Network Working Group D. Harrington Internet-Draft Futurewei Technologies Expires: September 5, 2006 J. Salowey Cisco Systems March 4, 2006 Secure Shell Security Model for SNMP draft-ietf-isms-secshell-02.txt Status of This Memo By submitting this Internet-Draft, each author represents that any applicable patent or other IPR claims of which he or she is aware have been or will be disclosed, and any of which he or she becomes aware will be disclosed, in accordance with Section 6 of BCP 79. 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. This Internet-Draft will expire on September 5, 2006. Copyright Notice Copyright (C) The Internet Society (2006). Abstract This memo describes a Security Model for the Simple Network Management Protocol, using the Secure Shell protocol within a Transport Mapping. Harrington & Salowey Expires September 5, 2006 [Page 1] Internet-Draft Secure Shell Security Model for SNMP March 2006 Table of Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.1. The Internet-Standard Management Framework . . . . . . . . 4 1.2. Modularity . . . . . . . . . . . . . . . . . . . . . . . . 4 1.3. Motivation . . . . . . . . . . . . . . . . . . . . . . . . 5 1.4. The Secure Shell Protocol . . . . . . . . . . . . . . . . 6 1.5. Constraints . . . . . . . . . . . . . . . . . . . . . . . 7 1.6. Conventions . . . . . . . . . . . . . . . . . . . . . . . 7 2. How SSHSM Fits into the TMSM Architecture . . . . . . . . . . 8 2.1. Security Capabilities of this Model . . . . . . . . . . . 8 2.1.1. Threats . . . . . . . . . . . . . . . . . . . . . . . 8 2.1.2. SSHSM Sessions . . . . . . . . . . . . . . . . . . . . 11 2.1.3. Authentication Protocol . . . . . . . . . . . . . . . 12 2.1.4. Privacy Protocol . . . . . . . . . . . . . . . . . . . 13 2.1.5. Protection against Message Replay, Delay and Redirection . . . . . . . . . . . . . . . . . . . . . 13 2.1.6. Security Protocol Requirements . . . . . . . . . . . . 13 2.2. Security Parameter Passing Requirement . . . . . . . . . . 15 2.3. Requirements for Notifications . . . . . . . . . . . . . . 15 3. RFC 3411 Abstract Service Interfaces . . . . . . . . . . . . . 16 3.1. Public Abstract Service Interfaces . . . . . . . . . . . . 16 3.1.1. Public ASIs for Outgoing Messages . . . . . . . . . . 16 3.1.2. Public ASIs for Incoming Messages . . . . . . . . . . 18 3.2. SNMP Messages Using this Security Model . . . . . . . . . 20 3.2.1. SNMPv1 and SNMPv2c Messages Using this Security Model . . . . . . . . . . . . . . . . . . . . . . . . 20 3.2.2. SNMPv3 Messages Using this Security Model . . . . . . 20 3.2.3. Passing Security Parameters . . . . . . . . . . . . . 23 3.2.4. MIB Module for SSH Security Model . . . . . . . . . . 25 3.2.5. [todo] Notifications . . . . . . . . . . . . . . . . . 26 3.3. Elements of Procedure . . . . . . . . . . . . . . . . . . 26 3.3.1. Establishing a Session . . . . . . . . . . . . . . . . 26 3.3.2. Closing a Session . . . . . . . . . . . . . . . . . . 29 3.3.3. Discovery . . . . . . . . . . . . . . . . . . . . . . 29 3.3.4. Generating an Outgoing SNMP Message . . . . . . . . . 30 3.3.5. Sending an Outgoing SNMP Message to the Network . . . 32 3.3.6. [todo] Prepare Data Elements from an Incoming SNMP Message . . . . . . . . . . . . . . . . . . . . . . . 33 3.3.7. Processing an Incoming SNMP Message . . . . . . . . . 33 3.4. Overview . . . . . . . . . . . . . . . . . . . . . . . . . 35 3.5. Structure of the MIB Module . . . . . . . . . . . . . . . 35 3.5.1. Textual Conventions . . . . . . . . . . . . . . . . . 36 3.5.2. The sshsmStats Subtree . . . . . . . . . . . . . . . . 36 3.5.3. The sshsmsSession Subtree . . . . . . . . . . . . . . 36 3.5.4. Relationship to Other MIB Modules . . . . . . . . . . 36 3.6. MIB module definition . . . . . . . . . . . . . . . . . . 37 3.7. Implementation Considerations . . . . . . . . . . . . . . 45 Harrington & Salowey Expires September 5, 2006 [Page 2] Internet-Draft Secure Shell Security Model for SNMP March 2006 3.8. Security Considerations . . . . . . . . . . . . . . . . . 45 3.9. IANA Considerations . . . . . . . . . . . . . . . . . . . 47 3.10. Acknowledgements . . . . . . . . . . . . . . . . . . . . . 47 4. References . . . . . . . . . . . . . . . . . . . . . . . . . . 47 4.1. Normative References . . . . . . . . . . . . . . . . . . . 47 4.2. Informative References . . . . . . . . . . . . . . . . . . 48 Appendix A. Open Issues . . . . . . . . . . . . . . . . . . . . . 49 A.1. Issues with Resolutions nearing Consensus . . . . . . . . 51 A.2. Closed Issues . . . . . . . . . . . . . . . . . . . . . . 51 Appendix B. Change Log . . . . . . . . . . . . . . . . . . . . . 53 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 53 Intellectual Property and Copyright Statements . . . . . . . . . . 53 Harrington & Salowey Expires September 5, 2006 [Page 3] Internet-Draft Secure Shell Security Model for SNMP March 2006 1. Introduction This memo describes a Security Model for the Simple Network Management Protocol, using the Secure Shell protocol within a Transport Mapping, referred to as the Secure Shell Security Model (SSHSM). This memo defines a portion of the Management Information Base (MIB) for use with network management protocols in TCP/IP based internets. In particular it defines objects for monitoring and managing the Secure Shell Security Model for SNMP. It is important to understand the SNMP architecture and the terminology of the architecture to understand where the Security Model described in this memo fits into the architecture and interacts with other subsystems within the architecture. 1.1. The Internet-Standard Management Framework For a detailed overview of the documents that describe the current Internet-Standard Management Framework, please refer to section 7 of RFC 3410 [RFC3410]. Managed objects are accessed via a virtual information store, termed the Management Information Base or MIB. MIB objects are generally accessed through the Simple Network Management Protocol (SNMP). Objects in the MIB are defined using the mechanisms defined in the Structure of Management Information (SMI). This memo specifies a MIB module that is compliant to the SMIv2, which is described in STD 58, RFC 2578 [RFC2578], STD 58, RFC 2579 [RFC2579] and STD 58, RFC 2580 [RFC2580]. 1.2. Modularity The reader is expected to have read and understood the description of the SNMP architecture, as defined in [RFC3411],and the "Transport Mapping Security Model (TMSM) for the Simple Network Management Protocol" architecture extension defined in [I-D.ietf-isms-tmsm], which enables the use of external "lower layer" protocols to provide message security, tied into the SNMP architecture through the transport mapping subsystem. One such external protocol is the Secure Shell protocol [RFC4251]. This memo describes the Secure Shell Security Model for SNMP, a specific SNMP security model to be used within the SNMP Architecture, to provide authentication, encryption, and integrity checking of SNMP messages. Harrington & Salowey Expires September 5, 2006 [Page 4] Internet-Draft Secure Shell Security Model for SNMP March 2006 In keeping with the RFC 3411 design decisions to use self-contained documents, this memo includes the elements of procedure plus associated MIB objects which are needed for processing the Secure Shell Security Model for SNMP. These MIB objects SHOULD not be referenced in other documents. This allows the Secure Shell Security Model for SNMP to be designed and documented as independent and self- contained, having no direct impact on other modules, and allowing this module to be upgraded and supplemented as the need arises, and to move along the standards track on different time-lines from other modules. This modularity of specification is not meant to be interpreted as imposing any specific requirements on implementation. 1.3. Motivation Version 3 of the Simple Network Management Protocol (SNMPv3) added security to the previous versions of the protocol. The User Security Model (USM) [RFC3414] was designed to be independent of other existing security infrastructures, to ensure it could function when third party authentication services were not available, such as in a broken network. As a result, USM typically utilizes a separate user and key management infrastructure. Operators have reported that deploying another user and key management infrastructure in order to use SNMPv3 is a reason for not deploying SNMPv3 at this point in time. This memo describes a security model that will make use of the existing and commonly deployed Secure Shell security infrastructure. It is designed to meet the security and operational needs of network administrators, maximize useability in operational environments to achieve high deployment success and at the same time minimize implementation and deployment costs to minimize the time until deployment is possible. The work will address the requirement for the SSH client to authenticate the SSH server, for the SSH server to authenticate the SSH client (the user), and how SNMP can make use of the authenticated identities in authentication and auditing. . The work will include the ability to use any of the user authentication methods described in "SSH Authentication Protocol" [RFC4252] - public key, password, and host-based. Local accounts may be supported through the use of the public key, host-based or password based mechanisms. The password based mechanism allows for integration with deployed password infrastructure such as AAA servers using the RADIUS protocol [RFC2865]. It should be able to take advantage of other defined authentication mechanism such as those Harrington & Salowey Expires September 5, 2006 [Page 5] Internet-Draft Secure Shell Security Model for SNMP March 2006 defined in [I-D.ietf-secsh-gsskeyex] and future mechanism such as those that make use of X.509 certificate credentials. This will allow SSHSM to utilize user authentication and key exchange mechanisms which support different security infrastructures and provide different security properties. It is desirable to use mechanisms that could unify the approach for administrative security for SNMPv3 and Command Line intrfaces (CLI) and other management interfaces. The use of security services provided by Secure Shell is the approach commonly used for the CLI, and is the approach being adopted for use with NETCONF [I-D.ietf- netconf-prot]. Similar to NETCONF over SSH [I-D.ietf-netconf-ssh], this memo describes a method for invoking and running the SNMP protocol within a Secure Shell (SSH) session as an SSH subsystem. This memo defines how SNMP can be used within a Secure Shell (SSH) session, using the SSH connection protocol [RFC4254] over the SSH transport protocol [RFC4253], using SSH user-auth [RFC4252]for authentication. There are a number of challenges to be addressed to map Secure Shell authentication method parameters into the SNMP architecture so that SNMP continues to work without any surprises. These are discussed in detail below. Sections requiring further editing are identified by [todo] markers in the text. Points requiring further WG research and discussion are identified by [discuss] markers in the text. 1.4. The Secure Shell Protocol SSH is a protocol for secure remote login and other secure network services over an insecure network. It consists of three major components: o The Transport Layer Protocol [[RFC4253] provides server authentication, confidentiality, and integrity. It may optionally also provide compression. The transport layer will typically be run over a TCP/IP connection, but might also be used on top of any other reliable data stream. o The User Authentication Protocol [RFC4252] authenticates the client-side user to the server. It runs over the transport layer protocol. o The Connection Protocol [RFC4254] multiplexes the encrypted tunnel into several logical channels. It runs over the transport after succesfully authenticating the user. The client sends a service request once a secure transport layer connection has been established. A second service request is sent after user authentication is complete. This allows new protocols to be defined and coexist with the protocols listed above. Harrington & Salowey Expires September 5, 2006 [Page 6] Internet-Draft Secure Shell Security Model for SNMP March 2006 The connection protocol provides channels that can be used for a wide range of purposes. Standard methods are provided for setting up secure interactive shell sessions and for forwarding ("tunneling") arbitrary TCP/IP ports and X11 connections. 1.5. Constraints The design of this SNMP Security Model is also influenced by the following constraints: 1. When the requirements of effective management in times of network stress are inconsistent with those of security, the design of this model gives preference to the former. 2. In times of network stress, the security protocol and its underlying security mechanisms SHOULD NOT depend upon the ready availability of other network services (e.g., Network Time Protocol (NTP) or AAA protocols). 3. When the network is not under stress, the security model and its underlying security mechanisms MAY depend upon the ready availability of other network services. 4. It may not be possible for the security model to determine when the network is under stress. 5. A security mechanism should entail no changes to the basic SNMP network management philosophy. 1.6. Conventions The terms "manager" and "agent" are not used in this document, because in the RFC 3411 architecture, all SNMP entities have the capability of acting as either manager or agent or both depending on the SNMP applications included in the engine. Where distinction is required, the application names of Command Generator, Command Responder, Notification Generator, Notification Responder, and Proxy Forwarder are used. See "SNMP Applications" [RFC3413] for further information. Throughout this document, the terms "client" and "server" are used to refer to the two ends of the SSH transport connection. The client actively opens the SSH connection, and the server passively listens for the incoming SSH connection. Either SNMP entity may act as client or as server, as discussed further below. The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119]. Harrington & Salowey Expires September 5, 2006 [Page 7] Internet-Draft Secure Shell Security Model for SNMP March 2006 2. How SSHSM Fits into the TMSM Architecture SSH is a security layer which is plugged into the TMSM architecture between the underlying transport layer and the message dispatcher. The SSHSM model will establish an encrypted tunnel between the transport mappings of two SNMP engines. The sending transport mapping security model instance encrypts outgoing messages, and the receiving transport mapping security model instance decrypts the messages. After the transport layer tunnel is established, then SNMP messages can conceptually be sent through the tunnel from one SNMP message dispatcher to another SNMP message dispatcher. Once the tunnel is established, multiple SNMP messages may be able to be passed through the same tunnel. Within an engine, outgoing SNMP messages are passed unencrypted from the message dispatcher to the transport mapping, and incoming messages are passed unencrypted from the transport mapping to the message dispatcher. SSHSM follows the TMSM approach, in which the security-model has two separate areas of security processing - the TMSP performs transport- mapping-related security processing, and the MPSP performs security processing within the security model subsystem of the messaging model (MPSP). SSHSM security processing will be called from within the Transport Mapping functionality of an SNMP engine dispatcher to perform the translation of transport security parameters to/from security-model- independent parameters. Some SSHSM security processing will also be performed within a message processing portion of the model, for compatibility with the ASIs between the RFC 3411 Security Subsystem and the Message Processing Subsystem. 2.1. Security Capabilities of this Model 2.1.1. Threats The security protocols used in this memo are considered acceptably secure at the time of writing. However, the procedures allow for new authentication and privacy methods to be specified at a future time if the need arises. The Secure Shell Security Model provides protection against the threats identified by the RFC 3411 architecture [RFC3411]: Harrington & Salowey Expires September 5, 2006 [Page 8] Internet-Draft Secure Shell Security Model for SNMP March 2006 1. Message stream modification - Provide for verification that each received SNMP message has not been modified during its transmission through the network. . 2. Information modification - Provide for verification that the contents of each received SNMP message has not been modified during its transmission through the network. Data has not been altered or destroyed in an unauthorized manner, nor have data sequences been altered to an extent greater than can occur non- maliciously 3. Masquerade - Provide for both verification of the identity of the user on whose behalf a received SNMP message claims to have been generated, and the verification of the identity of the MIB owner. For the protocols specified in this memo, it is not possible to assure the specific user that originated a received SNMP message; rather, it is the user on whose behalf the message was originated that is authenticated. SSH provides verification of the identity of the MIB owner through the SSH Transport Protocol server authentication [RFC4253] 4. Verification of user identity is important for use with the SNMP access control subsystem, to ensure that only authorized users have access to potentially sensitive data. The SSH user identity will be used to map to an SNMP model-independent securityname for use with SNMP access control. 5. Authenticating the server ensures the authenticity of the SSH server that is associated with the SNMP engine that provides MIB data. Operators or management applications could act upon the data they receive (e.g. raise an alarm for an operator, modify the configuration of the device that sent the notification, modify the configuration of other devices in the network as the result of the notification, and so on), so it is important to know that the data is authentic. SSH allows for authentication of the SSH server using the SSH public key credentials described in [RFC4253] and mechanisms such as those described in [I-D.ietf- secsh-gsskeyex]. 6. Disclosure - Provide, when necessary, that the contents of each received SNMP message are protected from disclosure to unauthorized persons. 7. Replay - Provide for detection of received SNMP messages, which request or contain management information, whose time of generation was not recent. A message whose generation time is outside of a time window is not accepted. Note that message reordering is not dealt with and can occur in normal conditions 2.1.1.1. Data Origin Authentication Issues The RFC 3411 architecture recognizes three levels of security: Harrington & Salowey Expires September 5, 2006 [Page 9] Internet-Draft Secure Shell Security Model for SNMP March 2006 - without authentication and without privacy (noAuthNoPriv) - with authentication but without privacy (authNoPriv) - with authentication and with privacy (authPriv) SSH provides support for encryption and data integrity. While it is technically possible to support noAuthNoPriv and authNoPriv in SSH it is NOT RECOMMENDED by [RFC4253]. This means that an SSH connection SHOULD provide authPriv, which is the highest level of security defined in RFC 3411. It is possible for SSH to skip entity authenticaiton of the client through the "none" authentication method to support anonymous clients, however in this case an implementation MUST still support data integrity within the SSH transport protocol. The security protocols used in [RFC4253] are considered acceptably secure at the time of writing. However, the procedures allow for new authentication and privacy methods to be specified at a future time if the need arises. Implementations SHOULD support whatever authentications are provided by SSH. This includes anonymous access; if SSH supports anonymous access, and SSHSM can extract a username, then anonymous access SHOULD be supported. The authNoPriv security level may be important to accommodate governmental regulation (e.g. export laws) regarding encryption technologies. The transport layer algorithms used to provide data integrity and encryption SHOULD NOT be exposed to the SSHSM layer. In SNMP, we deliberately avoided this, and settled for an assertion, using msgFlags, that auth and priv were applied according to the rules of the security model. However, there should probably be an SSH-MIB, so the algorithms used to achieve the security level should be accessible to authorized administrators via a management interface. SSH should provide the identity of the authenticated principal. From this information it should be possible for the SNMP subsystem to determine if the session is allowed access to the subsystem. 2.1.1.1.1. noAuthPriv SSH provides the "none" userauth method, which is normally rejected by servers and used only to find out what userauth methods are supported. However, it is legal for a server to accept this method, which has the effect of not authenticating the ssh client to the ssh server. Doing this does not compromise authentication of the ssh server to the ssh client, nor does it compromise data confidentiality Harrington & Salowey Expires September 5, 2006 [Page 10] Internet-Draft Secure Shell Security Model for SNMP March 2006 or data integrity. The RFC 3411 architecture does not permit noAuthPriv. SSHSM should refuse a noAuthPriv session. 2.1.1.1.2. skipping public key verification Most key exchange algorithms are able to authenticate the SSH server's identity to the client. However, for the common case of DH signed by public keys, this requires the client to know the host's public key a priori and to verify that the correct key is being used. If this step is skipped, then authentication of the ssh server to the ssh client is not done. Data confidentiality and data integrity protection to the server still exist, but these are of dubious value when an attacker can insert himself between the client and the real ssh server. Note that some userauth methods may defend against this situation, but many of the common ones (including password and keyboard-interactive) do not, and in fact depend on the fact that the server's identity has been verified (so passwords are not disclosed to an attacker). 2.1.1.1.3. the 'none' MAC algorithm SSH provides the "none" MAC algorithm, which would allow you to turn off data integrity while maintaining confidentiality. However, if you do this, then an attacker may be able to modify the data in flight, which means you effectively have no authentication. SSH must not be configured using the "none" MAC algorithm for use with the SSHSM security model. 2.1.2. SSHSM Sessions The Secure Shell security model will utilize sessions, with a single user and security level associated with each session. All SSHSM sessions will utilize authPriv securityLevels, and all SNMP messages will be authenticated and encrypted. SSHSM sessions are established during the elements of procedure for an outgoing message, never during the elements of procedure for an incoming message. Implementations MAY choose to instantiate sessions in anticipation of outgoing messages. [todo] Say more about how sessions are initiated, how session state is made visible and so on. An SSHSM session is associated with state information that is maintained for its lifetime. SSH ensures that cryptographic keys Harrington & Salowey Expires September 5, 2006 [Page 11] Internet-Draft Secure Shell Security Model for SNMP March 2006 established at the beginning of the SSH session and stored in the SSH session state are fresh new session keys generated for each session. Thes eare used to authenticate and encrypt data, to prevent replay across sessions. SSH uses sequence information to prevent the replay and reordering of messages within a session. 2.1.2.1. Message security versus session security As part of session creation, the client and server entities are typically authenticated and authorized access to the session. In addition, as part of session establishment, cryptographic key material is exchanged and is then used to control access to the session on a message by message basis. Messages that fail the basic data origin authenticaiton/ data integrity checks will be rejected. Entities receiving the messages that do not have the correct encryption keys established during session creation will not be able to read the messages. In order for an entity to process messages, it must maintain certain state associated with the session. This includes, but is not limited to, cryptographic encryption and data integrity keys, entity identities and authorization information associated with the authenticated identites. After a message is received and passes integrity and authentication checks, the state stored in the session is used to provide further authorization for the message. 2.1.3. Authentication Protocol SSHSM should support any user authentication mechanism supported by SSH. This includes the three authentication methods described in the SSH Authentication Protocol document - publickey, password, and host- based. The password authentication mechanism allows for integration with deployed password based infrastructure. It is possible to hand a password to a service such as RADIUS [RFC2865] or Diameter [RFC3588] for validation. The validation could be done using the user-name and user-password attributes. It is also possible to use a different password validation protcol such as CHAP [RFC1994] or digest authentication [RFC 2617, draft-ietf-radext-digest-auth-04] to integrate with RADIUS or Diameter. Any of these mechanism leave the password in the clear on the device that is authenticating the password which introduces threats on the authentication infrastructure which is less than ideal. GSSKeyex [I-D.ietf-secsh-gsskeyex] provides a framework for the addition of user authentication mechanisms which support different security infrastructures and provide different security properties. Additional authentication mechanisms, such as one that supports X.509 Harrington & Salowey Expires September 5, 2006 [Page 12] Internet-Draft Secure Shell Security Model for SNMP March 2006 certificates, may be added to SSH in the future. 2.1.4. Privacy Protocol The Secure Shell Security Model uses the SSH transport layer protocol, which provides strong encryption, server authentication, and integrity protection. 2.1.5. Protection against Message Replay, Delay and Redirection The Secure Shell Security Model uses the SSH transport layer protocol. SSH uses sequence numbers and integrity checks to protect against replay and reordering of messages within a connection. SSH also provides protection against replay of entire sessions. In a properly-implemented DH exchange, both sides will generate new random numbers for each exchange, which means the exchange hash and thus the encryption and integrity keys will be distinct for every session. This would prevent capturing an SNMP message and redirecting it to another SNMP engine. Message delay is not as important an issue with SSH as it is with USM. USM checks the timeliness of messages because it does not provide session protection or message sequence ordering. The only delay that would seem to be possible would be to delay the transmission of all packets from a particular point in a session since SSH protects the ordering of packets. 2.1.6. Security Protocol Requirements Modifying the Secure Shell protocol, or configuring it in a particular manner, may change its security characteristics in ways that would impact other existing usages. If a change is necessary, the change should be an extension that has no impact on the existing usages. This document will describe the use of an SSH subsytem for SNMP. 2.1.6.1. Troubleshooting SSHSM will likely not work in conditions where access to the CLI has stopped working and, in situations where SNMP access has to work when the CLI has stopped working, the use of USM should be considered instead of SSHSM. [todo] establish a mechanism to determine when session establishment is repeatedly failing, and how to determine whether to fallback to USM. Harrington & Salowey Expires September 5, 2006 [Page 13] Internet-Draft Secure Shell Security Model for SNMP March 2006 2.1.6.2. Coexistence The Secure Shell security model can coexist with the USM security model, the only other currently defined security model. [discuss] #6: Are there are any wrinkles to coexistence with SNMPv1/v2c/USM? Note that RFC3584 discusses how to transfer fields between SNMPv3 and SNMPv1 messages. [todo] this area needs detailed analysis. 2.1.6.3. Mapping SSH to EngineID In the RFC3411 architecture, there are three use cases for an engineID: snmpEngineID - RFC3411 includes the SNMP-FRAMEWORK-MIB, which defines a snmpEngineID object. An snmpEngineID is the unique and unambiguous identifier of an SNMP engine. Since there is a one- to-one association between SNMP engines and SNMP entities, it also uniquely and unambiguously identifies the SNMP entity within an administrative domain. contextEngineID - Management information resides at an SNMP entity where a Command Responder Application has local access to potentially multiple contexts. This application uses a contextEngineID equal to the snmpEngineID of its associated SNMP engine. securityEngineID - The RFC3411 architecture defines ASIs that include a securityEngineID - the authoritative SNMP entity - which is either the local snmpEngineID or the target snmpEngineID, depending on the type of operation. Since a security model might utilize shared credentials and integrity-checking parameters, and the datastores of the two endpoints could get out of sync, the "authoritative" engineID indicates which end has the values to be used. The securityEngineID is used by USM when performing integrity checking and authentication, to look up values in the USM tables, and to synchronize "clocks". The securityEngineID is not needed by SSHSM, since integrity checking and authentication are handled outside the SNMP engine. [discuss] #7: is there still a need for an "authoritative SNMP engine"? Does authoritative have any meaning in a TMSM/SSHSM environment? In SNMPv3, the authoritative engine is usually the engine with the command responder, i.e. the agent; in non-proxy situations, securityEngineID equals contextEngineID. in client-server terms, the authoritative engine is usually the server. So, should the SNMP engine associated with the SSH server be authoritative? Would Infoms change that? Would bidirectional messaging change that? Would call-home change that? Do we need to set the securityEngineID Harrington & Salowey Expires September 5, 2006 [Page 14] Internet-Draft Secure Shell Security Model for SNMP March 2006 to indicate which side is the SSH server? 2.2. Security Parameter Passing Requirement Specific parameters for an incoming message can be determined from the transport layer by the transport mapping security processing (TMSP), before the message processing begins, and for outgoing messages, the security-model-specific parameters are gathered by the messaging-security-processing (MPSP) and passed with the outgoing message to the transport mapping. For outgoing messages, the MPSP portion of the security model creates the WholeMsg from its component parts. In the SSHSM model, an SNMPv3 message is built without any content in the SecurityParameters field of the message, and the WholeMsg is passed unencrypted back to the Message Processing Model for forwarding to the Transport Mapping. The MPSP takes input provided by the SNMP application, converts that information into suitable security parameters for SSHSM, and passes these in a cache referenced by tmStateReference to the TMSP (via the dispatcher). The TMSP establishes sessions as needed and passes messages to the SSH subsystem for processing. For incoming messages, the TMSP accepts (decrypted) messages from the SSH subsystem, and records the transport-related information and the security-related information, including authenticated identity, in a cache referenced by tmStateReference, and passes the WholeMsg and the tmStateReference to the MPSP (via the dispatcher). The cache reference could be thought of as an additional parameter in the ASIs between the transport mapping and the messaging security model. This approach does create dependencies between a model-specific TPSP and a corresponding specific MPSP. If a TMSM-model-independent ASI parameter is passed, this approach would be consistent with the securityStateReference cache already being passed around in the ASI. 2.3. Requirements for Notifications SSH connections may be initiated by command generators or by notification originators. Command generators are frequently operated by a human, but notification originators frequently are unmanned automated processes. As a result, it will be necessary to provision authentication credentials on the SNMP engine containing the notification originator so it can successfully authenticate to an engine containing a notification receiver. [discuss] #9: Can an existing R/R session be reused for Harrington & Salowey Expires September 5, 2006 [Page 15] Internet-Draft Secure Shell Security Model for SNMP March 2006 notifications? There is some text in Appendix A in RFC 3430 [RFC3430]which captured some of these discussions when RFC 3430 was written. 3. RFC 3411 Abstract Service Interfaces Abstract service interfaces have been defined by RFC 3411 to describe the conceptual data flows between the various subsystems within an SNMP entity. The Secure Shell Security Model uses some of these conceptual data flows when communicating with other subsystems, such as the Message Processing Subsystem. These RFC 3411-defined data flows are referred to here as public interfaces. 3.1. Public Abstract Service Interfaces 3.1.1. Public ASIs for Outgoing Messages The IN parameters of the prepareOutgoingMessage() ASI are used to pass information from the dispatcher (application subsystem) to the message processing subsystem. The OUT parameters are used to pass information from the message processing subsystem to the dispatcher and on to the transport mapping: statusInformation = -- success or errorIndication prepareOutgoingMessage( IN transportDomain -- transport domain to be used IN transportAddress -- transport address to be used IN messageProcessingModel -- typically, SNMP version IN securityModel -- Security Model to use IN securityName -- on behalf of this principal IN securityLevel -- Level of Security requested IN contextEngineID -- data from/at this entity IN contextName -- data from/in this context IN pduVersion -- the version of the PDU IN PDU -- SNMP Protocol Data Unit IN expectResponse -- TRUE or FALSE IN sendPduHandle -- the handle for matching -- incoming responses OUT destTransportDomain -- destination transport domain OUT destTransportAddress -- destination transport address OUT outgoingMessage -- the message to send OUT outgoingMessageLength -- its length ) The abstract service primitive from a Message Processing Model to a Security Model to generate the components of a Request message is: Harrington & Salowey Expires September 5, 2006 [Page 16] Internet-Draft Secure Shell Security Model for SNMP March 2006 statusInformation = -- success or errorIndication generateRequestMsg( IN messageProcessingModel -- typically, SNMP version IN globalData -- message header, admin data IN maxMessageSize -- of the sending SNMP entity IN securityModel -- for the outgoing message IN securityEngineID -- authoritative SNMP entity IN securityName -- on behalf of this principal IN securityLevel -- Level of Security requested IN scopedPDU -- message (plaintext) payload OUT securityParameters -- filled in by Security Module OUT wholeMsg -- complete generated message OUT wholeMsgLength -- length of generated message ) The abstract service primitive from a Message Processing Model to a Security Model to generate the components of a Response message is: statusInformation = -- success or errorIndication generateResponseMsg( IN messageProcessingModel -- typically, SNMP version IN globalData -- message header, admin data IN maxMessageSize -- of the sending SNMP entity IN securityModel -- for the outgoing message IN securityEngineID -- authoritative SNMP entity IN securityName -- on behalf of this principal IN securityLevel -- Level of Security requested IN scopedPDU -- message (plaintext) payload IN securityStateReference -- reference to security state -- information from original -- request OUT securityParameters -- filled in by Security Module OUT wholeMsg -- complete generated message OUT wholeMsgLength -- length of generated message ) The abstract data elements passed as parameters in the abstract service primitives are as follows: [todo] check each parameter and determine if it is necessary for SSHSM and whether the description is accurate o statusInformation - An indication of whether the encoding and securing of the message was successful. If not it is an indication of the problem. o messageProcessingModel - The SNMP version number for the message to be generated. This data is not used by the User-based Security module. Harrington & Salowey Expires September 5, 2006 [Page 17] Internet-Draft Secure Shell Security Model for SNMP March 2006 o globalData - The message header (i.e., its administrative information). This data is not used by the User-based Security module. o maxMessageSize - The maximum message size as included in the message. This data is not used by the User-based Security module. o securityParameters - These are the security parameters. They will be filled in by the SSH Security module. o securityModel - The securityModel in use. Should be SSH Security Model. o securityName - identifies a principal to be used for securing an outgoing message. The securityName has a format that is independent of the Security Model. In case of a response this parameter is ignored and the value from the cache is used. o securityLevel - The Level of Security from which the SSH Security module determines if the message needs to be protected from disclosure and if the message needs to be authenticated. o securityEngineID - The snmpEngineID of the authoritatvie SNMP engine to which a dateRequest message is to be sent. In case of a response it is implied to be the processing SNMP engine's snmpEngineID and so if it is specified, then it is ignored. o scopedPDU - The message payload. The data is opaque as far as the SSH Security Model is concerned. o securityStateReference - A handle/reference to cachedSecurityData to be used when securing an outgoing Response message. This is the exact same hsecurityStateReference as was generated by the SSH Security module when processing the incoming Request message to which this is the Response message. o wholeMsg - The fully encoded SNMP message ready for sending on the wire. o wholeMsgLength - The length of the encoded SNMP message (wholeMsg). Upon completion of the process, the SSH Security module returns statusInformation. If the process was successful, the completed message is returned, without the privacy and authentication applied yet. If the process was not successful, then an errorIndication is returned. 3.1.2. Public ASIs for Incoming Messages The abstract service primitive from a Transport Mapping (in the dispatcher) to a Message Processing Model for a received message is:: Harrington & Salowey Expires September 5, 2006 [Page 18] Internet-Draft Secure Shell Security Model for SNMP March 2006 result = -- SUCCESS or errorIndication prepareDataElements( IN transportDomain -- origin transport domain IN transportAddress -- origin transport address IN wholeMsg -- as received from the network IN wholeMsgLength -- as received from the network OUT messageProcessingModel -- typically, SNMP version OUT securityModel -- Security Model to use OUT securityName -- on behalf of this principal OUT securityLevel -- Level of Security requested OUT contextEngineID -- data from/at this entity OUT contextName -- data from/in this context OUT pduVersion -- the version of the PDU OUT PDU -- SNMP Protocol Data Unit OUT pduType -- SNMP PDU type OUT sendPduHandle -- handle for matched request OUT maxSizeResponseScopedPDU -- maximum size sender can accept OUT statusInformation -- success or errorIndication -- error counter OID/value if error OUT stateReference -- reference to state information -- to be used for possible Response ) The abstract service primitive from a Message Processing Model to the Security Subsystem for a received message is:: statusInformation = -- errorIndication or success -- error counter OID/value if error processIncomingMsg( IN messageProcessingModel -- typically, SNMP version IN maxMessageSize -- of the sending SNMP entity IN securityParameters -- for the received message IN securityModel -- for the received message IN securityLevel -- Level of Security IN wholeMsg -- as received on the wire IN wholeMsgLength -- length as received on the wire OUT securityEngineID -- authoritative SNMP entity OUT securityName -- identification of the principal OUT scopedPDU, -- message (plaintext) payload OUT maxSizeResponseScopedPDU -- maximum size sender can handle OUT securityStateReference -- reference to security state ) -- information, needed for response Harrington & Salowey Expires September 5, 2006 [Page 19] Internet-Draft Secure Shell Security Model for SNMP March 2006 3.2. SNMP Messages Using this Security Model The syntax of an SNMP message using this Security Model adheres to the message format defined in the version-specific Message Processing Model document (for example [RFC3412]). At the time of this writing, there are three defined message formats - SNMPv1, SNMPv2c, and SNMPv3. 3.2.1. SNMPv1 and SNMPv2c Messages Using this Security Model Since message security is provided by a "lower layer", the message does not need to carry message security parameters. The securityModel and securityName parameters are determined by the Secure Shell Security Model from the SSH service. SSHSM requires that transport always be authenticated and integrity-checked and encrypted, so all SSHSM messages are authPriv. Since an incoming SNMPv1 or SNMPv2c message lacks a msgFlags field, the msgFlags is always treated as authPriv. The community string is not used as an authentication mechansism, since user authentication is provided by SSH userauth. The community string is still used to provide context information. To be clear, the community strin gi snot touched, and just shipped opaquely, so people who use the community string in proprietary ways to identify contexts should not be impacted. The SNMPv1 and SNMPv2c message formats do not contain a contextEngineID, but do contain an IP Address field that can be used to perform proxy, and where implemented by the agent, the snmpEngineID at the IP address can be learned by querying the device with a GET request. 3.2.2. SNMPv3 Messages Using this Security Model RFC 3412 defines two primitives, generateRequestMsg() and processIncomingMsg() which require the specification of an authoritative SNMP entity. [discuss] #10: which securityparameters must be supported for the SSHSM model, and why? Which services provided in USM are needed in TMSM/SSHSM? How does the Message Processing model provides this information to the security model via generateRequestMsg() and processIncomingMsg() primitives? The SNMPv3Message SEQUENCE is defined in [RFC3412]. The following fields are specific to the Secure Shell Security Model: Harrington & Salowey Expires September 5, 2006 [Page 20] Internet-Draft Secure Shell Security Model for SNMP March 2006 SNMPv3MessageSyntax DEFINITIONS IMPLICIT TAGS ::= BEGIN SNMPv3Message ::= SEQUENCE { -- identify the layout of the SNMPv3Message -- this element is in same position as in SNMPv1 -- and SNMPv2c, allowing recognition -- the value 3 is used for snmpv3 msgVersion INTEGER ( 0 .. 2147483647 ), -- administrative parameters msgGlobalData HeaderData, -- security model-specific parameters -- format defined by Security Model msgSecurityParameters OCTET STRING, msgData ScopedPduData } HeaderData ::= SEQUENCE { msgID INTEGER (0..2147483647), msgMaxSize INTEGER (484..2147483647), msgFlags OCTET STRING (SIZE(1)), -- .... ...1 authFlag -- .... ..1. privFlag -- .... .1.. reportableFlag -- Please observe: -- .... ..00 is OK, means noAuthNoPriv -- .... ..01 is OK, means authNoPriv -- .... ..10 reserved, MUST NOT be used. -- .... ..11 is OK, means authPriv msgSecurityModel INTEGER (1..2147483647) } ScopedPduData ::= CHOICE { plaintext ScopedPDU, encryptedPDU OCTET STRING -- encrypted scopedPDU value } ScopedPDU ::= SEQUENCE { contextEngineID OCTET STRING, contextName OCTET STRING, data ANY -- e.g., PDUs as defined in [RFC3416] } END Harrington & Salowey Expires September 5, 2006 [Page 21] Internet-Draft Secure Shell Security Model for SNMP March 2006 3.2.2.1. msgGlobalData SSHSM requires that transport always be authenticated, integrity- checked, and encrypted, so all SSHSM messages are authPriv. The msgFlags MUST always be set to authPriv. msgSecurityModel is set to the IANA-assigned value for the Secure Shell Security Model. See http://www.iana.org/assignments/snmp-number-spaces. 3.2.2.1.1. msgSecurityParameters Since message security is provided by a "lower layer", and the securityName parameter is always determined from the SSH authentication method, the SNMP message does not need to carry message security parameters within the msgSecurityParameters field. To prevent its being used in a manner that could be damaging, such as for carrying a virus or worm, when used with SSHSM, it is an empty field. The field msgSecurityParameters in SNMPv3 messages has a data type of OCTET STRING. Its value MUST be the BER serialization of the following ASN.1 sequence: SSHSMSecurityParametersSyntax DEFINITIONS IMPLICIT TAGS ::= BEGIN SSHsmSecurityParameters :: SEQUENCE { OCTET STRING } END 3.2.2.1.2. msgFlags For an outgoing message, msgFlags is the requested security for the message; if a SSHSM cannot provide the requested securityLevel, the request MUST be discarded and SHOULD notify the message processing model that the request failed. For an outgoing message, it is acceptable for the SSHSM to provide stronger than requested security. To avoid the need to mess with the ASN.1 encoding, the SNMPv3 message carries the requested msgFlags, not the actual securityLevel applied to the message. If a message format other than SNMPv3 is used, then the new message may carry the more accurate securityLevel in the SNMP message. Harrington & Salowey Expires September 5, 2006 [Page 22] Internet-Draft Secure Shell Security Model for SNMP March 2006 3.2.3. Passing Security Parameters For each message received, the Security Model caches the state information such that a Response message can be generated using the same security information, even if the Configuration Datastore is altered between the time of the incoming request and the outgoing response. For SSHSM, there are three levels of state that need to be maintained: the session, the message, and the model-independent translations. The tmStateReference is used to pass model- and mechanism-specific parameters to coordinate the session-related activities and specific message pair processing between the TMSP and MPSP. The SSHSM has the responsibility for explicitly releasing the complete tmStateReference when the session is destroyed. The SSHSM has the responsibility for releasing the message-specific parameters in the tmStateReference once a response message has been sent, or the data is no longer needed. The MPSP translates select parameters from the tmStateReference cache into model-independent parameters subsequently passed in the securityStateReference cache to a Message Processing Model. The Message Processing Model has the responsibility for explicitly releasing the securityStateReference if such data is no longer needed. The securityStateReference cached data may be implicitly released via the generation of a response, or explicitly released by using the stateRelease primitive, as described in RFC 3411 section 4.5.1." SSH does not require that a session be maintained nor that it be closed when the keys associated with the host or client associated with the session are changed. Some SSH implementations may close an existing session if the keys associated with the session change. For SSHSM, if the session is closed between the time a Request is received and a Response message is being prepared, then the Response should be discarded. 3.2.3.1. Transport Session Parameters SSHSM will create a session between the TMSM of one SNMP entity and the TMSM of another SNMP entity. The created SSH "tunnel" MUST provide authentication of the client and server, and MUST integrity- check and encrypt the messages. Upon establishment of an SSH session, the TMSP will cache the transport parameters in the tmStateReference for subsequent usage. This information should be stored in a local datastore. Harrington & Salowey Expires September 5, 2006 [Page 23] Internet-Draft Secure Shell Security Model for SNMP March 2006 The tmStateReference cache for use with the SSH Authentication Protocol [RFC4252] will include the following transport-related information: [discuss] #15: What data needs to be stored in the tmStateReference, and how does SSHSM get the information from SSH, for the various authentication and transport options? tmSessionID = a unique local identifier tmTransportDomain = tDomainSSH tmTransportAddress = x.x.x.x:y tmSecurityModel - SSHSM tmSecurityLevel = "authPriv" Additional information will be added to the tmStateReference by the authentication portion of the SSHSM. [discuss] #16B: Passing a securityname might be useful for passing as a hint to RADIUS or other authorization mechanism to indicate which identity we want to use when doing access control, and RADIUS,etc. can tell us whether the username being authenticated is allowed to be mapped to that authorization/accounting identity. Should we provide securityname when establishing a session, so the authentication machanisms can use it as a hint? 3.2.3.1.1. Authenticating Servers and Clients tmSecurityName = the user name authenticated by SSH tmSecurityName is the name that has been successfully authenticated by SSH, from the user name field of the SSH_MSG_USERAUTH_REQUEST message. How this data is extracted from the SSH environment to put into the SNMP environment is implementation-dependent. [todo] #18: I currently have multiple sections, one for each known auth mechanism. We need to discuss the parameters that need to be cached for each. Once we are complete, I will collapse this into one section. 3.2.3.2. [discuss] Using Passwords to Authenticate SNMP Principals Upon creation of a SSH session, the TMSP will cache the session authentication information in the tmStateReference: tmSecurityName is the name extracted from the user name field of the SSH_MSG_USERAUTH_REQUEST message, after authentication has completed successfully. Harrington & Salowey Expires September 5, 2006 [Page 24] Internet-Draft Secure Shell Security Model for SNMP March 2006 tmAuthMechanism = "password" tmAuthProtocol = "password" tmSecurityLevel = appropriate choice from SnmpSecurityLevel tmAuthzRef = "[todo] authorization data obtained during the exchange" 3.2.3.3. [discuss] Using Public keys to Authenticate SNMP Principals Upon creation of a SSH session, the TMSP will cache the session authentication information in the tmStateReference: tmSecurityName is the name extracted from the user name field of the SSH_MSG_USERAUTH_REQUEST message tmAuthMechanism = "publickey" tmAuthProtocol = public key algorithm name tmSecurityLevel = appropriate choice from SnmpSecurityLevel tmAuthzRef = "[todo] authorization data obtained during the exchange" 3.2.3.4. [discuss] Using Host-based Authentication of SNMP Principals Upon creation of a SSH session, the TMSP will cache the session authentication information in the tmStateReference: tmSecurityName is the name used in user name field of the SSH_MSG_USERAUTH_REQUEST message tmAuthMechanism = "hostbased" tmAuthProtocol = public key algorithm for host key tmSecurityLevel = appropriate choice from SnmpSecurityLevel tmAuthzRef = "[todo] authorization data obtained during the exchange" 3.2.3.5. securityStateReference for SSHSM The parameters associated with an incoming request message to be applied to the outgoing response. messageProcessingModel = SNMPv3 securityModel = SSHSM sessionID = tmSessionID 3.2.4. MIB Module for SSH Security Model Each security model should use its own MIB module, rather than utilizing the USM MIB, to eliminate dependencies on a model that could be replaced some day. See RFC 3411 section 4.1.1. [todo] the mapping from model-specific identity to a model independent securityName for storage in an LCD is implementation- Harrington & Salowey Expires September 5, 2006 [Page 25] Internet-Draft Secure Shell Security Model for SNMP March 2006 dependent. This is Implementation-dependent, both in the case of extracting tmSecurityname from SSH for an incoming message, and for providing an LCD mapping. [todo] Module needs to be worked out once things become stable.. 3.2.5. [todo] Notifications For notifications, if no session has yet been created, or the session has been closed, then the TMSP will establish a session and populate the cache for subsequent usage. [discuss] #21: we need to determine what data should be persistent and stored in the LCD for notification purposes. 3.3. Elements of Procedure 3.3.1. Establishing a Session The Secure Shell Security Model provides the following primitive to pass data back and forth between the Transport Mapping portion of the Security Model and the SSH service: statusInformation establishSession( IN destTransportDomain -- transport domain to be used IN destTransportAddress -- transport address to be used IN securityModel -- Security Model to use IN securityEngineID -- SNMP entity IN securityName -- on behalf of this principal IN securityLevel -- Level of Security requested IN subsystem OUT sessionID ) The following describes the procedure to follow to establish a session between a client and sever to run SNMP over SSH. This process is followed by any SNMP engine establishing a session for subsequent use. In practice, this is done by an application that initiates a transaction, such as a Command Generator or a Notification Originator or a Proxy Forwarder. It is never triggered by an application preparing a response message, such as a Command Responder or Notification Receiver, because securityStatereference will always have session information for a response message The parameters necessary to establish a session are provided by the Harrington & Salowey Expires September 5, 2006 [Page 26] Internet-Draft Secure Shell Security Model for SNMP March 2006 Secure Shell Security Model to the SSH client code, using the establishSession() ASI. 1) If the securityLevel specifies that the message is to be authenticated, but the SSH implementation does not support an authentication protocol, then the message cannot be sent. An error indication (unsupportedSecurityLevel) is returned to the calling module. 2) If the securityLevel specifies that the message is to be protected from disclosure, but the SSH implementation does not support encryption, then the message cannot be sent. An error indication (unsupportedSecurityLevel) is returned to the calling module. 3) Using destTransportDomain and destTransportAddress, the client will establish an SSH transport connection using the SSH transport protocol, and the client and server will mutually authenticate, and exchange keys for message integrity and encryption. if the attempt to establish a connection is successful, then tmStateReference is created, and the values of transportDomain and transportAddress are saved. If the attempt to establish a connection is unsuccessful, then an error indication [todo] will be returned, and [todo] processing stops. [discuss] #22: There are a significant number of security problems associated with mapping to a transport address which may need to be discussed in the security considerations section. 4) The provided securityEngineID and securityName and securityLevel are used to lookup the associated entry in the Local Configuration Datastore (LCD), and the model-specific information concerning the principal at the destination is extracted. This step allows preconfiguration of model-specific principals mapped to the engine/ name/level, for example, for sending notifications using host-only authentication. Set the username in the SSH_MSG_USERAUTH_REQUEST to the username extracted from the LCD. If information about the user is absent from the LCD, then set the username in the SSH_MSG_USERAUTH_REQUEST to the value of securityName. This allows a deployment without preconfigured mappings between model-specific and model-independent names, but the securityName will need to contain a username recognized by the authentication mechanism. 5)The client will then invoke the "ssh-userauth" service to authenticate the user, as described in the SSH authentication protocol [RFC4252]. Harrington & Salowey Expires September 5, 2006 [Page 27] Internet-Draft Secure Shell Security Model for SNMP March 2006 6) If the authentication is unsuccessful, then the transport connection should be closed, tmStateReference is discarded, the message is discarded, an error indication (unknownSecurityName) is returned to the calling module, and processing stops for this message. 7) Once the user has been successfully authenticated, the client will invoke the "ssh- connection" service, also known as the SSH connection protocol [RFC4254]. 8) After the ssh-connection service is established, the client will use an SSH_MSG_CHANNEL_OPEN message to open a channel of type "session", providing a selected sender channel number, and a maximum packet size based on maxMessageSize. 9) If successful, this will result in an SSH session. The destTransportDomain nd the destTransportAddress, plus the "recipient channel" and "sender channel" and other relevant data from the SSH_MSG_CHANNEL_OPEN_CONFIRMATION are added to the tmStateReference for subsequent use. 10) Running SNMP as an SSH subsystem avoids the need for the script to recognize shell prompts or skip over extraneous information, such as a system message that is printed at shell start-up. Once the SSH session has been established, the SNMP engine will invoke SNMP as an SSH subsystem, as indicated in the "subsystem" parameter. In order to allow SNMP traffic to be easily identified and filtered by firewalls and other network devices, servers associated with SNMP entities using the Secure Shell Security Model MUST default to providing access to the "SNMP" SSH subsystem only when the SSH session is established using the IANA-assigned TCP port (TBD). Servers SHOULD be configurable to allow access to the SNMP SSH subsystem over other ports. [todo] check whether there is a better way to establish a tunnel for SNMP messages. [discuss] We must perform some type of engineID discovery to provide the mapping between transport address, session, and engineID at this point in the session establishment procedure? We have an established channel; can we simply send a GET of snmpEngineID and record the value in the tmStateReference? 11) [todo] the engine will perform an SNMP GET command requesting the value of the remote engine's snmpEngineID object, and create a tmStateReference cache recording the following information: Harrington & Salowey Expires September 5, 2006 [Page 28] Internet-Draft Secure Shell Security Model for SNMP March 2006 the remote engine's snmpEngineID the transport address the recipient and sender channels 3.3.2. Closing a Session The Secure Shell Security Model provides the following primitive to pass data back and forth between the Security Model and the SSH service: statusInformation closeSession( IN sessionID ) The following describes the procedure to follow to close a session between a client and sever to run SNMP over SSH. This process is followed by any SNMP engine closing the corresponding SNMP session. The Secure Shell Security Model identifies which session should be closed to the SSH client code, using the closeSession() ASI. [discuss] #23: We need to discuss the circumstances under which a session should be closed, and how an SNMP engine should determine if, and respond if the SSH session is closed by other means. 3.3.3. Discovery Since snmpEngineID isn't really needed for authentication and integrity checking, it becomes useful primarily for contextEngineID. contextEngineID is useful for proxy, and for a management application to uniquely identify an SNMP entity. Since snmpEngineID is an object in the SNMP-FRAMEWORK-MIB, the mapping between engineID and transport address could be established after a tunnel is established, or could be determined using noAuthNoPriv (with suitable caveats). [discuss] #24: How should we enable auto-discovery? Auto-discovery of SNMP devices is an important feature of many NMS platforms. Should we simply use a noAuthNoPriv request, and recommend an associated access control configuration that only makes accessible relatively benign data such as sysOID, sysDescription, and snmpEngineID? Should we standardize this approach for all TMSM models, including a "named policy" for what can be discovered (a policy to be configured within whatever access control system is used)? Harrington & Salowey Expires September 5, 2006 [Page 29] Internet-Draft Secure Shell Security Model for SNMP March 2006 Alternatively, can we let USM perform discovery so we don't have to attenpt to establish an SSH connection first? USM is the mandatory- to-implement security model, so this could make sense. 3.3.4. Generating an Outgoing SNMP Message This section describes the procedure followed by the Secure Shell Security Model whenever it generates a message containing a management operation (like a request, a response, a notification, or a report) on behalf of a user. The parameters needed are supplied by the Message Processing Model via the generateRequestMsg() or the generateResponseMsg() ASI statusInformation = -- success or errorIndication generateRequestMsg( IN messageProcessingModel -- typically, SNMP version IN globalData -- message header, admin data IN maxMessageSize -- of the sending SNMP entity IN securityModel -- for the outgoing message IN securityEngineID -- authoritative SNMP entity IN securityName -- on behalf of this principal IN securityLevel -- Level of Security requested IN scopedPDU -- message (plaintext) payload OUT securityParameters -- filled in by Security Module OUT wholeMsg -- complete generated message OUT wholeMsgLength -- length of generated message OUT tmStateReference -- reference to session info ) Harrington & Salowey Expires September 5, 2006 [Page 30] Internet-Draft Secure Shell Security Model for SNMP March 2006 statusInformation = -- success or errorIndication generateResponseMsg( IN messageProcessingModel -- typically, SNMP version IN globalData -- message header, admin data IN maxMessageSize -- of the sending SNMP entity IN securityModel -- for the outgoing message IN securityEngineID -- authoritative SNMP entity IN securityName -- on behalf of this principal IN securityLevel -- Level of Security requested IN scopedPDU -- message (plaintext) payload IN securityStateReference -- reference to security state -- information from original -- request OUT securityParameters -- filled in by Security Module OUT wholeMsg -- complete generated message OUT wholeMsgLength -- length of generated message OUT tmStateReference -- reference to session info ) 1) verify securityModel = sshsmSecurityModel determine whether we need to use the SSH subsystem for Request/ Responses ("SNMP"), or for Notifications ("SNMPNotification") or Reports. [discuss] #34 - how do we determine this? [discuss] #35 - which subsystem is used for Reports? 2) If there is a securityStateReference, extract the tmStateReference information from the cachedSecurityData from the Request message. At this point, the cachedSecurityData can now be discarded. [todo] clarify which data can be discarded. 2b) [todo] #13 - If the message is a Response, and a session never existed or has been closed, or the Request/Response subsystem never existed or was closed, then discard the message, and generate a Report 3) If there is no securityStateReference, then lookup the session info indexed by {securityModel, securityName, securityLevel}, and set tmStateReference. [todo] insert check for msgflags versus session/transport characterstics here, and in the transport-mapping portion. 4) If there is no session info for this index, then create an incomplete tmStateReference indexed by the provided {securityName, securityLevel}. Store the securityModel and maxMessageSize information. When the TMSP gets the incomplete tmStateReference, it will recognize that it needs to establish a new session, and fill in the rest of the information for subsequent use. 5) fill in the securityParameters with the serialization of a zero-length OCTET STRING. 6) The wholeMsg is now serialized and then represents the unauthenticated message being prepared. Harrington & Salowey Expires September 5, 2006 [Page 31] Internet-Draft Secure Shell Security Model for SNMP March 2006 7) The completed message (wholeMsg) with its length (wholeMsgLength) and securityParameters (a zero-length octet string) and tmStateReference is returned to the calling module with the statusInformation set to success. The Message Processing Model then passes information to the disptacher for forwarding to the Transport Mapping. 3.3.5. Sending an Outgoing SNMP Message to the Network The TMSP portion of the Secure Shell Security Model performs the following tasks: 8) Uses tmStateReference to lookup session information. 9) [todo] verifies that auth and priv can be provided, as requested, and error-out if not. [todo] insert check for msgflags versus session/transport characterstics here. 10) If the session information is incomplete (i.e, has no tmTransportAddress), then call establishSession() using the destTransportDomain and destTransportAddress (the output of the PrepareOutgoingMessage() ASI) and the securityModel, securityEngineID, securityName, securityLevel from the tmStateReference. Store all information in the tmStateReference for subsequent use. [discuss] #25: Where is the best place to call establishSession()? Note that the whole message is completely put together within the message-processing portion of the security model, in the hopes that a session will be able to be established when the message gets to the transport mapping portion of the architecture. It is done this way because the RFC3411 arcitecture doesn't pass the transport addressing info into the security model via messaging model. It would seem a much more efficient approach to verify that the session can be established, while still in the security model portion of the messaging model. If we don't establish the session until we get to the transport mapping, we've done a lot of work for nothing. And thus far, there is no place to record failed attempts to establish a session, so an engine doesn't learn to not try to open a session. In an environment where the SNMP engine might be a daemon used by multiple applications, an attacker could use this to cause a denial of service attack at the NMS. This would likely occur on the NMS side. I don't know if there's any way to cause it to happen on the agent side. I suppose a rogue agent with callhome functionality might be able to cause a denial of service for an NMS by repeatedly requesting callhome and then refusing the connections. 11) An SSH_MSG_CHANNEL_DATA message is sent, indicating the recipient channel and encapsulating the wholeMessage. Harrington & Salowey Expires September 5, 2006 [Page 32] Internet-Draft Secure Shell Security Model for SNMP March 2006 [discuss] #26: According to RFC 3411, section 4.1.1, the application provides the transportDomain and transportAddress to the PDU dispatcher via the sendPDU() primitive. If we permit multiple sessions per transportAddress, then we would need to define how session identifiers get passed from the application to the PDU dispatcher (and then to the MP model). [discuss] #28: For notification tables, how do we predefine the dynamic session identifiers? We might have a MIB module that records the session information for subsequent use by the applications and other subsytems, or it might be passed in the tmStateReference cache. For notifications, I assume the SNMPv3 notification tables would be a place to find the address, but I'm not sure how to identify the presumably-dynamic session identifiers. The MIB module could identify whether the session was initiated by the remote engine or initiated by the current engine, and possibly assigned a purpose (incoming request/response or outgoing notifications). 3.3.6. [todo] Prepare Data Elements from an Incoming SNMP Message For an incoming message, the TMSP will need to put information from the transport mechanisms used into the tmStateReference so the MPSP can extract the information and add it conceptually to the securityStateReference. 3.3.7. Processing an Incoming SNMP Message This section describes the procedure followed by an SNMP engine whenever it receives a message containing a management operation on behalf of a user. To simplify the elements of procedure, the release of state information is not always explicitly specified. As a general rule, if state information is available when a message gets discarded, the message-state information should also be released, and if state information is available when a session is closed, the session state information should also be released. Also, an error indication can return an OID and value for an incremented counter and optionally a value for securityLevel, and values for contextEngineID or contextName for the counter. In addition, the securityStateReference data is returned if any such information is available at the point where the error is detected. [todo] this paragraph may no longer be accurate because of persistent session state information. The abstract service primitive from a Message Processing Model to the Security Subsystem for a received message is:: Harrington & Salowey Expires September 5, 2006 [Page 33] Internet-Draft Secure Shell Security Model for SNMP March 2006 statusInformation = -- errorIndication or success -- error counter OID/value if error processIncomingMsg( IN messageProcessingModel -- typically, SNMP version IN maxMessageSize -- of the sending SNMP entity IN securityParameters -- for the received message IN securityModel -- for the received message IN securityLevel -- Level of Security IN wholeMsg -- as received on the wire IN wholeMsgLength -- length as received on the wire OUT securityEngineID -- authoritative SNMP entity OUT securityName -- identification of the principal OUT scopedPDU, -- message (plaintext) payload OUT maxSizeResponseScopedPDU -- maximum size sender can handle OUT securityStateReference -- reference to security state ) -- information, needed for response 1) If the received securityParameters is not the serialization of an OCTET STRING formatted according to the SSHsmSecurityParameters, then the snmpInASNParseErrs counter [RFC3418] is incremented, and an error indication (parseError) is returned to the calling module. Note that we return without the OID and value of the incremented counter, which may be important if this security model supports generating a Report PDU (which SSHSM doesn't so far), because in this case there is not enough information to generate a Report PDU. [todo] check whether this field parses correctly and report errors through Reports 2) The SSHSM queries the associated SSH engine, in an implementation- dependent manner, to determine the transport and security parameters for the received message: a) the transportDomain and transportAddress b) tmSecurityName - an identifier for the authenticated entity c) whether authentication is on or off, d) whether encryption is on or off, e) integrity-checking options 3) The securityEngineID to be returned to the caller is determined in an implementation-dependent manner, such as by using the transport address to perform a lookup in its Local Configuration Datastore (LCD). If the securityEngineID is unknown, then an SNMP engine may perform discovery to create a new entry in its LCD and continue processing. Note that securityEngineID is required by the SNMPv3 message processing model in RFC 3412 section 7.2 13a) 4) If the information about the message security indicates that the security options do not match the securityLevel requested by the Harrington & Salowey Expires September 5, 2006 [Page 34] Internet-Draft Secure Shell Security Model for SNMP March 2006 caller, then the SSHsmStatsUnsupportedSecLevels counter is incremented and an error indication (unsupportedSecurityLevel) together with the OID and value of the incremented counter is returned to the calling module. 5) The scopedPDU component is assumed to be in plain text and is the message payload to be returned to the calling module. 7) The maxSizeResponseScopedPDU is calculated. This is the maximum size allowed for a scopedPDU for a possible Response message. Provision is made for a message header that allows the same securityLevel as the received Request. 10) Information about the value of tmSecurityName is extracted from the Local Configuration Datastore (LCD) to provide conversion from the SSH authentication-method-specific tmSecurityName to a model- independent securityName. If no information is available for the username in the LCD, then the securityName is set to the username associated with the session. 11) The security data is cached as cachedSecurityData, so that a possible response to this message can and will use the same authentication and privacy parameters. Information to be saved/ cached is as follows: [todo] copy from the "Passing Security Parameters" section above. transportDomain, transportAddress securityEngineID SSH username, auth options encryption options Integrity checking options 12) The statusInformation is set to success and a return is made to the calling module passing back the OUT parameters as specified in the processIncomingMsg primitive. 3.4. Overview 3.5. Structure of the MIB Module Objects in this MIB module are arranged into subtrees. Each subtree is organized as a set of related objects. The overall structure and assignment of objects to their subtrees, and the intended purpose of each subtree, is shown below. Harrington & Salowey Expires September 5, 2006 [Page 35] Internet-Draft Secure Shell Security Model for SNMP March 2006 3.5.1. Textual Conventions Generic and Common Textual Conventions used in this document can be found summarized at http://www.ops.ietf.org/mib-common-tcs.html 3.5.2. The sshsmStats Subtree This subtree contains SSHSM security-model-dependent counters. This subtree provides information for identifying fault conditions and performance degradation. 3.5.3. The sshsmsSession Subtree This subtree contains SSHSM security-model-dependent information about sessions. 3.5.4. Relationship to Other MIB Modules Some management objects defined in other MIB modules are applicable to an entity implementing this MIB. In particular, it is assumed that an entity implementing the TMSM-MIB module will also implement the SNMPv2-MIB [RFC3418] and the TMSM-MIB [I-D.ietf-isms-tmsm]. This MIB module is for managing SSHSM-specific information. 3.5.4.1. Relationship to the SNMPv2-MIB The 'system' group in the SNMPv2-MIB [RFC3418] is defined as being mandatory for all systems, and the objects apply to the entity as a whole. The 'system' group provides identification of the management entity and certain other system-wide data. The SSHSM-MIB does not duplicate those objects. 3.5.4.2. Relationship to the TMSM-MIB The 'tmsmSession' group in the TMSM-MIB [I-D.ietf-isms-tmsm] is defined as being applicable to all Transport-Mapping Security Models that use sessions. 3.5.4.3. MIB Modules Required for IMPORTS The following MIB module imports items from [RFC2578], [RFC2579], [RFC2580], [RFC3411], [RFC3419], and [I-D.ietf-isms-tmsm] Harrington & Salowey Expires September 5, 2006 [Page 36] Internet-Draft Secure Shell Security Model for SNMP March 2006 3.6. MIB module definition ** Is AES the only officially required to support SSH encryption ** mechanisms? It seems RFC 4344 has much more to offer. BTW, is it ** useful to export all this information in an SSHSM MIB module? Some ** of the stuff seems generic SSH... SSHSM-MIB DEFINITIONS ::= BEGIN IMPORTS MODULE-IDENTITY, OBJECT-TYPE, OBJECT-IDENTITY, mib-2, Counter32, Integer32 FROM SNMPv2-SMI TestAndIncr, AutonomousType FROM SNMPv2-TC MODULE-COMPLIANCE, OBJECT-GROUP FROM SNMPv2-CONF snmpAuthProtocols, snmpPrivProtocols, SnmpAdminString, SnmpSecurityLevel, SnmpEngineID FROM SNMP-FRAMEWORK-MIB TransportAddress, TransportAddressType FROM TRANSPORT-ADDRESS-MIB tmsmSessionID FROM TMSM-MIB ; sshsmMIB MODULE-IDENTITY LAST-UPDATED "200509020000Z" ORGANIZATION "ISMS Working Group" CONTACT-INFO "WG-EMail: isms@lists.ietf.org Subscribe: isms-request@lists.ietf.org Chairs: Juergen Quittek NEC Europe Ltd. Network Laboratories Kurfuersten-Anlage 36 69115 Heidelberg Germany +49 6221 90511-15 quittek@netlab.nec.de Juergen Schoenwaelder International University Bremen Campus Ring 1 28725 Bremen Germany +49 421 200-3587 Harrington & Salowey Expires September 5, 2006 [Page 37] Internet-Draft Secure Shell Security Model for SNMP March 2006 j.schoenwaelder@iu-bremen.de Co-editors: David Harrington Effective Software 50 Harding Rd Portsmouth, New Hampshire 03801 USA +1 603-436-8634 ietfdbh@comcast.net Joseph Salowey Cisco Systems 2901 3rd Ave Seattle, WA 98121 USA jsalowey@cisco.com " DESCRIPTION "The Secure Shell Security Model MIB Copyright (C) The Internet Society (2005). This version of this MIB module is part of RFC XXXX; see the RFC itself for full legal notices. -- NOTE to RFC editor: replace XXXX with actual RFC number -- for this document and remove this note " REVISION "200509020000Z" -- 02 September 2005 DESCRIPTION "The initial version, published in RFC XXXX. -- NOTE to RFC editor: replace XXXX with actual RFC number -- for this document and remove this note " ::= { mib-2 xxxx } -- RFC Ed.: replace xxxx with IANA-assigned number and -- remove this note -- ---------------------------------------------------------- -- -- subtrees in the SSHSM-MIB -- ---------------------------------------------------------- -- sshsmNotifications OBJECT IDENTIFIER ::= { sshsmMIB 0 } sshsmObjects OBJECT IDENTIFIER ::= { sshsmMIB 1 } sshsmConformance OBJECT IDENTIFIER ::= { sshsmMIB 2 } -- ------------------------------------------------------------- -- Objects -- ------------------------------------------------------------- Harrington & Salowey Expires September 5, 2006 [Page 38] Internet-Draft Secure Shell Security Model for SNMP March 2006 TransportAddressSSH ::= TEXTUAL-CONVENTION DISPLAY-HINT "1a" STATUS current DESCRIPTION "[discuss] Represents either a hostname encoded in ASCII using the IDNA protocol [RFC3490] followed by a colon ':' (ASCII character 0x3A) and a decimal port number in ASCII, or an IP address followed by a colon ':' (ASCII character 0x3A) and a decimal port number in ASCII. The name SHOULD be fully qualified whenever possible. Values of this textual convention are not directly useable as transport-layer addressing information, and require runtime resolution. As such, applications that write them must be prepared for handling errors if such values are not supported, or cannot be resolved (if resolution occurs at the time of the management operation). The DESCRIPTION clause of TransportAddress objects that may have TransportAddressSSH values must fully describe how (and when) such names are to be resolved to IP addresses and vice versa. This textual convention SHOULD NOT be used directly in object definitions since it restricts addresses to a specific format. However, if it is used, it MAY be used either on its own or in conjunction with TransportAddressType or TransportDomain as a pair. When this textual convention is used as a syntax of an index object, there may be issues with the limit of 128 sub-identifiers specified in SMIv2, STD 58. In this case, the OBJECT-TYPE declaration MUST include a 'SIZE' clause to limit the number of potential instance sub-identifiers." SYNTAX OCTET STRING (SIZE (1..255)) transportDomainSSH OBJECT-IDENTITY STATUS current DESCRIPTION "The SSH transport domain. The corresponding transport address is of type TransportAddressSSH." ::= { snmpDomains xxxx } -- RFC Ed.: replace xxxx with IANA-assigned number and -- remove this note sshsmPasswordAuthProtocol OBJECT-IDENTITY STATUS current Harrington & Salowey Expires September 5, 2006 [Page 39] Internet-Draft Secure Shell Security Model for SNMP March 2006 DESCRIPTION "The Secure Shell Password Authentication Method" REFERENCE "RFC 4252" ::= { snmpAuthProtocols 4 } sshsmPublickeyAuthProtocol OBJECT-IDENTITY STATUS current DESCRIPTION "The Secure Shell Public Key Authentication Method" REFERENCE "RFC 4252" ::= { snmpAuthProtocols 5 } sshsmHostbasedAuthProtocol OBJECT-IDENTITY STATUS current DESCRIPTION "The Secure Shell Host-based Authentication Method" REFERENCE "RFC 4252" ::= { snmpAuthProtocols 6 } sshsmAESPrivProtocol OBJECT-IDENTITY STATUS current DESCRIPTION "The AES Encryption Protocol." ::= { snmpPrivProtocols 5 } -- Statistics for the Secure Shell Security Model sshsmStats OBJECT IDENTIFIER ::= { sshsmObjects 1 } -- [todo] do we need any of these? or other stats? sshsmStatsUnsupportedSecLevels OBJECT-TYPE SYNTAX Counter32 MAX-ACCESS read-only STATUS current DESCRIPTION "The total number of packets received by the SNMP engine which were dropped because they requested a securityLevel that was unknown to the SNMP engine or otherwise unavailable. [todo] we should never hit any of these because they should never be sent by the remote SNMP engine if an appropriate session does not exist. We also do not know what was requested by the remote session. " ::= { sshsmStats 1 } Harrington & Salowey Expires September 5, 2006 [Page 40] Internet-Draft Secure Shell Security Model for SNMP March 2006 sshsmStatsUnknownUserNames OBJECT-TYPE SYNTAX Counter32 MAX-ACCESS read-only STATUS current DESCRIPTION "The total number of packets received by the SNMP engine which were dropped because they referenced a user that was not known to the SNMP engine. [discuss] In SSHSM, we do no preconfiguration, so we don't know any SSH users. If authentication is based on principals defined in the SSH authentication, if the user is not known by SSH, the message wouldn't reach the SNMP engine, so this count would always be zero. " ::= { sshsmStats 3 } sshsmStatsUnknownEngineIDs OBJECT-TYPE SYNTAX Counter32 MAX-ACCESS read-only STATUS current DESCRIPTION "The total number of packets received by the SNMP engine which were dropped because they referenced an snmpEngineID that was not known to the SNMP engine. [todo] We don't use the engineID during authentication, encryption, or integrity checking, so there is never an error condition related to unknown securityEngineID. (But check the RFC3413 and the RFC3584 dependencies on knowing the securityEngineID.) " ::= { sshsmStats 4 } -- The sshsmSession Group sshsmSession OBJECT IDENTIFIER ::= { sshsmObjects 2 } sshsmSessionSpinLock OBJECT-TYPE SYNTAX TestAndIncr MAX-ACCESS read-write STATUS current DESCRIPTION "An advisory lock used to allow several cooperating Command Generator Applications to coordinate their use of facilities to create sessions in the usmUserTable. " ::= { sshsmSession 1 } Harrington & Salowey Expires September 5, 2006 [Page 41] Internet-Draft Secure Shell Security Model for SNMP March 2006 sshsmSessionTable OBJECT-TYPE SYNTAX SEQUENCE OF SshsmSessionEntry MAX-ACCESS not-accessible STATUS current DESCRIPTION "The table of currently available sessions configured in the SNMP engine's Local Configuration Datastore (LCD) for SNMP over SSH sessions. Sessions are created as needed, and do not persist across network management system reboots. " ::= { sshsmSession 2 } sshsmSessionEntry OBJECT-TYPE SYNTAX SshsmSessionEntry MAX-ACCESS not-accessible STATUS current DESCRIPTION "A session configured in the SNMP engine's Local Configuration Datastore (LCD) for the Secure Shell Security Model. " INDEX { sshsmSessionID } ::= { sshsmSessionTable 1 } SshsmSessionEntry ::= SEQUENCE { sshsmSessionID Integer32, sshsmTMSMSession tmsmSessionID, sshsmSessionTDomain transportDomain, sshsmSessionTAddress transportAddress, sshsmSessionUserName SnmpAdminString, sshsmSessionSecurityName SnmpAdminString, sshsmSessionSecurityLevel SnmpSecurityLevel } sshsmSessionID OBJECT-TYPE SYNTAX Integer32 (1..65535) MAX-ACCESS not-accessible STATUS current DESCRIPTION "A locally-unique identifier for a session. " ::= { sshsmSessionEntry 1 } sshsmSessionID OBJECT-TYPE SYNTAX Integer32 (1..65535) MAX-ACCESS not-accessible STATUS current Harrington & Salowey Expires September 5, 2006 [Page 42] Internet-Draft Secure Shell Security Model for SNMP March 2006 DESCRIPTION "A locally-unique identifier for a TMSM session. This is the associated tmsmSessionID from TMSM-MIB. " ::= { sshsmSessionEntry 2 } sshsmSessionTDomain OBJECT-TYPE SYNTAX TransportDoaminSSH MAX-ACCESS read-only STATUS current DESCRIPTION "The transport domain associated with this session. " ::= { sshsmSessionEntry 3 } sshsmSessionTAddress OBJECT-TYPE SYNTAX TransportAddressSSH MAX-ACCESS read-only STATUS current DESCRIPTION "The hostname and port, or the transport address associated with this session. " ::= { sshsmSessionEntry 4 } sshsmSessionUserName OBJECT-TYPE SYNTAX SnmpAdminString MAX-ACCESS read-only STATUS current DESCRIPTION "A human readable string representing the principal in Security Model dependent format, such as the the user name used in the SSH-USERAUTH-REQUEST message for a successful authentication. " ::= { sshsmSessionEntry 5 } sshsmSessionSecurityName OBJECT-TYPE SYNTAX SnmpAdminString MAX-ACCESS read-only STATUS current DESCRIPTION "A human readable string representing the principal in Security Model independent format. The default transformation of the Secure Shell Security Model dependent security ID to the securityName and vice versa is the identity function so that the securityName is the same as the SSH user name. " Harrington & Salowey Expires September 5, 2006 [Page 43] Internet-Draft Secure Shell Security Model for SNMP March 2006 ::= { sshsmSessionEntry 6 } sshsmSessionSecurityLevel OBJECT-TYPE SYNTAX SnmpSecurityLevel MAX-ACCESS read-only STATUS current DESCRIPTION "The Level of Security at which SNMP messages can be sent using this session, in particular, one of: noAuthNoPriv - without authentication and without privacy, authNoPriv - with authentication but without privacy, authPriv - with authentication and with privacy. " DEFVAL { authPriv } ::= { sshsmSessionEntry 7 } -- ------------------------------------------------------------- -- sshsmMIB - Conformance Information -- ------------------------------------------------------------- sshsmGroups OBJECT IDENTIFIER ::= { sshsmConformance 1 } sshsmCompliances OBJECT IDENTIFIER ::= { sshsmConformance 2 } -- ------------------------------------------------------------- -- Units of conformance -- ------------------------------------------------------------- sshsmGroup OBJECT-GROUP OBJECTS { sshsmStatsUnsupportedSecLevels, sshsmStatsUnknownUserNames, sshsmStatsUnknownEngineIDs, sshsmSessionTMSession, sshsmSessionTDomain, sshsmSessionTAddress, sshsmSessionTransportDomain, sshsmSessionAddress, sshsmSessionUserName, sshsmSessionSecurityName, sshsmSessionSecurityLevel, sshsmSessionAuthProtocol, sshsmSessionPrivProtocol, sshsmSessionEngineID, sshsmSessionPrivProtocol, Harrington & Salowey Expires September 5, 2006 [Page 44] Internet-Draft Secure Shell Security Model for SNMP March 2006 sshsmSessionSpinLock } STATUS current DESCRIPTION "A collection of objects for maintaining session information of an SNMP engine which implements the SNMP Secure Shell Security Model. " ::= { sshsmGroups 2 } -- ------------------------------------------------------------- -- Compliance statements -- ------------------------------------------------------------- sshsmCompliance MODULE-COMPLIANCE STATUS current DESCRIPTION "The compliance statement for SNMP engines that support the SSHSM-MIB" MODULE MANDATORY-GROUPS { sshsmGroup } ::= { sshsmCompliances 1 } END 3.7. Implementation Considerations [discuss] #27: The SNMP over TCP Transport Mapping document [RFC3430]says that TCP connections can be recreated dynamically or kept for future use and actually leaves all that to the transport mapping. Do we need to discuss these issues? Where? 3.8. Security Considerations This document describes a security model that would permit SNMP to utilize SSH security services. [todo] expand as needed. SSHSM relies on SSH mutual authentication, binding of keys, confidentiality and integrity. Any authentication method that meets the requirements of the SSH architecture will provide the properties of mutual authentication and binding of keys. While SSH does support turning off confidentiality and integrity, they SHOULD NOT be turned off when used with SSHSM. SSHv2 provides Perfect Forward Security (PFS) for encryption keys. PFS is a major design goal of SSH, and any well-designed keyex algorithm will provide it. Harrington & Salowey Expires September 5, 2006 [Page 45] Internet-Draft Secure Shell Security Model for SNMP March 2006 [todo] We will probably need to discuss the security implications of password based authentication methods. SSHSM has no way to verify that server authentication was performed, to learn the host's public key in advance, or verify that the correct key is being used. SSHSM simply trusts that these are properly handled by the implementer and deployer. There are a number of management objects defined in this MIB module with a MAX-ACCESS clause of read-write and/or read-create. Such objects may be considered sensitive or vulnerable in some network environments. The support for SET operations in a non-secure environment without proper protection can have a negative effect on network operations. These are the tables and objects and their sensitivity/vulnerability: o [todo] There are no management objects defined in this MIB module that have a MAX-ACCESS clause of read-write and/or read-create. So, if this MIB module is implemented correctly, then there is no risk that an intruder can alter or create any management objects of this MIB module via direct SNMP SET operations. Some of the readable objects in this MIB module (i.e., objects with a MAX-ACCESS other than not-accessible) may be considered sensitive or vulnerable in some network environments. It is thus important to control even GET and/or NOTIFY access to these objects and possibly to even encrypt the values of these objects when sending them over the network via SNMP. These are the tables and objects and their sensitivity/vulnerability: o [todo] SNMP versions prior to SNMPv3 did not include adequate security. Even if the network itself is secure (for example by using IPSec), even then, there is no control as to who on the secure network is allowed to access and GET/SET (read/change/create/delete) the objects in this MIB module. It is RECOMMENDED that implementers consider the security features as provided by the SNMPv3 framework (see [RFC3410], section 8), including full support for the SNMPv3 cryptographic mechanisms (for authentication and privacy). Further, deployment of SNMP versions prior to SNMPv3 is NOT RECOMMENDED. Instead, it is RECOMMENDED to deploy SNMPv3 and to enable cryptographic security. It is then a customer/operator responsibility to ensure that the SNMP entity giving access to an instance of this MIB module is properly configured to give access to Harrington & Salowey Expires September 5, 2006 [Page 46] Internet-Draft Secure Shell Security Model for SNMP March 2006 the objects only to those principals (users) that have legitimate rights to indeed GET or SET (change/create/delete) them. 3.9. IANA Considerations IANA is requested to assign: 1. a TCP port number in the range 1..1023 in the http://www.iana.org/assignments/port-numbers registry which will be the default port for SNMP over SSH sessions as defined in this document, 2. an SMI number under mib-2, for the MIB module in this document, 3. an SnmpSecurityModel for the Secure Shell Security Model, as documented in the MIB module in this document, 4. "snmp" as an SSH Service Name in the http://www.iana.org/assignments/ssh-parameters registry. 3.10. Acknowledgements The editors would like to thank Jeffrey Hutzelman for sharing his SSH insights. 4. References 4.1. Normative References [RFC2578] McCloghrie, K., Ed., Perkins, D., Ed., and J. Schoenwaelder, Ed., "Structure of Management Information Version 2 (SMIv2)", STD 58, RFC 2578, April 1999. [RFC2579] McCloghrie, K., Ed., Perkins, D., Ed., and J. Schoenwaelder, Ed., "Textual Conventions for SMIv2", STD 58, RFC 2579, April 1999. [RFC2580] McCloghrie, K., Perkins, D., and J. Schoenwaelder, "Conformance Statements for SMIv2", STD 58, RFC 2580, April 1999. [RFC2865] Rigney, C., Willens, S., Rubens, A., and W. Simpson, "Remote Authentication Dial In User Service (RADIUS)", RFC 2865, June 2000. [RFC3411] Harrington, D., Presuhn, R., and B. Wijnen, "An Architecture for Describing Simple Network Management Protocol (SNMP) Management Frameworks", STD 62, RFC 3411, December 2002. [RFC3412] Case, J., Harrington, D., Presuhn, R., and B. Wijnen, Harrington & Salowey Expires September 5, 2006 [Page 47] Internet-Draft Secure Shell Security Model for SNMP March 2006 "Message Processing and Dispatching for the Simple Network Management Protocol (SNMP)", STD 62, RFC 3412, December 2002. [RFC3414] Blumenthal, U. and B. Wijnen, "User-based Security Model (USM) for version 3 of the Simple Network Management Protocol (SNMPv3)", STD 62, RFC 3414, December 2002. [RFC3418] Presuhn, R., "Management Information Base (MIB) for the Simple Network Management Protocol (SNMP)", STD 62, RFC 3418, December 2002. [RFC3419] Daniele, M. and J. Schoenwaelder, "Textual Conventions for Transport Addresses", RFC 3419, December 2002. [RFC3430] Schoenwaelder, J., "Simple Network Management Protocol Over Transmission Control Protocol Transport Mapping", RFC 3430, December 2002. [RFC4251] Ylonen, T. and C. Lonvick, "The Secure Shell (SSH) Protocol Architecture", RFC 4251, January 2006. [RFC4252] Ylonen, T. and C. Lonvick, "The Secure Shell (SSH) Authentication Protocol", RFC 4252, January 2006. [RFC4253] Ylonen, T. and C. Lonvick, "The Secure Shell (SSH) Transport Layer Protocol", RFC 4253, January 2006. [RFC4254] Ylonen, T. and C. Lonvick, "The Secure Shell (SSH) Connection Protocol", RFC 4254, January 2006. [I-D.ietf-isms-tmsm] Harrington, D. and J. Schoenwaelder, "Transport Mapping Security Model (TMSM) for the Simple Network Management Protocol", draft-ietf-isms-tmsm-00 (work in progress), October 2005. 4.2. Informative References [RFC3410] Case, J., Mundy, R., Partain, D., and B. Stewart, "Introduction and Applicability Statements for Internet- Standard Management Framework", RFC 3410, December 2002. [RFC3413] Levi, D., Meyer, P., and B. Stewart, "Simple Network Management Protocol (SNMP) Applications", STD 62, RFC 3413, December 2002. [RFC3588] Calhoun, P., Loughney, J., Guttman, E., Zorn, G., and J. Harrington & Salowey Expires September 5, 2006 [Page 48] Internet-Draft Secure Shell Security Model for SNMP March 2006 Arkko, "Diameter Base Protocol", RFC 3588, September 2003. [I-D.ietf-netconf-prot] Enns, R., "NETCONF Configuration Protocol", draft-ietf-netconf-prot-11 (work in progress), February 2006. [I-D.ietf-netconf-ssh] Wasserman, M. and T. Goddard, "Using the NETCONF Configuration Protocol over Secure Shell (SSH)", draft-ietf-netconf-ssh-05 (work in progress), October 2005. [I-D.ietf-secsh-gsskeyex] Hutzelman, J., "GSSAPI Authentication and Key Exchange for the Secure Shell Protocol", draft-ietf-secsh-gsskeyex-10 (work in progress), August 2005. Appendix A. Open Issues We need to reach consensus on some issues. I numbered the [discuss] markers in the text for easy correlation to the issue discussions. *** When discussing these issues, please use the provided # in the subject line, and please limit the message to one topic at a time. *** Here is the current list of issues from the SSHSM document where we need to reach consensus. #3: we need some text contributed to discuss the implications of sessions on SNMP. #4: Should the SSHSM document include a discussion of the operational expectations of this model for use in troubleshooting a broken network, or can this be covered in the TMSM document? (Either way, we could use some contributed text on the topic) #5: Should the SSHSM document include a discussion of ways SNMP could be extended to better support management/monitoring needs when a network is running just fine, or can this be covered in the TMSM document, or in an applicability document? #6: Are there are any wrinkles to coexistence with SNMPv1/v2c/USM? #7: is there still a need for an "authoritative SNMP engine"? #9: Can an existing R/R session be reused for notifications? #10: a) which securityparameters must be supported for the SSHSM model? b) Which services provided in USM are needed in TMSM/SSHSM? C) How does the Message Processing model provide this information to the security model via generateRequestMsg() and processIncomingMsg() primitives? Harrington & Salowey Expires September 5, 2006 [Page 49] Internet-Draft Secure Shell Security Model for SNMP March 2006 #12: a) how does SSHSM determine whether SSH can provide the security services requested in msgFlags? B) There were discussions about whether it was acceptable for a transport- mapping-model to provide stronger security than requested. Does this need to be discussed in the SSHSM document, or should we discuss this in the TMSM document? c) when sending a message into an environment where encryption is not legal, how do we ensure that encryption is not provided? #15: What data needs to be stored in the tmStateReference, and how does SSHSM get the information from SSH, for the various authentication and transport options? #16 B) passing a securityname might be useful for passing as a hint to RADIUS or other authorization mechanism to indicate which identity we want to use when doing access control, and RADIUS,etc. can tell us whether the username being authenticated is allowed to be mapped to that authorization/accounting identity. Should we provide securityname when establishing a session, so the authentication machanisms can use it as a hint? #17: I believe somebody suggested we require mutual authentication. I'm not sure I understand the edits. #21: we need to determine what data should be persistent and stored in the LCD for notification purposes. #22: Joe: There are a significant number of security problems associated with mapping to a transport address which may need to be discussed in the security considerations section. #23: We need to discuss the circumstances under which a session should be closed, and how an SNMP engine should determine if, and respond if the SSH session is closed by other means #24: How should we enable auto-discovery? #25: Where is the best place to call establishSession()? See the "Sending an Outgoing Message to the Network" section for more details on this issue. #26: According to RFC 3411, section 4.1.1, the application provides the transportDomain and transportAddress to the PDU dispatcher via the sendPDU() primitive. If we permit multiple sessions per transportAddress, then we would need to define how session identifiers get passed from the application to the PDU dispatcher (and then to the MP model). #27: The SNMP over TCP Transport Mapping document (RFC3430) says that TCP connections can be recreated dynamically or kept for future use and actually leaves all that to the transport mapping. Do we need to discuss these issues? Where? in the security considerations? #28: For notification tables, how do we predefine the dynamic session identifiers? #31: Is maxSizeResponseScopedPDU relevant? Can it be calculated once for the session? Do we need to take into consideration the SSH window size? Harrington & Salowey Expires September 5, 2006 [Page 50] Internet-Draft Secure Shell Security Model for SNMP March 2006 #33: does the mib need to be writable, so sessions can be preconfigured, such as for callhome, or would it be populated at creation time by the underlying instrumentation, and not writable by SNMP? [discuss] #34 - how do we determine whether a PDU contains a Request /Responseor a Notification? ** Reports are a reaction to a previously received message and thus they go wherever the previous message triggering the report came from. [discuss] #35 - which subsystem is used for Reports? [todo] We need to define somewhere what the max message size is that needs to be supported over the SSH transport. RFC 3430 says in 2.2 that implementations have to support 8192 octets... A.1. Issues with Resolutions nearing Consensus A.2. Closed Issues #1: is it important to support anonymous user access to SNMP? Resolution: We should support whatever authorizations are provided by SSH; if SSH supports anonymous access, and SSHSM can extract a username, then it should be supported. #2: a) is server authentication a requirement that SNMP will require of the client? b) how can we verify that server authentication was performed, or do we take simply trust the SSH client layer to perform such authentication? c) for the common case of DH signed by public keys, how does the client learn the host's public key in advance, and verify that the correct key is being used? #8: Do we need a mapping between the SSH key (or other SSH engine identifier) and SNMP engineID? What happens if an agent "spoofs" another engineID, and an NMS perfoms a SET of sensitive parameters to the agent? Resolution: we do not need to address this for local SSH and local snmpEngineID, unless smebody can show a use case requirement. There is likely to be a need to map, in an implementation-dependent manner, the remote engineIDs with the associated SSH host (mapping of engineID/transport address/host key). #11: If we eliminate all msgSecurityParameters, should the msgSecurityParameters field in the SNMPv3 message simply be a zero- length OCTET STRING, or should it be an ASN.1 NULL? It MUST be a BER-encoded OCTET STRING #13: will SSHSM be impacted by keychanges to the SSH local datastore? Resolution: if the session is closed whe the Response is being prepared, discard the Response. #14: MUST the SSHSM model provide mutual authentication of the client Harrington & Salowey Expires September 5, 2006 [Page 51] Internet-Draft Secure Shell Security Model for SNMP March 2006 and server, and MUST it authenticate, integrity-check, and encrypt the messages? Resolution: yes. #16: The SSH server doesn't necessarily authorize the name carried in the SSH_MSG_USERAUTH_REQUEST message, but may return a different name or list of names that are authorized to be used given the authentication of the provided username. Resolution: this is mistaken; the username from the SSH_MSG_USERAUTH_REQUEST SHOULD be used. A) What should be the source of the SSHSM mechanism-specific username for mapping to securityname? Resolution: the username from the SSH_MSG_USERAUTH_REQUEST SHOULD be used. #18: I currently have multiple sections, one for each known auth mechanism. We need to discuss the parameters that need to be cached for each, and determine whether we can collapse this into one section. a) Using Passwords to Authenticate SNMP Principals B) Using Public keys to Authenticate SNMP Principals C) Using Host-based Authentication of SNMP Principals Resolution: I will collapse this later, after we have verified we have considered all current/likely scenarios. #19: RADIUS is just an instance of the password authentication protocol. The details of RADIUS are within the SSH layer. I don't think it is a good idea to expose this outside of SSH. Resolution: If possible, the details of RADIUS should not be exposed in SSHSM. There may be an issue with receiving authorization without exposing the details. #20: How do we get the mapping from model-specific identity to a model independent securityName?. Resolution: Implementation- dependent, both in the case of extracting tmSecurityname from SSH for an incoming message, and for providing an LCD mapping. #29: do we need to support reports? For what purpose? Yes, reports are used from application processing and for contextEngine discovery. #30: If we actually do not extract anything from securityParameters, do we need to check whether this field parses correctly? It apparently parsed well enough to pass the parse test in the messaging model. Could we simply ignore the securityParameters being passed in? The only argument I see for checking to ensure this is empty is to ensure somebody isn't using the filed for non-standard purposes, such as passing a virus in the field. If we do check it, do we need to report it through Reports? Resolution: yes; it won't hurt to check it. #32: For an incoming message (Processing an Incoming Message section 10), is using a default securityName mapping the right thing to do? Harrington & Salowey Expires September 5, 2006 [Page 52] Internet-Draft Secure Shell Security Model for SNMP March 2006 Resolution: Yes, it is the right thing to do. Appendix B. Change Log "From -00- to -01-" -00- initial draft as ISMS work product: updated references to SecSH RFCs Modified text related to issues# 1, 2, 8, 11, 13, 14, 16, 18, 19, 20, 29, 30, and 32. updated security considerations removed Juergen Schoenwaelder from authors, at his request ran the mib module through smilint "From -01- to -02-" Added TransportDomain and Address Authors' Addresses David Harrington Futurewei Technologies 1700 Alma Dr. Suite 100 Plano, TX 75075 USA Phone: +1 603 436 8634 EMail: dharrington@huawei.com Joseph Salowey Cisco Systems 2901 3rd Ave Seattle, WA 98121 USA EMail: jsalowey@cisco.com Full Copyright Statement Copyright (C) The Internet Society (2006). This document is subject to the rights, licenses and restrictions contained in BCP 78, and except as set forth therein, the authors retain all their rights. This document and the information contained herein are provided on an Harrington & Salowey Expires September 5, 2006 [Page 53] Internet-Draft Secure Shell Security Model for SNMP March 2006 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Intellectual Property The IETF takes no position regarding the validity or scope of any Intellectual Property Rights or other rights that might be claimed to pertain to the implementation or use of the technology described in this document or the extent to which any license under such rights might or might not be available; nor does it represent that it has made any independent effort to identify any such rights. Information on the procedures with respect to rights in RFC documents can be found in BCP 78 and BCP 79. Copies of IPR disclosures made to the IETF Secretariat and any assurances of licenses to be made available, or the result of an attempt made to obtain a general license or permission for the use of such proprietary rights by implementers or users of this specification can be obtained from the IETF on-line IPR repository at http://www.ietf.org/ipr. The IETF invites any interested party to bring to its attention any copyrights, patents or patent applications, or other proprietary rights that may cover technology that may be required to implement this standard. Please address the information to the IETF at ietf-ipr@ietf.org. Acknowledgement Funding for the RFC Editor function is currently provided by the Internet Society. Harrington & Salowey Expires September 5, 2006 [Page 54]