Network Working Group J. Galbraith Request for Comments: 4819 J. Van Dyke Category: Standards Track VanDyke Software J. Bright Silicon Circus March 2007
Secure Shell Public Key Subsystem
Status of This Memo
This document specifies an Internet standards track protocol for the Internet community, and requests discussion and suggestions for improvements. Please refer to the current edition of the "Internet Official Protocol Standards" (STD 1) for the standardization state and status of this protocol. Distribution of this memo is unlimited.
Copyright Notice
Copyright (C) The IETF Trust (2007).
Abstract
Secure Shell defines a user authentication mechanism that is based on public keys, but does not define any mechanism for key distribution. No common key management solution exists in current implementations. This document describes a protocol that can be used to configure public keys in an implementation-independent fashion, allowing client software to take on the burden of this configuration.
The Public Key Subsystem provides a server-independent mechanism for clients to add public keys, remove public keys, and list the current public keys known by the server. Rights to manage public keys are specific and limited to the authenticated user.
A public key may also be associated with various restrictions, including a mandatory command or subsystem.
Galbraith, et al. Standards Track [Page 1]
RFC 4819 Secure Shell Public Key Subsystem March 2007
Secure Shell (SSH) is a protocol for secure remote login and other secure network services over an insecure network. Secure Shell defines a user authentication mechanism that is based on public keys, but does not define any mechanism for key distribution. Common practice is to authenticate once with password authentication and transfer the public key to the server. However, to date no two implementations use the same mechanism to configure a public key for use.
This document describes a subsystem that can be used to configure public keys in an implementation-independent fashion. This approach allows client software to take on the burden of this configuration. The Public Key Subsystem protocol is designed for extreme simplicity in implementation. It is not intended as a Public Key Infrastructure for X.509 Certificates (PKIX) replacement.
The Secure Shell Public Key Subsystem has been designed to run on top of the Secure Shell transport layer [2] and user authentication protocols [3]. It provides a simple mechanism for the client to manage public keys on the server.
This document should be read only after reading the Secure Shell architecture [1] and Secure Shell connection [4] documents.
This protocol is intended to be used from the Secure Shell Connection Protocol [4] as a subsystem, as described in the section "Starting a Shell or a Command". The subsystem name used with this protocol is "publickey".
This protocol requires that the user be able to authenticate in some fashion before it can be used. If password authentication is used, servers SHOULD provide a configuration option to disable the use of password authentication after the first public key is added.
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 RFC 2119 [5].
The Public Key Subsystem provides a server-independent mechanism for clients to add public keys, remove public keys, and list the current public keys known by the server. The subsystem name is "publickey".
Galbraith, et al. Standards Track [Page 3]
RFC 4819 Secure Shell Public Key Subsystem March 2007
The public keys added, removed, and listed using this protocol are specific and limited to those of the authenticated user.
The operations to add, remove, and list the authenticated user's public keys are performed as request packets sent to the server. The server sends response packets that indicate success or failure as well as provide specific response data.
The format of public key blobs are detailed in Section 6.6, "Public Key Algorithms" of the SSH Transport Protocol document [2].
Client implementations SHOULD reject this request; it is normally sent only by the client.
If want reply is TRUE, the server MUST respond with SSH_MSG_CHANNEL_SUCCESS if the Public Key Subsystem was successfully started, or SSH_MSG_CHANNEL_FAILURE if the server failed to start or does not support the Public Key Subsystem.
The server SHOULD respond with SSH_MSG_CHANNEL_FAILURE if the user is not allowed access to the Public Key Subsystem (for example, because the user authenticated with a restricted public key).
It is RECOMMENDED that clients request and check the reply for this request.
Galbraith, et al. Standards Track [Page 4]
RFC 4819 Secure Shell Public Key Subsystem March 2007
All Public Key Subsystem requests and responses are sent in the following form:
uint32 length string name ... request/response specific data follows
The length field describes the length of the name field and of the request/response-specific data, but does not include the length of the length field itself. The client MUST receive acknowledgement of each request prior to sending a new request.
The version packet, as well as all requests and responses described in Section 4, are a description of the 'name' field and the data part of the packet.
A request is acknowledged by sending a status packet. If there is data in response to the request, the status packet is sent after all data has been sent.
string "status" uint32 status code string description [7] string language tag [6]
A status message MUST be sent for any unrecognized packets, and the request SHOULD NOT close the subsystem.
RFC 4819 Secure Shell Public Key Subsystem March 2007
If a request completed successfully, the server MUST send the status code SSH_PUBLICKEY_SUCCESS. The meaning of the failure codes is as implied by their names.
Both sides MUST start a connection by sending a version packet that indicates the version of the protocol they are using.
string "version" uint32 protocol-version-number
This document describes version 2 of the protocol. Version 1 was used by an early draft of this document. The version number was incremented after changes in the handling of status packets.
Both sides send the highest version that they implement. The lower of the version numbers is the version of the protocol to use. If either side can't support the lower version, it should close the subsystem and notify the other side by sending an SSH_MSG_CHANNEL_CLOSE message. Before closing the subsystem, a status message with the status SSH_PUBLICKEY_VERSION_NOT_SUPPORTED SHOULD be sent. Note that, normally, status messages are only sent by the server (in response to requests from the client). This is the only occasion on which the client sends a status message.
Both sides MUST wait to receive this version before continuing. The "version" packet MUST NOT be sent again after this initial exchange. The SSH_PUBLICKEY_VERSION_NOT_SUPPORTED status code must not be sent in response to any other request.
Implementations MAY use the first 15 bytes of the version packet as a "magic cookie" to avoid processing spurious output from the user's shell (as described in Section 6.5 of [4]). These bytes will always be:
If the client wishes to add a public key, the client sends:
string "add" string public key algorithm name string public key blob boolean overwrite uint32 attribute-count string attrib-name string attrib-value bool critical repeated attribute-count times
The server MUST attempt to store the public key for the user in the appropriate location so the public key can be used for subsequent public key authentications. If the overwrite field is false and the specified key already exists, the server MUST return SSH_PUBLICKEY_KEY_ALREADY_PRESENT. If the server returns this, the client SHOULD provide an option to the user to overwrite the key. If the overwrite field is true and the specified key already exists, but cannot be overwritten, the server MUST return SSH_PUBLICKEY_ACCESS_DENIED.
Attribute names are defined following the same scheme laid out for algorithm names in [1]. If the server does not implement a critical attribute, it MUST fail the add, with the status code SSH_PUBLICKEY_ATTRIBUTE_NOT_SUPPORTED. For the purposes of a critical attribute, mere storage of the attribute is not sufficient -- rather, the server must understand and implement the intent of the attribute.
The following attributes are currently defined:
"comment"
The value of the comment attribute contains user-specified text about the public key. The server SHOULD make every effort to preserve this value and return it with the key during any subsequent list operation. The server MUST NOT attempt to interpret or act upon the content of the comment field in any way. The comment attribute must be specified in UTF-8 format [7].
Galbraith, et al. Standards Track [Page 7]
RFC 4819 Secure Shell Public Key Subsystem March 2007
The comment field is useful so the user can identify the key without resorting to comparing its fingerprint. This attribute SHOULD NOT be critical.
"comment-language"
If this attribute is specified, it MUST immediately follow a "comment" attribute and specify the language for that attribute [6]. The client MAY specify more than one comment if it additionally specifies a different language for each of those comments. The server SHOULD attempt to store each comment with its language attribute. This attribute SHOULD NOT be critical.
"command-override"
"command-override" specifies a command to be executed when this key is in use. The command should be executed by the server when it receives an "exec" or "shell" request from the client, in place of the command or shell which would otherwise have been executed as a result of that request. If the command string is empty, both "exec" and "shell" requests should be denied. If no "command-override" attribute is specified, all "exec" and "shell" requests should be permitted (as long as they satisfy other security or authorization checks the server may perform). This attribute SHOULD be critical.
"subsystem"
"subsystem" specifies a comma-separated list of subsystems that may be started (using a "subsystem" request) when this key is in use. This attribute SHOULD be critical. If the value is empty, no subsystems may be started. If the "subsystem" attribute is not specified, no restrictions are placed on which subsystems may be started when authenticated using this key.
"x11"
"x11" specifies that X11 forwarding may not be performed when this key is in use. The attribute-value field SHOULD be empty for this attribute. This attribute SHOULD be critical.
"shell"
"shell" specifies that session channel "shell" requests should be denied when this key is in use. The attribute-value field SHOULD be empty for this attribute. This attribute SHOULD be critical.
Galbraith, et al. Standards Track [Page 8]
RFC 4819 Secure Shell Public Key Subsystem March 2007
"exec"
"exec" specifies that session channel "exec" requests should be denied when this key is in use. The attribute-value field SHOULD be empty for this attribute. This attribute SHOULD be critical.
"agent"
"agent" specifies that session channel "auth-agent-req" requests should be denied when this key is in use. The attribute-value field SHOULD be empty for this attribute. This attribute SHOULD be critical.
"env"
"env" specifies that session channel "env" requests should be denied when this key is in use. The attribute-value field SHOULD be empty for this attribute. This attribute SHOULD be critical.
"from"
"from" specifies a comma-separated list of hosts from which the key may be used. If a host not in this list attempts to use this key for authorization purposes, the authorization attempt MUST be denied. The server SHOULD make a log entry regarding this. The server MAY provide a method for administrators to disallow the appearance of a host in this list. The server should use whatever method is appropriate for its platform to identify the host -- e.g., for IP- based networks, checking the IP address or performing a reverse DNS lookup. For IP-based networks, it is anticipated that each element of the "from" parameter will take the form of a specific IP address or hostname.
"port-forward"
"port-forward" specifies that no "direct-tcpip" requests should be accepted, except those to hosts specified in the comma-separated list supplied as a value to this attribute. If the value of this attribute is empty, all "direct-tcpip" requests should be refused when using this key. This attribute SHOULD be critical.
"reverse-forward"
"reverse-forward" specifies that no "tcpip-forward" requests should be accepted, except for the port numbers in the comma-separated list supplied as a value to this attribute. If the value of this attribute is empty, all "tcpip-forward" requests should be refused when using this key. This attribute SHOULD be critical.
Galbraith, et al. Standards Track [Page 9]
RFC 4819 Secure Shell Public Key Subsystem March 2007
In addition to the attributes specified by the client, the server MAY provide a method for administrators to enforce certain attributes compulsorily.
If the client wishes to remove a public key, the client sends:
string "remove" string public key algorithm name string public key blob
The server MUST attempt to remove the public key for the user from the appropriate location, so that the public key cannot be used for subsequent authentications.
If the client wishes to list the known public keys, the client sends:
string "list"
The server will respond with zero or more of the following responses:
string "publickey" string public key algorithm name string public key blob uint32 attribute-count string attrib-name string attrib-value repeated attribute-count times
There is no requirement that the responses be in any particular order. Whilst some server implementations may send the responses in some order, client implementations should not rely on responses being in any order.
Following the last "publickey" response, a status packet MUST be sent.
If the client wishes to know which key attributes the server supports, it sends:
string "listattributes"
Galbraith, et al. Standards Track [Page 10]
RFC 4819 Secure Shell Public Key Subsystem March 2007
The server will respond with zero or more of the following responses:
string "attribute" string attribute name boolean compulsory
The "compulsory" field indicates whether this attribute will be compulsorily applied to any added keys (irrespective of whether the attribute has been specified by the client) due to administrative settings on the server. If the server does not support administrative settings of this nature, it MUST return false in the compulsory field. An example of use of the "compulsory" attribute would be a server with a configuration file specifying that the user is not permitted shell access. Given this, the server would return the "shell" attribute, with "compulsory" marked true. Whatever attributes the user subsequently asked the server to apply to their key, the server would also apply the "shell" attribute, rendering it impossible for the user to use a shell.
Following the last "attribute" response, a status packet MUST be sent.
An implementation MAY choose not to support this request.
This protocol assumes that it is run over a secure channel and that the endpoints of the channel have been authenticated. Thus, this protocol assumes that it is externally protected from network-level attacks.
This protocol provides a mechanism that allows client authentication data to be uploaded and manipulated. It is the responsibility of the server implementation to enforce any access controls that may be required to limit the access allowed for any particular user (the user being authenticated externally to this protocol, typically using the SSH User Authentication Protocol [3]). In particular, it is possible for users to overwrite an existing key on the server with this protocol, whilst at the same time specifying fewer restrictions for the new key than were previously present. Servers should take care that when doing this, clients are not able to override presets from the server's administrator.
This protocol requires the client to assume that the server will correctly implement and observe attributes applied to keys. Implementation errors in the server could cause clients to authorize keys for access they were not intended to have, or to apply fewer restrictions than were intended.
Galbraith, et al. Standards Track [Page 11]
RFC 4819 Secure Shell Public Key Subsystem March 2007
In the following sections, the values for the namespaces are textual. The conventions and instructions to the IANA for future assignments are given in this section. The initial assignments are given in their respective sections.
All names registered by the IANA in the following sections MUST be printable US-ASCII strings, and MUST NOT contain the characters at-sign ("@"), comma (","), or whitespace or control characters (ASCII codes 32 or less). Names are case-sensitive, and MUST NOT be longer than 64 characters.
A provision is made here for locally extensible names. The IANA will not register and will not control names with the at-sign in them. Names with the at-sign in them will have the format of "name@domainname" (without the double quotes) where the part preceding the at-sign is the name. The format of the part preceding the at-sign is not specified; however, these names MUST be printable US-ASCII strings, and MUST NOT contain the comma character (","), or whitespace, or control characters (ASCII codes 32 or less). The part following the at-sign MUST be a valid, fully qualified Internet domain name [10] controlled by the person or organization defining the name. Names are case-sensitive, and MUST NOT be longer than 64 characters. It is up to each domain how it manages its local namespace. It has been noted that these names resemble STD 11 [9] email addresses. This is purely coincidental and actually has nothing to do with STD 11 [9]. An example of a locally defined name is "our-attribute@example.com" (without the double quotes).
Galbraith, et al. Standards Track [Page 12]
RFC 4819 Secure Shell Public Key Subsystem March 2007
Attributes are used to define properties or restrictions for public keys. The following table lists the initial assignments of Public Key Subsystem Attribute names.
Galbraith, et al. Standards Track [Page 13]
RFC 4819 Secure Shell Public Key Subsystem March 2007
Attribute Name --------------- comment comment-language command-override subsystem x11 shell exec agent env from port-forward reverse-forward
Status responses have status codes in the range 0 to 255. These numbers are allocated as follows. Of these, the range 192 to 255 is reserved for use by local, private extensions.
RFC 4819 Secure Shell Public Key Subsystem March 2007
Full Copyright Statement
Copyright (C) The IETF Trust (2007).
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 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST 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.