001 AVP Code The AVP Code, combined with the Vendor-Id field, identifies the attribute uniquely. The first 256 AVP numbers are reserved for backward compatibility with RADIUS and are to be interpreted as per NASREQ [7]. AVP numbers 256 and above are used for Diameter, which are allocated by IANA (see section 15.1). AVP Flags The AVP Flags field informs the receiver how each attribute must be handled. Note that subsequent Diameter applications MAY define bits to be used within the AVP Header, and an unrecognized bit should be considered an error. The 'r' and the reserved bits are unused and should be set to 0 and ignored on receipt, while the 'P' bit is defined in [11]. The 'M' Bit, known as the Mandatory bit, indicates whether support of the AVP is required. If an unrecognized AVP with the 'M' bit set is received by a Diameter node, the message MUST be rejected. Diameter Relay and Redirector agents MUST NOT reject messages with unrecognized AVPs. A Diameter node that sets the 'M' bit in an AVP that is not defined in a given message's ABNF is at fault if the message is rejected. In order to provide service to the user, the node at fault MUST re-issue a request either without the AVP, or without setting its 'M' bit. A Diameter node that rejects a message due to an unrecognized AVP with the 'M' bit set, and the AVP in question is defined in the message's ABNF is at fault. In most cases the initiator of the failing request will not provide service to the user. AVPs with the 'M' bit cleared are informational only and a receiver that receives a message with such an AVP that is not supported MAY simply ignore the AVP. The 'V' bit, known as the Vendor-Specific bit, indicates whether the optional Vendor-ID field is present in the AVP header. When set the AVP Code belongs to the specific vendor code address space. Unless otherwise noted, AVPs will have the following default AVP Flags field settings: The 'M' bit MUST be set. The 'V' bit MUST NOT be set. AVP Length The AVP Length field is three octets, and indicates the length of this AVP including the AVP Code, AVP Length, AVP Flags, Reserved, Calhoun et al. expires December 2001 [Page 27] Internet-Draft June 2001 the Vendor-ID field (if present) and the AVP data. If a message is received with an invalid attribute length, the message SHOULD be rejected. 4.2 Optional Header Elements The AVP Header contains one optional field. This field is only present if the respective bit-flag is enabled. Vendor-ID The Vendor-ID field is present if the 'V' bit is set in the AVP Flags field. The optional four octet Vendor-ID field contains the IANA assigned "SMI Network Management Private Enterprise Codes" [2] value, encoded in network byte order. Any vendor wishing to implement a vendor-specific Diameter AVP MUST use their own Vendor-ID along with their privately managed AVP address space, guaranteeing that they will not collide with any other vendor's vendor-specific AVP, nor with future IETF applications. A vendor ID value of zero (0) corresponds to the IETF adopted AVP values, as managed by the IANA. Since the absence of the vendor ID field implies that the AVP in question is not vendor specific, implementations SHOULD not use the zero (0) vendor ID. 4.3 AVP Data Formats The Data field is zero or more octets and contains information specific to the Attribute. The format and length of the Data field is determined by the AVP Code and AVP Length fields. The format of the Data field MAY be one of the following data types. The interpretation of the values depends on the specification of the AVP. For example, an OctetString may be used to transmit human readable string data and Unsigned32 may be used to transmit a time value. Conventions for these common interpretations are described below. OctetString The data contains arbitrary data of variable length. Unless otherwise noted, the AVP Length field MUST be set to at least 8 (12 if the 'V' bit is enabled). Data used to transmit (human readable) character string data uses the UTF-8 [24] character set and is NOT NULL-terminated. The minimum Length field MUST be 9, but can be set to any value up to 65504 bytes. AVP Values of this type that do not align on a 32-bit boundary MUST have the necessary padding. Calhoun et al. expires December 2001 [Page 28] Internet-Draft June 2001 Address 32 bit (IPv4) [17] or 128 bit (IPv6) [16] address, most significant octet first. The format of the address (IPv4 or IPv6) is determined by the length. If the attribute value is an IPv4 address, the AVP Length field MUST be 12 (16 if 'V' bit is enabled), otherwise the AVP Length field MUST be set to 24 (28 if the 'V' bit is enabled) for IPv6 addresses. Integer32 32 bit signed value, in network byte order. The AVP Length field MUST be set to 12 (16 if the 'V' bit is enabled). Integer64 64 bit signed value, in network byte order. The AVP Length field MUST be set to 16 (20 if the 'V' bit is enabled). Unsigned32 32 bit unsigned value, in network byte order. The AVP Length field MUST be set to 12 (16 if the 'V' bit is enabled). Unsigned32 values used to transmit time data contains the four most significant octets returned from NTP [18], in network byte order. Unsigned64 64 bit unsigned value, in network byte order. The AVP Length field MUST be set to 16 (20 if the 'V' bit is enabled). Float32 This represents floating point values of single precision as described by [30]. The 32 bit value is transmitted in network byte order. The AVP Length field MUST be set to 12 (16 if the 'V' bit is enabled). Float64 This represents floating point values of double precision as described by [30]. The 64 bit value is transmitted in network byte order. The AVP Length field MUST be set to 16 (20 if the 'V' bit is enabled). Float128 This represents floating point values of quadruple precision as described by [30]. The 128 bit value is transmitted in network byte order. The AVP Length field MUST be set to 24 (28 if the 'V' bit is enabled). Grouped The Data field is specified as a sequence of AVPs. Each of these AVPs follows including their headers and padding. The Calhoun et al. expires December 2001 [Page 29] Internet-Draft June 2001 AVP Length field is set to 8 (12 if the 'V' bit is enabled) plus the total length of all included AVPs, including their headers and padding. 4.4 Grouped AVP Values The Diameter protocol allows AVP values of type 'Grouped.' This implies that the Data field is actually a sequence of AVPs. It is possible to include an AVP with a Grouped type within a Grouped type, that is, to nest them. AVPs within an AVP of type Grouped have the same padding requirements as non-Grouped AVPs, as defined in section 4.0. Every Grouped AVP defined MUST include a corresponding grammar, using ABNF [31] (with modifications), as defined below. avp-def = name "::=" avp name-fmt = ALPHA *(ALPHA / DIGIT / "-") name = name-fmt ; The name has to be the name of an AVP, ; defined in the base or extended Diameter ; specifications. avp = header [ *fixed] [ *required] [ *optional] [ *fixed] header = "" avpcode = 1*DIGIT ; The AVP Code assigned to the Grouped AVP fixed = [qual] "<" avp-spec ">" required = [qual] "{" avp-spec "}" optional = [qual] "[" avp-name "]" ; The avp-name in the 'optional' rule cannot ; evaluate to any AVP Name which is included ; in a fixed or required rule. qual = [min] "*" [max] ; See ABNF conventions, RFC 2234 section 6.6. ; The absence of any qualifiers implies that ; one and only one such AVP MUST be present. ; Calhoun et al. expires December 2001 [Page 30] Internet-Draft June 2001 ; NOTE: "[" and "]" have a different meaning ; than in ABNF (see the optional rule, above). ; These braces cannot be used to express ; optional fixed rules (such as an optional ; ICV at the end.) To do this, the convention ; is '0*1fixed'. min = 1*DIGIT ; The minimum number of times the element may ; be present. max = 1*DIGIT ; The maximum number of times the element may ; be present. avp-spec = name-fmt ; The avp-spec has to be an AVP Name, defined ; in the base or extended Diameter ; specifications. avp-name = avp-spec | "AVP" ; The string "AVP" stands for *any* arbitrary ; AVP Name, which does not conflict with the ; required or fixed position AVPs defined in ; the command code definition. 4.4.1 Example AVP with a Grouped Data type The Example AVP (AVP Code 999999) is of type Grouped and is used to clarify how Grouped AVP values work. The Grouped Data field has the following ABNF grammar: Example-AVP ::= < AVP Header: 999999 > { Origin-Host } 1*{ Session-Id } *[ AVP ] An Example AVP with Grouped Data follows. The Origin-Host AVP is required. In this case: Origin-Host = "example.com". One or more Session-Ids must follow. Here there are two: Session-Id = "grump.example.com:33041;23432;893;0AF3B81" Calhoun et al. expires December 2001 [Page 31] Internet-Draft June 2001 Session-Id = "grump.example.com:33054;23561;2358;0AF3B82" optional AVPs included are Recovery-Policy = 2163bc1d0ad82371f6bc09484133c3f09ad74a0dd5346d54195a7cf0b35 2cabc881839a4fdcfbc1769e2677a4c1fb499284c5f70b48f58503a45c5 c2d6943f82d5930f2b7c1da640f476f0e9c9572a50db8ea6e51e1c2c7bd f8bb43dc995144b8dbe297ac739493946803e1cee3e15d9b765008a1b2a cf4ac777c80041d72c01e691cf751dbf86e85f509f3988e5875dc905119 26841f00f0e29a6d1ddc1a842289d440268681e052b30fb638045f7779c 1d873c784f054f688f5001559ecff64865ef975f3e60d2fd7966b8c7f92 Futuristic-Acct-Record = fe19da5802acd98b07a5b86cb4d5d03f0314ab9ef1ad0b67111ff3b90a0 57fe29620bf3585fd2dd9fcc38ce62f6cc208c6163c008f4258d1bc88b8 17694a74ccad3ec69269461b14b2e7a4c111fb239e33714da207983f58c 41d018d56fe938f3cbf089aac12a912a2f0d1923a9390e5f789cb2e5067 d3427475e49968f841 The data for the optional AVPs is represented in hex since the format of these AVPs is neither known at the time of definition of the Example-AVP group, nor (likely) at the time when the example instance of this AVP is interpreted - except by Diameter implementations which support the same set of AVPs. The encoding example illustrates how padding is used, how length fields are calculated and how AVPs do not have to begin on 8 byte boundaries. Also note that AVPs may be present in the Grouped AVP value which the receiver cannot interpret (here, the Recover-Policy and Futuristic-Acct-Record AVPs). This AVP would be encoded as follows: Calhoun et al. expires December 2001 [Page 32] Internet-Draft June 2001 0 1 2 3 4 5 6 7 +-------+-------+-------+-------+-------+-------+-------+-------+ 0 | Example AVP Header (AVP Code = 999999), Length = 468 | +-------+-------+-------+-------+-------+-------+-------+-------+ 8 | Origin-Host AVP Header (AVP Code = 264), Length = 19 | +-------+-------+-------+-------+-------+-------+-------+-------+ 16 | 'e' | 'x' | 'a' | 'm' | 'p' | 'l' | 'e' | '.' | +-------+-------+-------+-------+-------+-------+-------+-------+ 24 | 'c' | 'o' | 'm' |Padding| Session-Id AVP Header | +-------+-------+-------+-------+-------+-------+-------+-------+ 32 | (AVP Code = 263), Length = 50 | 'g' | 'r' | 'u' | 'm' | +-------+-------+-------+-------+-------+-------+-------+-------+ . . . +-------+-------+-------+-------+-------+-------+-------+-------+ 64 | 'A' | 'F' | '3' | 'B' | '8' | '1' |Padding|Padding| +-------+-------+-------+-------+-------+-------+-------+-------+ 68 | Session-Id AVP Header (AVP Code = 263), Length = 51 | +-------+-------+-------+-------+-------+-------+-------+-------+ 72 | 'g' | 'r' | 'u' | 'm' | 'p' | '.' | 'e' | 'x' | +-------+-------+-------+-------+-------+-------+-------+-------+ . . . +-------+-------+-------+-------+-------+-------+-------+-------+ 104 | '0' | 'A' | 'F' | '3' | 'B' | '8' | '2' |Padding| +-------+-------+-------+-------+-------+-------+-------+-------+ 112 | Recovery-Policy Header (AVP Code = 8341), Length = 223 | +-------+-------+-------+-------+-------+-------+-------+-------+ 120 | 0x21 | 0x63 | 0xbc | 0x1d | 0x0a | 0xd8 | 0x23 | 0x71 | +-------+-------+-------+-------+-------+-------+-------+-------+ . . . +-------+-------+-------+-------+-------+-------+-------+-------+ 320 | 0x2f | 0xd7 | 0x96 | 0x6b | 0x8c | 0x7f | 0x92 |Padding| +-------+-------+-------+-------+-------+-------+-------+-------+ 328 | Futuristic-Acct-Record Header (AVP Code = 15930), Length = 137| +-------+-------+-------+-------+-------+-------+-------+-------+ 336 | 0xfe | 0x19 | 0xda | 0x58 | 0x02 | 0xac | 0xd9 | 0x8b | +-------+-------+-------+-------+-------+-------+-------+-------+ . . . +-------+-------+-------+-------+-------+-------+-------+-------+ 464 | 0x41 |Padding|Padding|Padding| +-------+-------+-------+-------+ 4.5 Diameter Base Protocol AVPs The following table describes the Diameter AVPs defined in the base protocol, their AVP Code values, types, possible flag values and whether the AVP MAY be encrypted. Calhoun et al. expires December 2001 [Page 33] Internet-Draft June 2001 +---------------------+ | AVP Flag rules | |----+-----+----+-----|----+ AVP Section | | |SHLD| MUST|MAY | Attribute Name Code Defined Data Type |MUST| MAY | NOT| NOT|Encr| -----------------------------------------|----+-----+----+-----|----| Accounting- 482 13.2 Unsigned32 | M | P | | V | Y | Interim-Interval | | | | | | Accounting- 50 13.5 OctetString| M | P | | V | Y | Multi-Session-Id | | | | | | Accounting- 485 13.3 Unsigned32 | M | P | | V | Y | Record-Number | | | | | | Accounting- 480 13.1 Unsigned32 | M | P | | V | Y | Record-Type | | | | | | Accounting- 44 13.4 OctetString| M | P | | V | Y | Session-Id | | | | | | Acct- 259 6.10 Integer32 | M | | | V | N | Application-Id | | | | | | Auth- 258 6.6 Integer32 | M | | | V | N | Application-Id | | | | | | Authorization- 291 10.4 Unsigned32 | M | | | V | N | Lifetime | | | | | | Destination-Host 293 5.6 OctetString| M | | | V | N | Destination- 283 5.7 OctetString| M | | | V | N | Realm | | | | | | Error-Message 281 9.3 OctetString| | | | V | N | Error-Reporting- 294 9.4 OctetString| | | | V | N | Host | | | | | | Failed-AVP 279 9.5 OctetString| M | | | V | Y | Firmware- 267 6.5 Unsigned32 | | | | V,M | N | Revision | | | | | | Host-IP-Address 257 6.7 Address | M | | | V | N | Origin-Host 264 5.4 OctetString| M | | | V | N | Origin-Realm 296 5.5 OctetString| M | | | V | N | Product-Name 269 6.9 OctetString| | | | | N | Proxy-Host 280 5.8.3 Address | M | | | V | N | Proxy-Info 284 5.8.2 Grouped | M | | | V | N | Proxy-State 33 5.8.4 OctetString| M | | | V | N | Redirect-Host 292 5.9 OctetString| M | | | V | Y | Result-Code 268 9.1 Unsigned32 | M | | | V | N | Route-Record 282 5.8.1 OctetString| M | | | V | N | Session-Id 263 10.3 OctetString| M | | | V | Y | Session-Timeout 27 10.5 Unsigned32 | M | | | V | N | Origin-State-Id 278 10.11 Unsigned32 | M | | | V | N | Supported- 265 6.8 Unsigned32 | M | | | V | N | Vendor-Id | | | | | | -----------------------------------------|----+-----+----+-----|----| Calhoun et al. expires December 2001 [Page 34] Internet-Draft June 2001 +---------------------+ | AVP Flag rules | |----+-----+----+-----|----+ AVP Section | | |SHLD| MUST|MAY | Attribute Name Code Defined Data Type |MUST| MAY | NOT| NOT|Encr| -----------------------------------------|----+-----+----+-----|----| Termination- 295 10.9 Unsigned32 | M | | | V | N | Cause | | | | | | User-Name 1 10.6 OctetString| M | | | V | Y | Vendor-Id 266 6.4 Unsigned32 | M | | | V,M | N | Vendor-Specific- 260 6.11 Grouped | M | | | V,M | N | Application-Id -----------------------------------------|----+-----+----+-----|----| 5.0 Diameter message processing All Diameter messages MUST include the Origin-Host and Origin-Realm AVPs, which are used to identify the source of the message. The Destination-Host AVP MAY be present in requests, and MUST be present in answers. The Destination-Host AVP is used when the destination of the message is fixed, which includes: - Authentication requests that span multiple round trips - A Diameter message that uses a security mechanism that makes use of a pre-established session key shared between the source and the final destination of the message. - Server initiated messages that MUST be received by a specific Diameter client (e.g. access device), such as the Abort- Session-Request message, which is used to request that a particular user's session be terminated. The Destination-Realm AVP MUST be present if the message is routable. A message that MUST NOT be relayed, proxied or redirected MUST NOT include the Destination-Realm in its ABNF. The value of the Destination-Realm AVP MAY be extracted from the User-Name AVP, or other application-specific methods. When a message is received, the message is processed in the following order: 1. If the message is destined for the local host, the procedures listed in section 5.1 are followed. 2. If the message is intended for a Diameter peer with whom the local host is able to directly communicate with, the procedures listed in section 5.2 are followed. This is known as Message Forwarding. 3. The procedures listed in section 5.3 are followed, which is known as Message Routing. Calhoun et al. expires December 2001 [Page 35] Internet-Draft June 2001 4. If none of the above are successful, an answer is returned with the Result-Code set to DIAMETER_UNABLE_TO_DELIVER. Note the processing rules contained in this section are intended to be used as general guidelines to Diameter developers. Certain implementations MAY use different methods than the ones described here, and still be in compliance with the protocol specification. 5.1 Processing Local Messages A request is known to be for local comsumption when one of the following conditions occur: - The Destination-Host AVP contains the local host's identity, - The Destination-Host AVP is not present, the Destination-Realm AVP contains a realm the server is configured to process locally, and the Diameter application is locally supported, or - The Destination-Realm AVP is not present. When a request is locally processed, the following procedures MUST be applied, in addition to any additional procedures that MAY be discussed in the Diameter application defining the command: - The same Hop-by-Hop identifier in the request is used in the answer. - The local host's identity is encoded in the Origin-Host and Origin-Host AVPs. - The value of the Origin-Host AVP in the request is included in the answer's Destination-Host AVP. - The Result-Code AVP is added with its value indicating success or failure. - If the Session-Id is present in the request, it MUST be included in the answer. - Any Route-Record or Proxy-Info AVPs in the request MUST be added to the answer message, in the same order they were present in the request. When the local message is an answer, no additional procedures beyond those listed in the specific Diameter application are to be followed. 5.2 Message Forwarding Message forwarding is done using the Diameter Peer Table. The Diameter peer table contains all of the peers that the local node is able to directly communicate with. When a request is received, and the host encoded in the Destination- Calhoun et al. expires December 2001 [Page 36] Internet-Draft June 2001 Host AVP is one that is present in the peer table, the message SHOULD be forwarded to the peer. If the message received is an answer, the host in the Destination- Host AVP is in the peer table, and there are no Route-Record AVPs in the message, the message MUST be forwarded to the peer. 5.2.1 Peer Table The Diameter Peer Table is used in message forwarding, and referenced by the Domain Routing Table. A Peer Table entry contains the following fields: - Peer name. The Fully Qualified Domain Name of the peer. This MAY be resolved locally, or known via the CER or CEA message. - Port Number. The port number the peer may be contacted on. - Protocol. Specifies whether TCP or SCTP is the protocol to use to communicate with the peer. - TLS Enabled. Specifies whether TLS is to be used when communicating with the peer. 5.3 Message Routing Diameter request message routing is done via realms. A Diameter message that is proxyable MUST include the target realm in the Destination-Realm AVP. The realm MAY be retrieved from the User-Name AVP, which is in the form of a Network Access Identifier (NAI). The realm portion of the NAI is inserted in the Destination-Realm AVP. Diameter agents have a list of locally supported realms, and MAY have a list of externally supported realms. When a request is received that includes a realm that is not locally supported, the message is routed to the peer configured in the Domain Routing Table table (see section 5.3.1). 5.3.1 Realm-Based Routing Table All Realm-Based routing lookups are performed against what is commonly known as the Domain Routing Table (see section 16.0). A Domain Routing Table Entry contains the following fields: - Domain Name. The Domain Name is analogous to the realm portion of the NAI. This is the field that is typically used as a primary key in the routing table lookups. Note that some implementations perform their lookups based on longest-match- from-the-right on the realm rather than requiring an exact match. Calhoun et al. expires December 2001 [Page 37] Internet-Draft June 2001 - Application Identifier. It is possible for a routing entry to have a different destination based on the Acct-Application-Id (for accounting messages) or Auth-Application-Id (for non- accounting messages) of the message. This field is typically used as a secondary key field in routing table lookups. - Local Action. The Local Action field is used to identify how a message should be treated. The following actions are supported: 1. LOCAL - Diameter messages that resolve to a routing entry with the Local Action set to Local can be satisfied locally, and do not need to be routed to another server. 2. RELAY - All Diameter messages that fall within this category MUST be routed to a next hop server, without modifying any non-routing AVPs. See sections 5.3.3 and 5.3.4 for relaying guidelines 3. PROXY - All Diameter messages that fall within this category MUST be routed to a next hop server. The local server MAY apply its local policies to the message by including new AVPs to the message prior to routing. See sections 5.3.3 and 5.3.4 for relaying guidelines. 4. REDIRECT - Diameter messages that fall within this category MUST have the identity of the home Diameter server(s) appended, and returned to the sender of the message. See section 5.3.2 for redirect guidelines. - Server Identifier - One or more servers the message is to be routed to. These servers MUST also be present in the Peer table. When the Local Action is set to RELAY or PROXY, this field contains the identity of the server(s) the message must be routed to. When the Local Action field is set to REDIRECT, this field contains the identity of one or more servers the message should be redirected to. It is important to note that Diameter agents MUST support at least one of the LOCAL, RELAY, PROXY or REDIRECT modes of operation. Agents do not need to support all modes of operation in order to conform with the protocol specification, but MUST follow the protocol compliance guidelines in section 2.0. Relay agents MUST NOT reorder AVPs, and proxies SHOULD NOT reorder AVPs. When a request is routed, the target server MUST have advertised the Application Identifier (see section 6.1) for the given message, or have advertised itself as a relay or proxy agent. 5.3.2 Redirecting requests When a redirector agent receives a request whose routing entry is set to REDIRECT, it MUST answer the request with Message-Reject-Answer, while maintaining the Hop-by-Hop Identifier in the header, and Calhoun et al. expires December 2001 [Page 38] Internet-Draft June 2001 include the Result-Code AVP to DIAMETER_REDIRECT_INDICATION. Each of the servers associated with the routing entry are added in separate Redirect-Host AVP. +------------------+ | Diameter | | Redirector Agent | +------------------+ ^ | 1. Request | | 2. MRA + joe@xyz.com | | Result-Code = DIAMETER_REDIRECT_INDICATION + | | Redirect-Host AVP(s) | v +---------+ 3. Request +----------+ | abc.net |------------->| xyz.net | | Relay | | Diameter | | Agent |<-------------| Server | +---------+ 4. Answer +----------+ Figure 7: Diameter Redirect Server Redirector agents MAY also include the certificate of the servers in the Redirect-Host AVP(s). These certificates are encapsulated in a CMS-Cert AVP [11]. The receiver of the MRA message with the Result-Code AVP set to DIAMETER_REDIRECT_INDICATION uses the hop-by-hop field in the Diameter header to identify the request in the pending message queue (see Section 7.3) that is to be redirected. If no transport connection exists with the new agent, one is created, and the request is sent directly to it. 5.3.3 Relaying and Proxying Requests A relay or proxy agent MUST check for forwarding loops before forwarding requests. A loop is detected if the server finds its own address in a Route-Record AVP. When such an event occurs, the agent MUST answer with the Result-Code AVP set to DIAMETER_LOOP_DETECTED. A relay or proxy agent MUST append a Route-Record AVP that includes its identity to all requests forwarded. The last Route-Record AVP in all requests received MUST be validated, by ensuring that the host encoded in the AVP is the same as the peer the message was received from. The Hop-by-Hop identifier in the request is saved, and replaced with a locally unique value. Calhoun et al. expires December 2001 [Page 39] Internet-Draft June 2001 Relay and Proxy agents MAY include the Proxy-Info AVP in requests if it requires access any local state information when the corresponding response is received. Alternatively, it MAY simply use local storage to store state information. The message is then forwarded to the next hop, as identified in the Domain Routing Table. Figure 6 provides an example of message routing using the procedures listed in these sections. (Origin-Host=nas.mno.net) (Origin-Host=nas.mno.net) (Origin-Realm=mno.net) (Origin-Realm=mno.net) (Destination-Realm=abc.com) (Destination-Realm=abc.com) (Route-Record=drl.mno.net) +------+ ------> +------+ ------> +------+ | | (Request) | | (Request) | | | NAS +-------------------+ DRL +-------------------+ HMS | | | | | | | +------+ <------ +------+ <------ +------+ mno.net (Answer) mno.net (Answer) abc.com (Origin-Host=hms.abc.com) (Origin-Host=hms.abc.com) (Origin-Realm=abc.com) (Origin-Realm=abc.com) (Destination-Host=nas.mno.net) (Destination-Host=nas.mno.net) (Route-Record=drl.mno.net) Figure 6: Routing of Diameter messages 5.3.4 Relaying and Proxying Answers A relay or proxy agent MUST only process Answers whose last Route- Record AVP matches one of its identities. Any answers that do not conform to this rule MUST be dropped. The last Route-Record AVP MUST be removed from the message before it is forwarded to the next hop, which is identified by the second to last Route-Record AVP. If the last Proxy-Info AVP in the message is targeted to the local Diameter server, the AVP MUST be removed. If a relay or proxy agent receives an answer with a Result-Code AVP indicating a failure, it MUST NOT modify the contents of the AVP. Any additional local errors detected SHOULD be logged, but not reflected in the Result-Code AVP. If the agent receives an answer message with a Result-Code AVP indicating success, and it wishes to modify the AVP to indicate an error, it MUST issue an STR on behalf of the access device. Prior to forwarding the answer, the agent MUST restore the original Calhoun et al. expires December 2001 [Page 40] Internet-Draft June 2001 value of the Diameter header's Hop-by-Hop Identifier field. 5.3.5 Hiding Network Topology A Relay or Proxy agent routing messages outside of their administrative domain MAY need to hide the internal Diameter topology. This is done by removing all Route-Record AVPs in a request, and later adding them back into the corresponding answer, in the same order. Such agents MUST take care to not assume that the absence of any Route-Record AVPs implies the message is for local comsumption. 5.4 Origin-Host AVP The Origin-Host AVP (AVP Code 264) is of type OctetString, encoded in the UTF-8 [24] format, according to the Diameter identity rules defined in section 2.7, and MUST be present in all Diameter messages. This AVP identifies the endpoint which originated the Diameter message, i.e. the access device, home server, or broker. Relay agents MUST NOT modify this AVP. Note that the Origin-Host AVP may resolve to more than one address as the Diameter peer may support more than one address. This AVP SHOULD be placed as close to the Diameter header as possible. 5.5 Origin-Realm AVP The Origin-Realm AVP (AVP Code 296) is of type OctetString, encoded in the UTF-8 [24] format. This AVP contains the Realm of the originator of any Diameter message and MUST be present in all messages This AVP SHOULD be placed as close to the Diameter header as possible. 5.6 Destination-Host AVP The Destination-Host AVP (AVP Code 293) is of type OctetString, encoded in the UTF-8 [24] format, according to the Diameter identity rules defined in section 2.7. This AVP MUST be present in all unsolicited agent initiated messages, MAY be present in request messages, and MUST be present in Answer messages. The value of the Calhoun et al. expires December 2001 [Page 41] Internet-Draft June 2001 Destination-Host AVP is set to the value of the Origin-Host AVP found in a message from the intended target host. This AVP SHOULD be placed as close to the Diameter header as possible. 5.7 Destination-Realm AVP The Destination-Realm AVP (AVP Code 283) is of type OctetString, encoded in the UTF-8 [24] format, and contains the realm the message is to be routed to. The Destination-Realm AVP MUST NOT be present in Answer messages. Diameter Clients insert the realm portion of the User-Name AVP. Diameter servers initiating a request message use the value of the Origin-Realm AVP from a previous message received from the intended target host (unless it is known a priori). When present, the Destination-Realm AVP is used to perform message routing decisions. Request messages whose ABNF does not list the Destination-Realm AVP as a mandatory AVP are inherently non-routable messages. This AVP SHOULD be placed as close to the Diameter header as possible. 5.8 Routing AVPs The AVPs defined in this section are Diameter AVPs used for routing purposes. These AVPs change as Diameter messages are processed by agents, and therefore MUST NOT be protected using the Diameter CMS Security application [11]. 5.8.1 Route-Record AVP The Route-Record AVP (AVP Code 282) is of type OctetString, encoded in the UTF-8 [24] format, according to the Diameter identity rules defined in section 2.7. The identity added in this AVP MUST be the same as the identity sent in the Origin-Host of the Capabilities- Exchange-Request message. 5.8.2 Proxy-Info AVP The Proxy-Info AVP (AVP Code = 284) is of type Grouped. The Grouped Data field has the following ABNF grammar: Calhoun et al. expires December 2001 [Page 42] Internet-Draft June 2001 Proxy-Info ::= < AVP Header: 284 > { Proxy-Host } { Proxy-State } * [ AVP ] 5.8.3 Proxy-Host AVP The Proxy-Host AVP (AVP Code = 280) is of type OctetString, encoded in the UTF-8 [24] format, according to the Diameter identity rules defined in section 2.7. This AVP contains the identity of the host that added the Proxy-Info AVP. 5.8.4 Proxy-State AVP The Proxy-State AVP (AVP Code = 33) is of type OctetString, and contains state local information, and MUST be treated as opaque data. 5.9 Redirect-Host AVP The Redirect-Host AVP (AVP Code 292) is of type OctetString, encoded in the UTF-8 [24] format, according to the Diameter identity rules defined in section 2.7. This AVP MUST be present in Message-Reject- Answer messages that include the Result-Code AVP set to DIAMETER_REDIRECT_INDICATION. Upon receiving the above, the receiving Diameter node SHOULD forward the request directly to the host identified in this AVP. 6.0 Capabilities Exchange When two Diameter peers establish a transport connection, they MUST exchange the Device Reboot messages, as specified in the peer state machine (see section 8.0). This message has two purposes. First it allows a peer's identity to be discovered, and allows for capabilities exchange, such as the supported protocol version number, the locally supported Diameter applications, etc. The receiver only issues commands to its peers that have advertised support for the Diameter application that defines the command. A Diameter node MUST cache the supported applications in order to ensure that unrecognized commands and/or AVPs are not unnecessarily sent to a peer. A receiver of a Capabilities-Exchange-Req message which does not have Calhoun et al. expires December 2001 [Page 43] Internet-Draft June 2001 any applications in common with the sender MUST return a Capabilities-Exchange-Answer with the Result-Code AVP set to DIAMETER_NO_COMMON_APPLICATION, and SHOULD disconnect the transport layer connection. The Capabilities-Exchange-Request and Capabilities-Exchange-Answer messages MUST NOT be proxied, or redirected. Since the CER/CEA messages cannot be proxied, it is still possible that an upstream proxy receives a message for which it has no available peers to handle the application that corresponds to the Command-Code. In such instances, the Message-Reject-Answer message is used (see Section 9.2.1) to inform the downstream to take action (e.g. re-routing request to an alternate peer). With the exception of the Capabilities-Exchange-Request message, a message of type Request that includes the Auth-Application-Id or Acct-Application-Id AVPs, or a message with an application-specific command code, MAY only be forwarded to a host that has explicitly advertised support for the application (or has advertised the Relay Application Identifier). 6.1 Application Identifiers Each Diameter application MUST have an IANA assigned Application Identifier (see section 15.3). The base protocol does not require an application Identifier since its support is mandatory. Application Identifiers are communicated via two separate AVPs; Auth-Application-Id and Acct-Application-Id. The Auth-Application-Id AVP is used to communicate support for the authentication and authorization portion of an application. The Acct-Application-Id AVP, on the other hand, communicates support for the accounting portion of an application. This separation of AVPs allows a server to communicate that it is willing to accept only accounting messages for a given application. The following Application Identifier values are defined: NASREQ 1 [7] End-to-End Security 2 [11] Mobile-IP 4 [10] Relay 0xffffffff Relay and redirect agents MUST advertise the Proxy application identifier, while all other Diameter nodes MUST advertise locally Calhoun et al. expires December 2001 [Page 44] Internet-Draft June 2001 supported applications. The receiver of a Device Reboot message advertising Relay service MUST assume that the sender supports all current and future applications. Diameter relay and proxy agents are responsible for finding a downstream server that supports the application of a particular message. If none can be found, a MRA message is returned with the Result-Code AVP set to DIAMETER_UNABLE_TO_DELIVER. 6.2 Capabilities-Exchange-Request The Capabilities-Exchange-Request (CER), indicated by the Command- Code set to 257 and the Command Flags' 'R' bit set, is sent to inform a peer that a reboot has occurred. When Diameter is run over SCTP [26], which allows for connections to span multiple interfaces, hence multiple IP addresses, the Capabilities-Exchange-Request message MUST contain one Host-IP- Address AVP for each potential IP address that MAY be locally used when transmitting Diameter messages. Message Format ::= < Diameter Header: 257, REQUEST > { Origin-Host } { Origin-Realm } 1* { Host-IP-Address } { Vendor-Id } { Product-Name } [ Origin-State-Id ] * [ Supported-Vendor-Id ] * [ Auth-Application-Id ] * [ Acct-Application-Id ] [ Destination-Host ] [ Firmware-Revision ] * [ AVP ] 6.3 Capabilities-Exchange-Answer The Capabilities-Exchange-Request (CEA), indicated by the Command- Code set to 257 and the Command Flags' 'R' bit cleared, is sent in response to a CER message. When Diameter is run over SCTP [26], which allows for connections to span multiple interfaces, hence multiple IP addresses, the Capabilities-Exchange-Answer message MUST contain one Host-IP-Address Calhoun et al. expires December 2001 [Page 45] Internet-Draft June 2001 AVP for each potential IP address that MAY be locally used when transmitting Diameter messages. Message Format ::= < Diameter Header: 257 > { Result-Code AVP } { Origin-Host } { Origin-Realm } 1* { Host-IP-Address } { Vendor-Id } { Product-Name } [ Origin-State-Id ] * [ Supported-Vendor-Id ] * [ Auth-Application-Id ] * [ Acct-Application-Id ] [ Destination-Host ] [ Firmware-Revision ] * [ AVP ] 6.4 Vendor-Id AVP The Vendor-Id AVP (AVP Code 266) is of type Unsigned32 and contains the IANA "SMI Network Management Private Enterprise Codes" [2] value assigned to the vendor of the Diameter device. In combination with the Supported-Vendor-Id AVP (section 6.8), this MAY be used in order to know which vendor specific attributes may be sent to the peer. It is also envisioned that the combination of the Vendor-Id, Product-Name (section 6.9) and the Firmware-Revision (section 6.5) AVPs MAY provide very useful debugging information. A Vendor-Id value of zero in the CER or CEA messages is reserved and indicates that the Diameter peer is in the experimental or concept stage and that an IANA Private Enterprise Number has yet to be obtained by the implementor. 6.5 Firmware-Revision AVP The Firmware-Revision AVP (AVP Code 267) is of type Unsigned32 and is used to inform a Diameter peer of the firmware revision of the issuing device. For devices that do not have a firmware revision (general purpose computers running Diameter software modules, for instance), the revision of the Diameter software module may be reported instead. Calhoun et al. expires December 2001 [Page 46] Internet-Draft June 2001 6.6 Auth-Application-Id AVP The Auth-Application-Id AVP (AVP Code 258) is of type Unsigned32 and is used in order to advertise support of the Authentication and Authorization portion of an application (see Section 6.1). The Auth- Application-Id MUST also be present in all Authentication and/or Authorization messages that are defined in a separate Diameter specification and have an Application ID assigned. This AVP SHOULD be placed as close to the Diameter header as possible. 6.7 Host-IP-Address AVP The Host-IP-Address AVP (AVP Code 257) is of type Address and is used to inform a Diameter peer of the sender's IP address. All source addresses that a Diameter node expects to use with SCTP [26] MUST be advertised in the CER and CEA messages by including a Host-IP-Address AVP for each address. This AVP MUST ONLY be used in the CER and CEA messages. 6.8 Supported-Vendor-Id AVP The Supported-Vendor-Id AVP (AVP Code 265) is of type Unsigned32 and contains the IANA "SMI Network Management Private Enterprise Codes" [2] value assigned to a vendor other than the device vendor. This is used in the CER and CEA messages in order to inform the peer that the sender supports a subset of the vendor-specific commands and/or AVPs defined by the vendor identified in this AVP. 6.9 Product-Name AVP The Product-Name AVP (AVP Code 269) is of type OctetString, encoded in the UTF-8 [24] format, and contains the vendor assigned name for the product. The Product-Name AVP SHOULD remain constant across firmware revisions for the same product. 6.10 Acct-Application-Id AVP The Acct-application-Id AVP (AVP Code 259) is of type Unsigned32 and is used in order to advertise support of the Accounting portion of an application (see Section 6.1). The Acct-Application-Id MUST also be present in all Accounting messages that are defined in a separate Diameter specification and have an Application ID assigned. Calhoun et al. expires December 2001 [Page 47] Internet-Draft June 2001 This AVP SHOULD be placed as close to the Diameter header as possible. 6.11 Vendor-Specific-Application-Id AVP The Vendor-Specific-Application-Id AVP (AVP Code 260) is of type Grouped and is used to advertise support of a vendor-specific Diameter Application. Either the Auth-Application-Id or the Acct- Application-Id AVP MAY be present. Both AVPs MAY be present if they both contain the same value. This AVP MUST also be present in all vendor-specific commands defined in the vendor-specific application. This AVP SHOULD be placed as close to the Diameter header as possible. AVP Format ::= < AVP Header: 260 > 1* [ Vendor-Id ] 0*1{ Auth-Application-Id } 0*1{ Acct-Application-Id } 7.0 Transport Failure Detection Given the nature of the Diameter protocol, it is recommended that transport failures be detected as soon as possible. Detecting such failures will minimize the occurrence of messages sent to unavailable servers, resulting in unnecessary delays, and will provide better failover performance. The Device-Watchdog-Request and Device- Watchdog-Answer messages, defined in this section, are used to pro- actively detect transport failures. The watchdog behavior is controlled by the Tw timer, which ranges between 30 and 60 seconds. In order to avoid synchronization behaviors that can occur with fixed timers among distributed systems, each time the watchdog interval is calculated with a jitter by using the Tw value (which defaults to 30 seconds) and randomly adding or subtracting a random value drawn between 0.5 and 2 seconds. Alternative calculations to create jitter MAY be used. These MUST be pseudo-random and not cyclic. When a response is received, Tw is reset. Receiving a watchdog from a peer constitutes activity, and Tw should be reset. On sending a message, if the queue is empty, then Tw is reset. If the watchdog Calhoun et al. expires December 2001 [Page 48] Internet-Draft June 2001 timer expires and the queue is empty, then a watchdog packet is sent. 7.1 Device-Watchdog-Request The Device-Watchdog-Request (DWR), indicated by the Command-Code set to 280 and the Command Flags' 'R' bit set, is sent to a peer when no traffic has been exchanged between two peers as defined in Section 7.0, and no requests are pending with the peer. Message Format ::= < Diameter Header: 280, REQUEST > { Origin-Host } { Origin-Realm } { Destination-Host } 7.2 Device-Watchdog-Answer The Device-Watchdog-Answer (DWA), indicated by the Command-Code set to 280 and the Command Flags' 'R' bit cleared, is sent as a response to the Device-Watchdog-Request message. A receiver of the DWA SHOULD perform RTT calculation in the event that the transport RTO information is not available. Message Format ::= < Diameter Header: 280 > { Result-Code } { Origin-Host } { Origin-Realm } { Destination-Host } 7.3 Failover/Failback Procedures In the event that a transport failure is detected with a peer, it is necessary for all pending request messages to be forwarded to an alternate agent, if possible. This is commonly referred to as failover. In order for a Diameter node to perform failover procedures, it is necessary for the node to maintain a pending message queue for a given peer. When an answer message is received, the corresponding request is removed from the queue. The Hop-by-Hop Identifier field MAY be used to match the answer with the queued request. Calhoun et al. expires December 2001 [Page 49] Internet-Draft June 2001 When a transport failure is detected, all messages in the queue are sent to an alternate agent, if possible. An example of a case where it is not possible for forward the message to an alternate server is when the message has a fixed destination, and the unavailable peer is the message's final destination (see Destination-Host AVP). Such an error requires that the agent return an MRA with the Result-Code AVP set to DIAMETER_UNABLE_TO_DELIVER. It is important to note that multiple identical request or answer MAY be received as a result of a failover. The End-to-End Identifier field in the Diameter header along with the Origin-Host AVP MUST be used to identify duplicate messages. As described in section 2.1, a connection request should be periodically attempted with the failed peer in order to re-establish the transport connection. Once a connection has been successfully established, messages can once again be forwarded to the peer. This is commonly referred to as failback. 8.0 Peer State Machine This section contains a finite state machine, that MUST be observed by all Diameter implementations. Each Diameter node MUST follow the state machine described below when communicating with each peer. Multiple actions are separated by commas, and may continue on succeeding lines as space requires. Similarly, state and next state may also span multiple lines as space requires. There may be at most one transport connection between any two peers over which Diameter messages may be passed. This state machine is intended to handle both the simple case, in which one peer initiates a connection to the other, and the complex case, in which each peer simultaneously initiates a connection to the other. In the complex case, an election occurs to determine which transport connection will survive. I- is used to represent the initiator (connecting) connection, while the R- is used to represent the responder (listening) connection. The lack of a prefix indicates that the event or action is the same regardless of the connection on which the event occurred. The stable states that a state machine may be in are Closed, I-Open and R-Open; all other states are intermediate. Note that I-Open and R-Open are equivalent except for whether the initiator or responder transport connection is used for communication. A CER message is always sent on the initiating connection immediately Calhoun et al. expires December 2001 [Page 50] Internet-Draft June 2001 after the connection request is successfully completed. The non- elected connection will close down. All subsequent messages are sent on the elected connection. The state machine constrains only the behavior of a Diameter implementation as seen by Diameter peers through events on the wire. Any implementation that produces equivalent results is considered compliant. state event action next state ----------------------------------------------------------------- Closed Start I-Snd-Conn-Req Wait-Conn-Ack R-Rcv-Conn-Req R-Snd-Conn-Ack Wait-R-CER Wait-Conn-Ack I-Rcv-Conn-Ack I-Snd-CER Wait-I-CEA I-Rcv-Conn-Nack Cleanup Closed R-Rcv-Conn-Req R-Snd-Conn-Ack Wait-Conn-Ack/ Wait-R-CER Timeout Error Closed Wait-I-CEA I-Rcv-CEA Process-CEA I-Open R-Rcv-Conn-Req R-Snd-Conn-Ack Wait-R-CER/ Elect I-Peer-Disc I-Disc Closed Timeout Error Closed Wait-Conn-Ack/ I-Rcv-Conn-Ack I-Snd-CER Wait-R-CER/ Wait-R-CER Elect I-Rcv-Conn-Nack Cleanup Wait-R-CER R-Rcv-CER Process-CER Wait-Conn-Ack/ Elect Timeout Error Closed Wait-R-CER/ R-Rcv-CER Process-CER, Wait-Returns Elect Elect I-Peer-Disc I-Disc Wait-R-CER Timeout Error Closed Wait-Conn-Ack/ I-Rcv-Conn-Ack I-Snd-CER,Elect Wait-Returns Elect I-Rcv-Conn-Nack R-Snd-CEA R-Open R-Peer-Disc R-Disc Wait-Conn-Ack-2 Timeout Error Closed Wait-Returns Win-Election I-Disc,R-Snd-CEA R-Open I-Peer-Disc I-Disc,R-Snd-CEA R-Open I-Rcv-CEA R-Disc I-Open R-Peer-Disc R-Disc Wait-I-CEA-2 Timeout Error Closed Calhoun et al. expires December 2001 [Page 51] Internet-Draft June 2001 Wait-Conn-Ack-2 I-Rcv-Conn-Ack I-Snd-CER Wait-I-CEA-2 I-Rcv-Conn-Nack Cleanup Closed R-Rcv-Conn-Req R-Disc Wait-Conn-Ack-2 Timeout Error Closed Wait-I-CEA-2 I-Rcv-CEA Process-CEA I-Open I-Peer-Disc I-Disc Closed R-Rcv-Conn-Req R-Disc Wait-I-CEA-2 Timeout Error Closed Wait-R-CER R-Rcv-CER Process_CER, R-Open R-Snd-CEA Timeout Error Closed R-Open Send-Message R-Snd-Non-DRI R-Open R-Rcv-Non-DRI Process R-Open R-WatchDog-Timer R-Snd-DWR R-Open R-Rcv-DWA Process-DWA R-Open Stop R-Snd-Disc Closed R-Peer-Disc R-Disc Closed R-Rcv-CER Error Closed I-Open Send-Message I-Snd-Non-DRI I-Open I-Rcv-Non-DRI Process I-Open I-WatchDog-Timer I-Snd-DWR I-Open I-Rcv-DWA Process-DWA I-Open Stop I-Disc Closed I-Peer-Disc I-Disc Closed I-Rcv-DRI Error Closed R-Rcv-Conn-Req R-Disc I-Open 8.1 States Following is a more detailed description of each automaton state. Closed A peer is initially in the closed state, and no transport connection exists with the peer. Wait-Conn-Ack A transport connection has been initiated with the peer, and an acknowledgement is pending. Wait-I-CEA The local Diameter node is waiting for the peer to issue a DRI. Wait-Conn-Ack/Wait-R-CER A transport connection indication from the peer was received, while a transport connection has already Calhoun et al. expires December 2001 [Page 52] Internet-Draft June 2001 been locally initiated. Wait-R-CER/Elect Two transport connections have been established with the peer, and a DRI is pending on the responder connection. Wait-Conn-Ack/Elect A transport connection exists on the responder connection, while an acknowledgment has yet to be received on the initiator connection. Wait-Returns Multiple transport connections caused an election to occur. Wait-Conn-Ack-2 While an acknowledgement to a locally initiated transport connection hasn't been received, an election has failed and the initiator connection will be used between the peers. Wait-I-CEA-2 Following an election, the initiator connection won, and a DRI has yet to be received by the peer. Wait-R-CER A transport connection indication has been received from the peer, and a DRI has yet to be received by the peer. R-Open The responder connection will be used to communicate with the peer. I-Open The initiator connection will be used to communicate with the peer. 8.2 Events Transitions and actions in the automaton are caused by events. In this section we will ignore the -I and -R prefix, since the actual event would be identical, but would occur on one of two possible connections. Start The Diameter application has signaled that a connection should be initiated with the peer. Rcv-Conn-Req A transport connection indication from the peer has been received. Calhoun et al. expires December 2001 [Page 53] Internet-Draft June 2001 Rcv-Conn-Ack A positive acknowledgement was received to a locally initiated transport connection. Rcv-Conn-Nack A negative acknowledgement was received to a locally initiated transport connection. Timeout An application-defined timer has expired while waiting for some event. Rcv-CER A CER message from the peer was received. Rcv-CEA A CEA message from the peer was received. Peer-Disc A disconnection indication from the peer was received. Win-Election An election was held, and the local node was the winner. Send-Message A Non-DRI message is to be sent. Rcv-Non-DRI A Non-DRI message was received. WatchDog-Timer The Watchdog timer expired, indicating that a DWR message is to be sent to the peer. Rcv-DWA A DWA message was received. Stop The Diameter application has signaled that a connection should be terminated (e.g., on system shutdown). 8.3 Actions Actions in the automaton are caused by events and typically indicate the transmission of packets and/or an action to be taken on the connection. In this section we will ignore the -I and -R prefix, since the actual action would be identical, but would occur on one of two possible connections. Snd-Conn-Req A transport connection is initiated with the peer. Snd-Conn-Ack an acknowledgement is sent in response to a connect request, confirming that the transport layer connection is open. Snd-CER A CER message is sent to the peer. Calhoun et al. expires December 2001 [Page 54] Internet-Draft June 2001 Snd-CEA A CEA message is sent to the peer. Cleanup If necessary, the connection is shutdown, and any local resources are freed. Error The transport layer connection is disconnected, either politely or abortively, in response to an error condition. Local resources are freed. Process-CER A received CER is processed. Process-CEA A received CEA is processed. Disc The transport layer connection is disconnected, and local resources are freed. Elect An election occurs (see Section 8.4 for more information). Snd-Non-DRI A non-DRI message is sent. Snd-DWR A DWR message is sent. Process-DWA The DWA message is serviced. Process A non-DRI Diameter message is serviced. 8.4 The Election Process The election is performed on the responder. The responder compares the Origin-Host received in the DRI sent by its peer with its own Origin-Host (which it may or may not have actually sent). The transport layer connection with the higher value of Origin-Host is the one that survives. The comparison proceeds by considering the shorter OctetString to be null-padded to the length of the longer, then performing an octet by octet unsigned comparison with the first octet being most significant. Hanging octets are assumed to have value 0x80, but dimpled octets are ignored. 9.0 Error Handling There are two different types of errors in Diameter; protocol and applications. A protocol error is one that occurs at the base protocol level, and MAY require per hop attention (e.g. message routing error). Application errors, on the other hand, are generally occur due to a problem with a function specified in a Diameter Calhoun et al. expires December 2001 [Page 55] Internet-Draft June 2001 application (e.g. user authentication, Missing AVP). Result-Code AVP values that are used to report protocol errors MUST be used in the Message-Reject-Answer command. Unlike most Diameter commands, the Message-Reject-Answer does not have a corresponding request. When a request message is received that causes a protocol error, the command code is changed to Message-Reject-Answer, and the Result-Code AVP is set to the appropriate protocol error value. As the answer is sent back towards the originator of the request, each proxy or relay agent MAY take action on the message. 1. Request +---------+ Link Broken +-------------------------->|Diameter |----///----+ | +---------------------| | v +------+--+ | 2. MRA | Relay 2 | +--------+ |Diameter |<-+ (Unable to Forward) +---------+ |Diameter| | | | Home | | Relay 1 |--+ +---------+ | Server | +---------+ | 3. Request |Diameter | +--------+ +-------------------->| | ^ | Relay 3 |-----------+ +---------+ Figure 4 - Example of Protocol Error causing MRA message Figure 4 provides an example of a message forwarded upstream by a Diameter relay. When the message is received by Relay 2, and it detects that it cannot forward the request to the home server, an MRA message is returned with the Result-Code AVP set to DIAMETER_UNABLE_TO_DELIVER. Given that this error falls within the protocol error category, Relay 1 would take special action, and given the error, attempt to route the message through its alternate Relay 3. +---------+ 1. Request +---------+ 2. Request +---------+ | Access |------------>|Diameter |------------>|Diameter | | | | | | Home | | Device |<------------| Relay |<------------| Server | +---------+ 4. Answer +---------+ 3. Answer +---------+ (Missing AVP) (Missing AVP) Figure 5 - Example of Application Error Answer message Figure 5 provides an example of a Diameter message that caused an application error. When application errors occur, the Diameter entity reporting the error clears the 'R' bit in the Command Flags, and adds the Result-Code AVP with the proper value. Application errors do not require any proxy or relay agent involvement, and therefore the Calhoun et al. expires December 2001 [Page 56] Internet-Draft June 2001 message would be forwarded back to the originator of the request. There are certain Result-Code AVP application errors that require additional AVPs to be present in the answer, such as: - An unrecognized AVP is received with the 'M' bit (Mandatory bit) set, causes an answer to be sent with the Result-Code