Request for Comments: 5802
Category: Standards Track
Oryx Mail Systems GmbH
Salted Challenge Response Authentication Mechanism (SCRAM)
SASL and GSS-API Mechanisms
The secure authentication mechanism most widely deployed and used by Internet application protocols is the transmission of clear-text passwords over a channel protected by Transport Layer Security (TLS). There are some significant security concerns with that mechanism, which could be addressed by the use of a challenge response authentication mechanism protected by TLS. Unfortunately, the challenge response mechanisms presently on the standards track all fail to meet requirements necessary for widespread deployment, and have had success only in limited use.
This specification describes a family of Simple Authentication and Security Layer (SASL; RFC 4422) authentication mechanisms called the Salted Challenge Response Authentication Mechanism (SCRAM), which addresses the security concerns and meets the deployability requirements. When used in combination with TLS or an equivalent security layer, a mechanism from this family could improve the status quo for application protocol authentication and provide a suitable choice for a mandatory-to-implement mechanism for future application protocol standards.
Status of This Memo
This is an Internet Standards Track document.
This document is a product of the Internet Engineering Task Force (IETF). It represents the consensus of the IETF community. It has received public review and has been approved for publication by the Internet Engineering Steering Group (IESG). Further information on Internet Standards is available in Section 2 of RFC 5741.
Information about the current status of this document, any errata, and how to provide feedback on it may be obtained at http://www.rfc-editor.org/info/rfc5802.
Copyright © 2010 IETF Trust and the persons identified as the document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.
Table of Contents
1. Introduction ....................................................4 2. Conventions Used in This Document ...............................5 2.1. Terminology ................................................5 2.2. Notation ...................................................6 3. SCRAM Algorithm Overview ........................................7 4. SCRAM Mechanism Names ...........................................8 5. SCRAM Authentication Exchange ...................................9 5.1. SCRAM Attributes ..........................................10 5.2. Compliance with SASL Mechanism Requirements ...............13 6. Channel Binding ................................................14 6.1. Default Channel Binding ...................................15 7. Formal Syntax ..................................................15 8. SCRAM as a GSS-API Mechanism ...................................19 8.1. GSS-API Principal Name Types for SCRAM ....................19 8.2. GSS-API Per-Message Tokens for SCRAM ......................20 8.3. GSS_Pseudo_random() for SCRAM .............................20 9. Security Considerations ........................................20 10. IANA Considerations ...........................................22 11. Acknowledgements ..............................................23 12. References ....................................................24 12.1. Normative References .....................................24 12.2. Normative References for GSS-API Implementors ............24 12.3. Informative References ...................................25 Appendix A. Other Authentication Mechanisms .......................27 Appendix B. Design Motivations ....................................27
This specification describes a family of authentication mechanisms called the Salted Challenge Response Authentication Mechanism (SCRAM) which addresses the requirements necessary to deploy a challenge- response mechanism more widely than past attempts (see Appendix A and Appendix B). When used in combination with Transport Layer Security (TLS; see [RFC5246]) or an equivalent security layer, a mechanism from this family could improve the status quo for application protocol authentication and provide a suitable choice for a mandatory-to-implement mechanism for future application protocol standards.
For simplicity, this family of mechanisms does not presently include negotiation of a security layer [RFC4422]. It is intended to be used with an external security layer such as that provided by TLS or SSH, with optional channel binding [RFC5056] to the external security layer.
SCRAM is specified herein as a pure Simple Authentication and Security Layer (SASL) [RFC4422] mechanism, but it conforms to the new bridge between SASL and the Generic Security Service Application Program Interface (GSS-API) called "GS2" [RFC5801]. This means that this document defines both, a SASL mechanism and a GSS-API mechanism.
SCRAM provides the following protocol features:
- The authentication information stored in the authentication database is not sufficient by itself to impersonate the client. The information is salted to prevent a pre-stored dictionary attack if the database is stolen.
- The server does not gain the ability to impersonate the client to other servers (with an exception for server-authorized proxies).
- The mechanism permits the use of a server-authorized proxy without requiring that proxy to have super-user rights with the back-end server.
- Mutual authentication is supported, but only the client is named (i.e., the server has no name).
- When used as a SASL mechanism, SCRAM is capable of transporting authorization identities (see [RFC4422], Section 2) from the client to the server.
For an in-depth discussion of why other challenge response mechanisms are not considered sufficient, see Appendix A. For more information about the motivations behind the design of this mechanism, see Appendix B.
2. Conventions Used in This Document
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].
Example lines prefaced by "C:" are sent by the client and ones prefaced by "S:" by the server. If a single "C:" or "S:" label applies to multiple lines, then the line breaks between those lines are for editorial clarity only, and are not part of the actual protocol exchange.
This document uses several terms defined in [RFC4949] ("Internet Security Glossary") including the following: authentication, authentication exchange, authentication information, brute force, challenge-response, cryptographic hash function, dictionary attack, eavesdropping, hash result, keyed hash, man-in-the-middle, nonce, one-way encryption function, password, replay attack, and salt. Readers not familiar with these terms should use that glossary as a reference.
Some clarifications and additional definitions follow:
- Authentication information: Information used to verify an identity claimed by a SCRAM client. The authentication information for a SCRAM identity consists of salt, iteration count, "StoredKey" and "ServerKey" (as defined in the algorithm overview) for each supported cryptographic hash function.
- Authentication database: The database used to look up the authentication information associated with a particular identity. For application protocols, LDAPv3 (see [RFC4510]) is frequently used as the authentication database. For network-level protocols such as PPP or 802.11x, the use of RADIUS [RFC2865] is more common.
- Base64: An encoding mechanism defined in [RFC4648] that converts an octet string input to a textual output string that can be easily displayed to a human. The use of base64 in SCRAM is restricted to the canonical form with no whitespace.
- Octet: An 8-bit byte.
- Octet string: A sequence of 8-bit bytes.
- Salt: A random octet string that is combined with a password before applying a one-way encryption function. This value is used to protect passwords that are stored in an authentication database.
The pseudocode description of the algorithm uses the following notations:
- ":=": The variable on the left-hand side represents the octet string resulting from the expression on the right-hand side.
- "+": Octet string concatenation.
- "[ ]": A portion of an expression enclosed in "[" and "]" may not be included in the result under some circumstances. See the associated text for a description of those circumstances.
o Normalize(str): Apply the SASLprep profile [RFC4013] of the "stringprep" algorithm [RFC3454] as the normalization algorithm to a UTF-8 [RFC3629] encoded "str". The resulting string is also in UTF-8. When applying SASLprep, "str" is treated as a "stored strings", which means that unassigned Unicode codepoints are prohibited (see Section 7 of [RFC3454]). Note that implementations MUST either implement SASLprep or disallow use of non US-ASCII Unicode codepoints in "str".
- HMAC(key, str): Apply the HMAC keyed hash algorithm (defined in [RFC2104]) using the octet string represented by "key" as the key and the octet string "str" as the input string. The size of the result is the hash result size for the hash function in use. For example, it is 20 octets for SHA-1 (see [RFC3174]).
- H(str): Apply the cryptographic hash function to the octet string "str", producing an octet string as a result. The size of the result depends on the hash result size for the hash function in use.
- XOR: Apply the exclusive-or operation to combine the octet string on the left of this operator with the octet string on the right of this operator. The length of the output and each of the two inputs will be the same for this use.
- Hi(str, salt, i):
U1 := HMAC(str, salt + INT(1)) U2 := HMAC(str, U1) ... Ui-1 := HMAC(str, Ui-2) Ui := HMAC(str, Ui-1) Hi := U1 XOR U2 XOR ... XOR Ui
where "i" is the iteration count, "+" is the string concatenation operator, and INT(g) is a 4-octet encoding of the integer g, most significant octet first.
Hi() is, essentially, PBKDF2 [RFC2898] with HMAC() as the pseudorandom function (PRF) and with dkLen == output length of HMAC() == output length of H().
3. SCRAM Algorithm Overview
The following is a description of a full, uncompressed SASL SCRAM authentication exchange. Nothing in SCRAM prevents either sending the client-first message with the SASL authentication request defined by an application protocol ("initial client response"), or sending the server-final message as additional data of the SASL outcome of authentication exchange defined by an application protocol. See [RFC4422] for more details.
Note that this section omits some details, such as client and server nonces. See Section 5 for more details.
To begin with, the SCRAM client is in possession of a username and password (*) (or a ClientKey/ServerKey, or SaltedPassword). It sends the username to the server, which retrieves the corresponding authentication information, i.e., a salt, StoredKey, ServerKey, and the iteration count i. (Note that a server implementation may choose to use the same iteration count for all accounts.) The server sends the salt and the iteration count to the client, which then computes the following values and sends a ClientProof to the server:
(*) Note that both the username and the password MUST be encoded in UTF-8 [RFC3629].
Informative Note: Implementors are encouraged to create test cases that use both usernames and passwords with non-ASCII codepoints. In particular, it's useful to test codepoints whose "Unicode Normalization Form C" and "Unicode Normalization Form KC" are different. Some examples of such codepoints include Vulgar Fraction One Half (U+00BD) and Acute Accent (U+00B4).
SaltedPassword := Hi(Normalize(password), salt, i) ClientKey := HMAC(SaltedPassword, "Client Key") StoredKey := H(ClientKey) AuthMessage := client-first-message-bare + "," + server-first-message + "," + client-final-message-without-proof ClientSignature := HMAC(StoredKey, AuthMessage) ClientProof := ClientKey XOR ClientSignature ServerKey := HMAC(SaltedPassword, "Server Key") ServerSignature := HMAC(ServerKey, AuthMessage)
The server authenticates the client by computing the ClientSignature, exclusive-ORing that with the ClientProof to recover the ClientKey and verifying the correctness of the ClientKey by applying the hash function and comparing the result to the StoredKey. If the ClientKey is correct, this proves that the client has access to the user's password.
Similarly, the client authenticates the server by computing the ServerSignature and comparing it to the value sent by the server. If the two are equal, it proves that the server had access to the user's ServerKey.
The AuthMessage is computed by concatenating messages from the authentication exchange. The format of these messages is defined in Section 7.
4. SCRAM Mechanism Names
A SCRAM mechanism name is a string "SCRAM-" followed by the uppercased name of the underlying hash function taken from the IANA "Hash Function Textual Names" registry (see http://www.iana.org), optionally followed by the suffix "-PLUS" (see below). Note that SASL mechanism names are limited to 20 octets, which means that only hash function names with lengths shorter or equal to 9 octets (20-length("SCRAM-")-length("-PLUS") can be used. For cases when the underlying hash function name is longer than 9 octets, an alternative 9-octet (or shorter) name can be used to construct the corresponding SCRAM mechanism name, as long as this alternative name doesn't conflict with any other hash function name from the IANA "Hash Function Textual Names" registry. In order to prevent future conflict, such alternative names SHOULD be registered in the IANA "Hash Function Textual Names" registry.
For interoperability, all SCRAM clients and servers MUST implement the SCRAM-SHA-1 authentication mechanism, i.e., an authentication mechanism from the SCRAM family that uses the SHA-1 hash function as defined in [RFC3174].
The "-PLUS" suffix is used only when the server supports channel binding to the external channel. If the server supports channel binding, it will advertise both the "bare" and "plus" versions of whatever mechanisms it supports (e.g., if the server supports only SCRAM with SHA-1, then it will advertise support for both SCRAM-SHA-1 and SCRAM-SHA-1-PLUS). If the server does not support channel binding, then it will advertise only the "bare" version of the mechanism (e.g., only SCRAM-SHA-1). The "-PLUS" exists to allow negotiation of the use of channel binding. See Section 6.
5. SCRAM Authentication Exchange
SCRAM is a SASL mechanism whose client response and server challenge messages are text-based messages containing one or more attribute- value pairs separated by commas. Each attribute has a one-letter name. The messages and their attributes are described in Section 5.1, and defined in Section 7.
SCRAM is a client-first SASL mechanism (see [RFC4422], Section 5, item 2a), and returns additional data together with a server's indication of a successful outcome.
This is a simple example of a SCRAM-SHA-1 authentication exchange when the client doesn't support channel bindings (username 'user' and password 'pencil' are used):
C: n,,n=user,r=fyko+d2lbbFgONRv9qkxdawL S: r=fyko+d2lbbFgONRv9qkxdawL3rfcNHYJY1ZVvWVs7j,s=QSXCR+Q6sek8bf92, i=4096 C: c=biws,r=fyko+d2lbbFgONRv9qkxdawL3rfcNHYJY1ZVvWVs7j, p=v0X8v3Bz2T0CJGbJQyF0X+HI4Ts= S: v=rmF9pqV8S7suAoZWja4dJRkFsKQ=
First, the client sends the "client-first-message" containing:
- a GS2 header consisting of a flag indicating whether channel binding is supported-but-not-used, not supported, or used, and an optional SASL authorization identity;
- SCRAM username and a random, unique nonce attributes.
Note that the client's first message will always start with "n", "y", or "p"; otherwise, the message is invalid and authentication MUST fail. This is important, as it allows for GS2 extensibility (e.g., to add support for security layers).
In response, the server sends a "server-first-message" containing the user's iteration count i and the user's salt, and appends its own nonce to the client-specified one.
The client then responds by sending a "client-final-message" with the same nonce and a ClientProof computed using the selected hash function as explained earlier.
The server verifies the nonce and the proof, verifies that the authorization identity (if supplied by the client in the first message) is authorized to act as the authentication identity, and, finally, it responds with a "server-final-message", concluding the authentication exchange.
The client then authenticates the server by computing the ServerSignature and comparing it to the value sent by the server. If the two are different, the client MUST consider the authentication exchange to be unsuccessful, and it might have to drop the connection.
5.1. SCRAM Attributes
This section describes the permissible attributes, their use, and the format of their values. All attribute names are single US-ASCII letters and are case-sensitive.
Note that the order of attributes in client or server messages is fixed, with the exception of extension attributes (described by the "extensions" ABNF production), which can appear in any order in the designated positions. See Section 7 for authoritative reference.
- a: This is an optional attribute, and is part of the GS2 [RFC5801] bridge between the GSS-API and SASL. This attribute specifies an authorization identity. A client may include it in its first message to the server if it wants to authenticate as one user, but subsequently act as a different user. This is typically used by an administrator to perform some management task on behalf of another user, or by a proxy in some situations.
Upon the receipt of this value the server verifies its
correctness according to the used SASL protocol profile.
Failed verification results in failed authentication exchange.
If this attribute is omitted (as it normally would be), the
authorization identity is assumed to be derived from the
username specified with the (required) "n" attribute.
The server always authenticates the user specified by the "n" attribute. If the "a" attribute specifies a different user, the server associates that identity with the connection after successful authentication and authorization checks.
The syntax of this field is the same as that of the "n" field with respect to quoting of '=' and ','.
- n: This attribute specifies the name of the user whose password is used for authentication (a.k.a. "authentication identity" [RFC4422]). A client MUST include it in its first message to the server. If the "a" attribute is not specified (which would normally be the case), this username is also the identity that will be associated with the connection subsequent to authentication and authorization.
Before sending the username to the server, the client SHOULD prepare the username using the "SASLprep" profile [RFC4013] of the "stringprep" algorithm [RFC3454] treating it as a query string (i.e., unassigned Unicode code points are allowed). If the preparation of the username fails or results in an empty string, the client SHOULD abort the authentication exchange (*).
(*) An interactive client can request a repeated entry of the username value.
Upon receipt of the username by the server, the server MUST either prepare it using the "SASLprep" profile [RFC4013] of the "stringprep" algorithm [RFC3454] treating it as a query string (i.e., unassigned Unicode codepoints are allowed) or otherwise be prepared to do SASLprep-aware string comparisons and/or index lookups. If the preparation of the username fails or results in an empty string, the server SHOULD abort the authentication exchange. Whether or not the server prepares the username using "SASLprep", it MUST use it as received in hash calculations.
The characters ',' or '=' in usernames are sent as '=2C' and '=3D' respectively. If the server receives a username that contains '=' not followed by either '2C' or '3D', then the server MUST fail the authentication.
- m: This attribute is reserved for future extensibility. In this version of SCRAM, its presence in a client or a server message MUST cause authentication failure when the attribute is parsed by the other end.
- r: This attribute specifies a sequence of random printable ASCII characters excluding ',' (which forms the nonce used as input to the hash function). No quoting is applied to this string. As described earlier, the client supplies an initial value in its first message, and the server augments that value with its own nonce in its first response. It is important that this value be different for each authentication (see [RFC4086] for more details on how to achieve this). The client MUST verify that the initial part of the nonce used in subsequent messages is the same as the nonce it initially specified. The server MUST verify that the nonce sent by the client in the second message is the same as the one sent by the server in its first message.
- c: This REQUIRED attribute specifies the base64-encoded GS2 header and channel binding data. It is sent by the client in its second authentication message. The attribute data consist of:
- the GS2 header from the client's first message (recall that the GS2 header contains a channel binding flag and an optional authzid). This header is going to include channel binding type prefix (see [RFC5056]), if and only if the client is using channel binding;
- followed by the external channel's channel binding data, if and only if the client is using channel binding.
- s: This attribute specifies the base64-encoded salt used by the server for this user. It is sent by the server in its first message to the client.
- i: This attribute specifies an iteration count for the selected hash function and user, and MUST be sent by the server along with the user's salt.
For the SCRAM-SHA-1/SCRAM-SHA-1-PLUS SASL mechanism, servers SHOULD announce a hash iteration-count of at least 4096. Note that a client implementation MAY cache ClientKey&ServerKey (or just SaltedPassword) for later reauthentication to the same service, as it is likely that the server is going to advertise the same salt value upon reauthentication. This might be useful for mobile clients where CPU usage is a concern.
- p: This attribute specifies a base64-encoded ClientProof. The client computes this value as described in the overview and sends it to the server.
- v: This attribute specifies a base64-encoded ServerSignature. It is sent by the server in its final message, and is used by the client to verify that the server has access to the user's authentication information. This value is computed as explained in the overview.
- e: This attribute specifies an error that occurred during authentication exchange. It is sent by the server in its final message and can help diagnose the reason for the authentication exchange failure. On failed authentication, the entire server- final-message is OPTIONAL; specifically, a server implementation MAY conclude the SASL exchange with a failure without sending the server-final-message. This results in an application-level error response without an extra round-trip. If the server-final-message is sent on authentication failure, then the "e" attribute MUST be included.
- As-yet unspecified mandatory and optional extensions. Mandatory extensions are encoded as values of the 'm' attribute (see ABNF for reserved-mext in section 7). Optional extensions use as-yet unassigned attribute names.
Mandatory extensions sent by one peer but not understood by the other MUST cause authentication failure (the server SHOULD send the "extensions-not-supported" server-error-value).
Unknown optional extensions MUST be ignored upon receipt.
5.2. Compliance with SASL Mechanism Requirements
This section describes compliance with SASL mechanism requirements specified in Section 5 of [RFC4422].
- "SCRAM-SHA-1" and "SCRAM-SHA-1-PLUS".
2a) SCRAM is a client-first mechanism.
2b) SCRAM sends additional data with success.
- SCRAM is capable of transferring authorization identities from the client to the server.
- SCRAM does not offer any security layers (SCRAM offers channel binding instead).
- SCRAM has a hash protecting the authorization identity.
6. Channel Binding
SCRAM supports channel binding to external secure channels, such as TLS. Clients and servers may or may not support channel binding, therefore the use of channel binding is negotiable. SCRAM does not provide security layers, however, therefore it is imperative that SCRAM provide integrity protection for the negotiation of channel binding.
Use of channel binding is negotiated as follows:
- Servers that support the use of channel binding SHOULD advertise both the non-PLUS (SCRAM-<hash-function>) and PLUS-variant (SCRAM- <hash-function>-PLUS) mechanism name. If the server cannot support channel binding, it SHOULD advertise only the non-PLUS- variant. If the server would never succeed in the authentication of the non-PLUS-variant due to policy reasons, it MUST advertise only the PLUS-variant.
- If the client supports channel binding and the server does not appear to (i.e., the client did not see the -PLUS name advertised by the server), then the client MUST NOT use an "n" gs2-cbind- flag.
- Clients that support mechanism negotiation and channel binding MUST use a "p" gs2-cbind-flag when the server offers the PLUS- variant of the desired GS2 mechanism.
- If the client does not support channel binding, then it MUST use an "n" gs2-cbind-flag. Conversely, if the client requires the use of channel binding then it MUST use a "p" gs2-cbind-flag. Clients that do not support mechanism negotiation never use a "y" gs2- cbind-flag, they use either "p" or "n" according to whether they require and support the use of channel binding or whether they do not, respectively.
- Upon receipt of the client-first message, the server checks the channel binding flag (gs2-cbind-flag).
- If the flag is set to "y" and the server supports channel binding, the server MUST fail authentication. This is because if the client sets the channel binding flag to "y", then the client must have believed that the server did not support channel binding -- if the server did in fact support channel binding, then this is an indication that there has been a downgrade attack (e.g., an attacker changed the server's mechanism list to exclude the -PLUS suffixed SCRAM mechanism name(s)).
- If the channel binding flag was "p" and the server does not support the indicated channel binding type, then the server MUST fail authentication.
The server MUST always validate the client's "c=" field. The server does this by constructing the value of the "c=" attribute and then checking that it matches the client's c= attribute value.
For more discussions of channel bindings, and the syntax of channel binding data for various security protocols, see [RFC5056].
6.1. Default Channel Binding
A default channel binding type agreement process for all SASL application protocols that do not provide their own channel binding type agreement is provided as follows.
'tls-unique' is the default channel binding type for any application that doesn't specify one.
Servers MUST implement the "tls-unique" [RFC5929] channel binding type, if they implement any channel binding. Clients SHOULD implement the "tls-unique" [RFC5929] channel binding type, if they implement any channel binding. Clients and servers SHOULD choose the highest-layer/innermost end-to-end TLS channel as the channel to which to bind.
Servers MUST choose the channel binding type indicated by the client, or fail authentication if they don't support it.
7. Formal Syntax
ALPHA = <as defined in RFC 5234 appendix B.1>
DIGIT = <as defined in RFC 5234 appendix B.1>
UTF8-2 = <as defined in RFC 3629 (STD 63)>
UTF8-3 = <as defined in RFC 3629 (STD 63)>
UTF8-4 = <as defined in RFC 3629 (STD 63)>
attr-val = ALPHA "=" value ;; Generic syntax of any attribute sent ;; by server or client value = 1*value-char value-safe-char = %x01-2B / %x2D-3C / %x3E-7F / UTF8-2 / UTF8-3 / UTF8-4 ;; UTF8-char except NUL, "=", and ",". value-char = value-safe-char / "=" printable = %x21-2B / %x2D-7E ;; Printable ASCII except ",". ;; Note that any "printable" is also ;; a valid "value". base64-char = ALPHA / DIGIT / "/" / "+" base64-4 = 4base64-char base64-3 = 3base64-char "=" base64-2 = 2base64-char "==" base64 = *base64-4 [base64-3 / base64-2]
posit-number = %x31-39 *DIGIT
;; A positive number.
saslname = 1*(value-safe-char / "=2C" / "=3D") ;; Conforms to <value>. authzid = "a=" saslname ;; Protocol specific. cb-name = 1*(ALPHA / DIGIT / "." / "-") ;; See RFC 5056, Section 7. ;; E.g., "tls-server-end-point" or ;; "tls-unique". gs2-cbind-flag = ("p=" cb-name) / "n" / "y" ;; "n" -> client doesn't support channel binding. ;; "y" -> client does support channel binding ;; but thinks the server does not. ;; "p" -> client requires channel binding. ;; The selected channel binding follows "p=". gs2-header = gs2-cbind-flag "," [ authzid ] "," ;; GS2 header for SCRAM ;; (the actual GS2 header includes an optional ;; flag to indicate that the GSS mechanism is not ;; "standard", but since SCRAM is "standard", we ;; don't include that flag). username = "n=" saslname ;; Usernames are prepared using SASLprep. reserved-mext = "m=" 1*(value-char) ;; Reserved for signaling mandatory extensions. ;; The exact syntax will be defined in ;; the future.
channel-binding = "c=" base64
;; base64 encoding of cbind-input.
proof = "p=" base64 nonce = "r=" c-nonce [s-nonce] ;; Second part provided by server. c-nonce = printable s-nonce = printable salt = "s=" base64 verifier = "v=" base64 ;; base-64 encoded ServerSignature.
iteration-count = "i=" posit-number
;; A positive number.
[reserved-mext ","] username "," nonce ["," extensions]
[reserved-mext ","] nonce "," salt "," iteration-count ["," extensions] client-final-message-without-proof = channel-binding "," nonce ["," extensions]
client-final-message-without-proof "," proof
server-error = "e=" server-error-value
server-error-value = "invalid-encoding" /
"extensions-not-supported" / ; unrecognized 'm' value "invalid-proof" / "channel-bindings-dont-match" / "server-does-support-channel-binding" / ; server does not support channel binding "channel-binding-not-supported" / "unsupported-channel-binding-type" / "unknown-user" / "invalid-username-encoding" / ; invalid username encoding (invalid UTF-8 or ; SASLprep failed) "no-resources" / "other-error" / server-error-value-ext ; Unrecognized errors should be treated as "other-error". ; In order to prevent information disclosure, the server ; may substitute the real reason with "other-error".
server-error-value-ext = value
; Additional error reasons added by extensions
; to this document.
server-final-message = (server-error / verifier) ["," extensions] extensions = attr-val *("," attr-val) ;; All extensions are optional, ;; i.e., unrecognized attributes ;; not defined in this document ;; MUST be ignored. cbind-data = 1*OCTET cbind-input = gs2-header [ cbind-data ] ;; cbind-data MUST be present for ;; gs2-cbind-flag of "p" and MUST be absent ;; for "y" or "n".
8. SCRAM as a GSS-API Mechanism
This section and its sub-sections and all normative references of it not referenced elsewhere in this document are INFORMATIONAL for SASL implementors, but they are NORMATIVE for GSS-API implementors.
SCRAM is actually also a GSS-API mechanism. The messages are the same, but a) the GS2 header on the client's first message and channel binding data is excluded when SCRAM is used as a GSS-API mechanism, and b) the RFC2743 section 3.1 initial context token header is prefixed to the client's first authentication message (context token).
The GSS-API mechanism OID for SCRAM-SHA-1 is 22.214.171.124.5.5.14 (see Section 10).
SCRAM security contexts always have the mutual_state flag (GSS_C_MUTUAL_FLAG) set to TRUE. SCRAM does not support credential delegation, therefore SCRAM security contexts alway have the deleg_state flag (GSS_C_DELEG_FLAG) set to FALSE.
8.1. GSS-API Principal Name Types for SCRAM
SCRAM does not explicitly name acceptor principals. However, the use of acceptor principal names to find or prompt for passwords is useful. Therefore, SCRAM supports standard generic name syntaxes for acceptors such as GSS_C_NT_HOSTBASED_SERVICE (see [RFC2743], Section 4.1). Implementations should use the target name passed to GSS_Init_sec_context(), if any, to help retrieve or prompt for SCRAM passwords.
SCRAM supports only a single name type for initiators: GSS_C_NT_USER_NAME. GSS_C_NT_USER_NAME is the default name type for SCRAM.
There is no name canonicalization procedure for SCRAM beyond applying SASLprep as described in Section 5.1.
The query, display, and exported name syntaxes for SCRAM principal names are all the same. There are no SCRAM-specific name syntaxes (SCRAM initiator principal names are free-form); -- applications should use generic GSS-API name types such as GSS_C_NT_USER_NAME and
8.2. GSS-API Per-Message Tokens for SCRAM
The per-message tokens for SCRAM as a GSS-API mechanism SHALL be the same as those for the Kerberos V GSS-API mechanism [RFC4121] (see Section 4.2 and sub-sections), using the Kerberos V "aes128-cts-hmac- sha1-96" enctype [RFC3962].
The replay_det_state (GSS_C_REPLAY_FLAG), sequence_state (GSS_C_SEQUENCE_FLAG), conf_avail (GSS_C_CONF_FLAG) and integ_avail (GSS_C_CONF_FLAG) security context flags are always set to TRUE.
The 128-bit session "protocol key" SHALL be derived by using the least significant (right-most) 128 bits of HMAC(StoredKey, "GSS-API session key" || ClientKey || AuthMessage). "Specific keys" are then derived as usual as described in Section 2 of [RFC4121], [RFC3961], and [RFC3962].
The terms "protocol key" and "specific key" are Kerberos V5 terms [RFC3961].
SCRAM does support PROT_READY, and is PROT_READY on the initiator side first upon receipt of the server's reply to the initial security context token.
8.3. GSS_Pseudo_random() for SCRAM
The GSS_Pseudo_random() [RFC4401] for SCRAM SHALL be the same as for the Kerberos V GSS-API mechanism [RFC4402]. There is no acceptor- asserted sub-session key for SCRAM, thus GSS_C_PRF_KEY_FULL and GSS_C_PRF_KEY_PARTIAL are equivalent for SCRAM's GSS_Pseudo_random(). The protocol key to be used for the GSS_Pseudo_random() SHALL be the same as the key defined in Section 8.2.
9. Security Considerations
If the authentication exchange is performed without a strong security layer (such as TLS with data confidentiality), then a passive eavesdropper can gain sufficient information to mount an offline dictionary or brute-force attack that can be used to recover the user's password. The amount of time necessary for this attack depends on the cryptographic hash function selected, the strength of the password, and the iteration count supplied by the server. An external security layer with strong encryption will prevent this attack.
If the external security layer used to protect the SCRAM exchange uses an anonymous key exchange, then the SCRAM channel binding mechanism can be used to detect a man-in-the-middle attack on the security layer and cause the authentication to fail as a result. However, the man-in-the-middle attacker will have gained sufficient information to mount an offline dictionary or brute-force attack. For this reason, SCRAM allows to increase the iteration count over time. (Note that a server that is only in possession of "StoredKey" and "ServerKey" can't automatically increase the iteration count upon successful authentication. Such an increase would require resetting the user's password.)
If the authentication information is stolen from the authentication database, then an offline dictionary or brute-force attack can be used to recover the user's password. The use of salt mitigates this attack somewhat by requiring a separate attack on each password. Authentication mechanisms that protect against this attack are available (e.g., the EKE class of mechanisms). RFC 2945 [RFC2945] is an example of such technology. The WG elected not to use EKE like mechanisms as a basis for SCRAM.
If an attacker obtains the authentication information from the authentication repository and either eavesdrops on one authentication exchange or impersonates a server, the attacker gains the ability to impersonate that user to all servers providing SCRAM access using the same hash function, password, iteration count, and salt. For this reason, it is important to use randomly generated salt values.
SCRAM does not negotiate a hash function to use. Hash function negotiation is left to the SASL mechanism negotiation. It is important that clients be able to sort a locally available list of mechanisms by preference so that the client may pick the appropriate mechanism to use from a server's advertised mechanism list. This preference order is not specified here as it is a local matter. The preference order should include objective and subjective notions of mechanism cryptographic strength (e.g., SCRAM with a successor to SHA-1 may be preferred over SCRAM with SHA-1).
Note that to protect the SASL mechanism negotiation applications normally must list the server mechanisms twice: once before and once after authentication, the latter using security layers. Since SCRAM does not provide security layers, the only ways to protect the mechanism negotiation are a) use channel binding to an external channel, or b) use an external channel that authenticates a user- provided server name.
SCRAM does not protect against downgrade attacks of channel binding types. The complexities of negotiating a channel binding type, and handling down-grade attacks in that negotiation, were intentionally left out of scope for this document.
A hostile server can perform a computational denial-of-service attack on clients by sending a big iteration count value.
See [RFC4086] for more information about generating randomness.
10. IANA Considerations
IANA has added the following family of SASL mechanisms to the SASL Mechanism registry established by [RFC4422]:
Subject: Registration of a new SASL family SCRAM
SASL mechanism name (or prefix for the family): SCRAM-* Security considerations: Section 7 of [RFC5802] Published specification (optional, recommended): [RFC5802] Person & email address to contact for further information: IETF SASL WG <firstname.lastname@example.org>
Intended usage: COMMON
Owner/Change controller: IESG <email@example.com> Note: Members of this family MUST be explicitly registered using the "IETF Review" [RFC5226] registration procedure. Reviews MUST be requested on the SASL mailing list <firstname.lastname@example.org> (or a successor designated by the responsible Security AD).
Note to future SCRAM-mechanism designers: each new SCRAM-SASL mechanism MUST be explicitly registered with IANA and MUST comply with SCRAM-mechanism naming convention defined in Section 4 of this document.
IANA has added the following entries to the SASL Mechanism registry established by [RFC4422]:
To: email@example.com Subject: Registration of a new SASL mechanism SCRAM-SHA-1 SASL mechanism name (or prefix for the family): SCRAM-SHA-1 Security considerations: Section 7 of [RFC5802] Published specification (optional, recommended): [RFC5802] Person & email address to contact for further information: IETF SASL WG <firstname.lastname@example.org> Intended usage: COMMON Owner/Change controller: IESG <email@example.com> Note: To: firstname.lastname@example.org Subject: Registration of a new SASL mechanism SCRAM-SHA-1-PLUS SASL mechanism name (or prefix for the family): SCRAM-SHA-1-PLUS Security considerations: Section 7 of [RFC5802] Published specification (optional, recommended): [RFC5802] Person & email address to contact for further information: IETF SASL WG <email@example.com> Intended usage: COMMON Owner/Change controller: IESG <firstname.lastname@example.org> Note: Per this document, IANA has assigned a GSS-API mechanism OID for SCRAM-SHA-1 from the iso.org.dod.internet.security.mechanisms prefix (see "SMI Security for Mechanism Codes" registry).
This document benefited from discussions on the SASL WG mailing list. The authors would like to specially thank Dave Cridland, Simon Josefsson, Jeffrey Hutzelman, Kurt Zeilenga, Pasi Eronen, Ben Campbell, Peter Saint-Andre, and Tobias Markmann for their contributions to this document. A special thank you to Simon Josefsson for shepherding this document and for doing one of the first implementations of this specification.
12.1. Normative References
[RFC2104] Krawczyk, H., Bellare, M., and R. Canetti, "HMAC: Keyed- Hashing for Message Authentication", RFC 2104, February 1997. [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. [RFC3174] Eastlake, D. and P. Jones, "US Secure Hash Algorithm 1 (SHA1)", RFC 3174, September 2001. [RFC3454] Hoffman, P. and M. Blanchet, "Preparation of Internationalized Strings ("stringprep")", RFC 3454, December 2002. [RFC3629] Yergeau, F., "UTF-8, a transformation format of ISO 10646", STD 63, RFC 3629, November 2003. [RFC4013] Zeilenga, K., "SASLprep: Stringprep Profile for User Names and Passwords", RFC 4013, February 2005. [RFC4422] Melnikov, A. and K. Zeilenga, "Simple Authentication and Security Layer (SASL)", RFC 4422, June 2006. [RFC4648] Josefsson, S., "The Base16, Base32, and Base64 Data Encodings", RFC 4648, October 2006. [RFC5056] Williams, N., "On the Use of Channel Bindings to Secure Channels", RFC 5056, November 2007. [RFC5234] Crocker, D. and P. Overell, "Augmented BNF for Syntax Specifications: ABNF", STD 68, RFC 5234, January 2008. [RFC5929] Altman, J., Williams, N., and L. Zhu, "Channel Bindings for TLS", RFC 5929, July 2010.
12.2. Normative References for GSS-API Implementors
[RFC2743] Linn, J., "Generic Security Service Application Program Interface Version 2, Update 1", RFC 2743, January 2000. [RFC3961] Raeburn, K., "Encryption and Checksum Specifications for Kerberos 5", RFC 3961, February 2005. [RFC3962] Raeburn, K., "Advanced Encryption Standard (AES) Encryption for Kerberos 5", RFC 3962, February 2005. [RFC4121] Zhu, L., Jaganathan, K., and S. Hartman, "The Kerberos Version 5 Generic Security Service Application Program Interface (GSS-API) Mechanism: Version 2", RFC 4121, July 2005. [RFC4401] Williams, N., "A Pseudo-Random Function (PRF) API Extension for the Generic Security Service Application Program Interface (GSS-API)", RFC 4401, February 2006. [RFC4402] Williams, N., "A Pseudo-Random Function (PRF) for the Kerberos V Generic Security Service Application Program Interface (GSS-API) Mechanism", RFC 4402, February 2006. [RFC5801] Josefsson, S. and N. Williams, "Using Generic Security Service Application Program Interface (GSS-API) Mechanisms in Simple Authentication and Security Layer (SASL): The GS2 Mechanism Family", RFC 5801, July 2010.
12.3. Informative References
Zeilenga, K., "CRAM-MD5 to Historic", Work in Progress, November 2008.
Melnikov, A., "Moving DIGEST-MD5 to Historic", Work in Progress, July 2008. [RFC2865] Rigney, C., Willens, S., Rubens, A., and W. Simpson, "Remote Authentication Dial In User Service (RADIUS)", RFC 2865, June 2000. [RFC2898] Kaliski, B., "PKCS #5: Password-Based Cryptography Specification Version 2.0", RFC 2898, September 2000. [RFC2945] Wu, T., "The SRP Authentication and Key Exchange System", RFC 2945, September 2000. [RFC4086] Eastlake, D., Schiller, J., and S. Crocker, "Randomness Requirements for Security", BCP 106, RFC 4086, June 2005. [RFC4510] Zeilenga, K., "Lightweight Directory Access Protocol (LDAP): Technical Specification Road Map", RFC 4510, June 2006. [RFC4616] Zeilenga, K., "The PLAIN Simple Authentication and Security Layer (SASL) Mechanism", RFC 4616, August 2006. [RFC4949] Shirey, R., "Internet Security Glossary, Version 2", RFC 4949, August 2007. [RFC5226] Narten, T. and H. Alvestrand, "Guidelines for Writing an IANA Considerations Section in RFCs", BCP 26, RFC 5226, May 2008. [RFC5246] Dierks, T. and E. Rescorla, "The Transport Layer Security (TLS) Protocol Version 1.2", RFC 5246, August 2008. [RFC5803] Melnikov, A., "Lightweight Directory Access Protocol (LDAP) Schema for Storing Salted Challenge Response Authentication Mechanism (SCRAM) Secrets", RFC 5803, July 2010.
IANA, "Registration of TLS server end-point channel bindings", available from http://www.iana.org, June 2008.
Appendix A. Other Authentication Mechanisms
The DIGEST-MD5 [DIGESTHISTORIC] mechanism has proved to be too complex to implement and test, and thus has poor interoperability. The security layer is often not implemented, and almost never used; everyone uses TLS instead. For a more complete list of problems with DIGEST-MD5 that led to the creation of SCRAM, see [DIGESTHISTORIC].
The CRAM-MD5 SASL mechanism, while widely deployed, also has some problems. In particular, it is missing some modern SASL features such as support for internationalized usernames and passwords, support for passing of authorization identity, and support for channel bindings. It also doesn't support server authentication. For a more complete list of problems with CRAM-MD5, see [CRAMHISTORIC].
The PLAIN [RFC4616] SASL mechanism allows a malicious server or eavesdropper to impersonate the authenticating user to any other server for which the user has the same password. It also sends the password in the clear over the network, unless TLS is used. Server authentication is not supported.
Appendix B. Design Motivations
The following design goals shaped this document. Note that some of the goals have changed since the initial version of the document.
- The SASL mechanism has all modern SASL features: support for internationalized usernames and passwords, support for passing of authorization identity, and support for channel bindings.
- The protocol supports mutual authentication.
- The authentication information stored in the authentication database is not sufficient by itself to impersonate the client.
- The server does not gain the ability to impersonate the client to other servers (with an exception for server-authorized proxies), unless such other servers allow SCRAM authentication and use the same salt and iteration count for the user.
- The mechanism is extensible, but (hopefully) not over-engineered in this respect.
- The mechanism is easier to implement than DIGEST-MD5 in both clients and servers.
Chris Newman Oracle 800 Royal Oaks Monrovia, CA 91016 USA EMail: email@example.com
Oryx Mail Systems GmbH
EMail: Alexey.Melnikov@isode.com Nicolas Williams Oracle 5300 Riata Trace Ct Austin, TX 78727 USA EMail: Nicolas.Williams@oracle.com