Network Working Group E. Rescorla
Request for Comments: 2660 RTFM, Inc.
Category: Experimental A. Schiffman
Terisa Systems, Inc.
August 1999
The Secure HyperText Transfer Protocol
Status of this Memo
This memo defines an Experimental Protocol for the Internet
community. It does not specify an Internet standard of any kind.
Discussion and suggestions for improvement are requested.
Distribution of this memo is unlimited.
Copyright Notice
Copyright (C) The Internet Society (1999). All Rights Reserved.
Abstract
This memo describes a syntax for securing messages sent using the
Hypertext Transfer Protocol (HTTP), which forms the basis for the
World Wide Web. Secure HTTP (S-HTTP) provides independently
applicable security services for transaction confidentiality,
authenticity/integrity and non-repudiability of origin.
The protocol emphasizes maximum flexibility in choice of key
management mechanisms, security policies and cryptographic algorithms
by supporting option negotiation between parties for each
transaction.
Table of Contents
1. Introduction .................................................. 31.1. Summary of Features ......................................... 31.2. Changes ..................................................... 41.3. Processing Model ............................................ 51.4. Modes of Operation .......................................... 61.5. Implementation Options ...................................... 72. Message Format ................................................ 72.1. Notational Conventions ...................................... 82.2. The Request Line ............................................ 82.3. The Status Line ............................................. 82.4. Secure HTTP Header Lines .................................... 82.5. Content .....................................................122.6. Encapsulation Format Options ................................13
Rescorla & Schiffman Experimental [Page 1]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
2.6.1. Content-Privacy-Domain: CMS ...............................132.6.2. Content-Privacy-Domain: MOSS ..............................142.6.3. Permitted HTTP headers ....................................142.6.3.2. Host ....................................................152.6.3.3. Connection ..............................................153. Cryptographic Parameters ......................................153.1. Options Headers .............................................153.2. Negotiation Options .........................................163.2.1. Negotiation Overview ......................................163.2.2. Negotiation Option Format .................................163.2.3. Parametrization for Variable-length Key Ciphers ...........183.2.4. Negotiation Syntax ........................................183.3. Non-Negotiation Headers .....................................233.3.1. Encryption-Identity .......................................233.3.2. Certificate-Info ..........................................233.3.3. Key-Assign ................................................243.3.4. Nonces ....................................................253.4. Grouping Headers With SHTTP-Cryptopts .......................263.4.1. SHTTP-Cryptopts ...........................................264. New Header Lines for HTTP .....................................264.1. Security-Scheme .............................................265. (Retriable) Server Status Error Reports .......................275.1. Retry for Option (Re)Negotiation ............................275.2. Specific Retry Behavior .....................................285.3. Limitations On Automatic Retries ............................296. Other Issues ..................................................306.1. Compatibility of Servers with Old Clients ...................306.2. URL Protocol Type ...........................................306.3. Browser Presentation ........................................317. Implementation Notes ..........................................327.1. Preenhanced Data ............................................327.2. Note:Proxy Interaction ......................................347.2.1. Client-Proxy Authentication ...............................348. Implementation Recommendations and Requirements ...............349. Protocol Syntax Summary .......................................3510. An Extended Example ..........................................36
Appendix: A Review of CMS ........................................40
Bibliography and References ......................................41
Security Considerations ..........................................43
Authors' Addresses ...............................................44
Full Copyright Statement..........................................45
Rescorla & Schiffman Experimental [Page 2]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
The World Wide Web (WWW) is a distributed hypermedia system which has
gained widespread acceptance among Internet users. Although WWW
browsers support other, preexisting Internet application protocols,
the native and primary protocol used between WWW clients and servers
is the HyperText Transfer Protocol (HTTP) [RFC-2616]. The ease of
use of the Web has prompted its widespread employment as a
client/server architecture for many applications. Many such
applications require the client and server to be able to authenticate
each other and exchange sensitive information confidentially. The
original HTTP specification had only modest support for the
cryptographic mechanisms appropriate for such transactions.
Secure HTTP (S-HTTP) provides secure communication mechanisms between
an HTTP client-server pair in order to enable spontaneous commercial
transactions for a wide range of applications. Our design intent is
to provide a flexible protocol that supports multiple orthogonal
operation modes, key management mechanisms, trust models,
cryptographic algorithms and encapsulation formats through option
negotiation between parties for each transaction.
Secure HTTP is a secure message-oriented communications protocol
designed for use in conjunction with HTTP. It is designed to coexist
with HTTP's messaging model and to be easily integrated with HTTP
applications.
Secure HTTP provides a variety of security mechanisms to HTTP clients
and servers, providing the security service options appropriate to
the wide range of potential end uses possible for the World-Wide Web.
The protocol provides symmetric capabilities to both client and
server (in that equal treatment is given to both requests and
replies, as well as for the preferences of both parties) while
preserving the transaction model and implementation characteristics
of HTTP.
Several cryptographic message format standards may be incorporated
into S-HTTP clients and servers, particularly, but in principle not
limited to, [CMS] and [MOSS]. S-HTTP supports interoperation among a
variety of implementations, and is compatible with HTTP. S-HTTP
aware clients can communicate with S-HTTP oblivious servers and
vice-versa, although such transactions obviously would not use S-HTTP
security features.
S-HTTP does not require client-side public key certificates (or
public keys), as it supports symmetric key-only operation modes.
Rescorla & Schiffman Experimental [Page 3]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
This is significant because it means that spontaneous private
transactions can occur without requiring individual users to have
an established public key. While S-HTTP is able to take advantage
of ubiquitous certification infrastructures, its deployment does
not require it.
S-HTTP supports end-to-end secure transactions, in contrast with the
original HTTP authorization mechanisms which require the client to
attempt access and be denied before the security mechanism is
employed. Clients may be "primed" to initiate a secure transaction
(typically using information supplied in message headers); this may
be used to support encryption of fill-out forms, for example. With
S-HTTP, no sensitive data need ever be sent over the network in the
clear.
S-HTTP provides full flexibility of cryptographic algorithms, modes
and parameters. Option negotiation is used to allow clients and
servers to agree on transaction modes (e.g., should the request be
signed or encrypted or both -- similarly for the reply?);
cryptographic algorithms (RSA vs. DSA for signing, DES vs.
RC2 for encrypting, etc.); and certificate selection
(please sign with your "Block-buster Video certificate").
S-HTTP attempts to avoid presuming a particular trust model, although
its designers admit to a conscious effort to facilitate
multiply-rooted hierarchical trust, and anticipate that principals may
have many public key certificates.
S-HTTP differs from Digest-Authentication, described in [RFC-2617] in
that it provides support for public key cryptography and consequently
digital signature capability, as well as providing confidentiality.
This document describes S-HTTP/1.4. It differs from the previous
memo in that it differs from the previous memo in its support of
the Cryptographic Message Syntax (CMS) [CMS], a successor to PKCS-7;
and hence now supports the Diffie-Hellman and the (NIST) Digital
Signature Standard cryptosystems. CMS used in RSA mode is bits on the
wire compatible with PKCS-7.
Rescorla & Schiffman Experimental [Page 4]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
The creation of an S-HTTP message can be thought of as a a function
with three inputs:
1. The cleartext message. This is either an HTTP message
or some other data object. Note that since the cleartext message
is carried transparently, headers and all, any version of HTTP
can be carried within an S-HTTP wrapper.
2. The receiver's cryptographic preferences and keying material.
This is either explicitly specified by the receiver or subject
to some default set of preferences.
3. The sender's cryptographic preferences and keying material.
This input to the function can be thought of as implicit
since it exists only in the memory of the sender.
In order to create an S-HTTP message, then, the sender integrates the
sender's preferences with the receiver's preferences. The result of
this is a list of cryptographic enhancements to be applied and keying
material to be used to apply them. This may require some user
intervention. For instance, there might be multiple keys available to
sign the message. (See Section 3.2.4.9.3 for more on this topic.)
Using this data, the sender applies the enhancements to the message
clear-text to create the S-HTTP message.
The processing steps required to transform the cleartext message into
the S-HTTP message are described in Sections 2 and 3. The processing
steps required to merge the sender's and receiver's preferences are
described in Sections 3.2.
The recovery of an S-HTTP message can be thought of as a function of
four distinct inputs:
1. The S-HTTP message.
2. The receiver's stated cryptographic preferences and keying
material. The receiver has the opportunity to remember what
cryptographic preferences it provided in order for this
document to be dereferenced.
3. The receiver's current cryptographic preferences and
keying material.
4. The sender's previously stated cryptographic options.
The sender may have stated that he would perform certain
cryptographic operations in this message. (Again, see
sections 4 and 5 for details on how to do this.)
Rescorla & Schiffman Experimental [Page 5]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
In order to recover an S-HTTP message, the receiver needs to read the
headers to discover which cryptographic transformations were
performed on the message, then remove the transformations using some
combination of the sender's and receiver's keying material, while
taking note of which enhancements were applied.
The receiver may also choose to verify that the applied enhancements
match both the enhancements that the sender said he would apply
(input 4 above) and that the receiver requested (input 2 above) as
well as the current preferences to see if the S-HTTP message was
appropriately transformed. This process may require interaction with
the user to verify that the enhancements are acceptable to the user.
(See Section 6.4 for more on this topic.)
Message protection may be provided on three orthogonal axes:
signature, authentication, and encryption. Any message may be signed,
authenticated, encrypted, or any combination of these (including no
protection).
Multiple key management mechanisms are supported, including
password-style manually shared secrets and public-key key exchange.
In particular, provision has been made for prearranged (in an earlier
transaction or out of band) symmetric session keys in order to send
confidential messages to those who have no public key pair.
Additionally, a challenge-response ("nonce") mechanism is provided to
allow parties to assure themselves of transaction freshness.
If the digital signature enhancement is applied, an appropriate
certificate may either be attached to the message (possibly along
with a certificate chain) or the sender may expect the recipient to
obtain the required certificate (chain) independently.
In support of bulk encryption, S-HTTP defines two key transfer
mechanisms, one using public-key enveloped key exchange and another
with externally arranged keys.
In the former case, the symmetric-key cryptosystem parameter is
passed encrypted under the receiver's public key.
Rescorla & Schiffman Experimental [Page 6]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
In the latter mode, we encrypt the content using a prearranged
session key, with key identification information specified on one of
the header lines.
Secure HTTP provides a means to verify message integrity and sender
authenticity for a message via the computation of a Message
Authentication Code (MAC), computed as a keyed hash over the document
using a shared secret -- which could potentially have been arranged
in a number of ways, e.g.: manual arrangement or 'inband' key
management. This technique requires neither the use of public key
cryptography nor encryption.
This mechanism is also useful for cases where it is appropriate to
allow parties to identify each other reliably in a transaction
without providing (third-party) non-repudiability for the
transactions themselves. The provision of this mechanism is motivated
by our bias that the action of "signing" a transaction should be
explicit and conscious for the user, whereas many authentication
needs (i.e., access control) can be met with a lighter-weight
mechanism that retains the scalability advantages of public-key
cryptography for key exchange.
The protocol provides a simple challenge-response mechanism, allowing
both parties to insure the freshness of transmissions. Additionally,
the integrity protection provided to HTTP headers permits
implementations to consider the Date: header allowable in HTTP
messages as a freshness indicator, where appropriate (although this
requires implementations to make allowances for maximum clock skew
between parties, which we choose not to specify).
In order to encourage widespread adoption of secure documents for the
World-Wide Web in the face of the broad scope of application
requirements, variability of user sophistication, and disparate
implementation constraints, Secure HTTP deliberately caters to a
variety of implementation options. See Section 8 for implementation
recommendations and requirements.
Syntactically, Secure HTTP messages are the same as HTTP, consisting
of a request or status line followed by headers and a body. However,
the range of headers is different and the bodies are typically
Rescorla & Schiffman Experimental [Page 7]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
cryptographically enhanced.
In order to differentiate S-HTTP messages from HTTP messages and
allow for special processing, the request line should use the special
Secure" method and use the protocol designator "Secure-HTTP/1.4".
Consequently, Secure-HTTP and HTTP processing can be intermixed on
the same TCP port, e.g. port 80. In order to prevent leakage of
potentially sensitive information Request-URI should be "*". For
example:
Secure * Secure-HTTP/1.4
When communicating via a proxy, the Request-URI should be consist of
the AbsoluteURI. Typically, the rel path section should be replaced
by "*" to minimize the information passed to in the clear. (e.g.
http://www.terisa.com/*); proxies should remove the appropriate
amount of this information to minimize the threat of traffic
analysis. See Section 7.2.2.1 for a situation where providing more
information is appropriate.
S-HTTP responses should use the protocol designator "Secure-
HTTP/1.4". For example:
Secure-HTTP/1.4 200 OK
Note that the status in the Secure HTTP response line does not
indicate anything about the success or failure of the unwrapped HTTP
request. Servers should always use 200 OK provided that the Secure
HTTP processing is successful. This prevents analysis of success or
failure for any request, which the correct recipient can determine
from the encapsulated data. All case variations should be accepted.
The header lines described in this section go in the header of a
Secure HTTP message. All except 'Content-Type' and 'Content-Privacy-
Domain' are optional. The message body shall be separated from the
header block by two successive CRLFs.
Rescorla & Schiffman Experimental [Page 8]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
All data and fields in header lines should be treated as case
insensitive unless otherwise specified. Linear whitespace [RFC-822]
should be used only as a token separator unless otherwise quoted.
Long header lines may be line folded in the style of [RFC-822].
This document refers to the header block following the S-HTTP
request/response line and preceding the successive CRLFs collectively
as "S-HTTP headers".
The two values defined by this document are 'MOSS' and 'CMS'. CMS
refers to the privacy enhancement specified in section 2.6.1. MOSS
refers to the format defined in [RFC-1847] and [RFC-1848].
Under normal conditions, the terminal encapsulated content (after all
privacy enhancements have been removed) would be an HTTP message. In
this case, there shall be a Content-Type line reading:
Content-Type: message/http
The message/http content type is defined in RFC-2616.
If the inner message is an S-HTTP message, then the content type
shall be 'application/s-http'. (See Appendix for the definition of
this.)
It is intended that these types be registered with IANA as MIME
content types.
The terminal content may be of some other type provided that the type
is properly indicated by the use of an appropriate Content-Type
header line. In this case, the header fields for the encapsulation of
the terminal content apply to the terminal content (the 'final
headers'). But in any case, final headers should themselves always be
S-HTTP encapsulated, so that the applicable S-HTTP/HTTP headers are
never passed unenhanced.
S-HTTP encapsulation of non-HTTP data is a useful mechanism for
passing pre-enhanced data (especially presigned data) without
requiring that the HTTP headers themselves be pre-enhanced.
Rescorla & Schiffman Experimental [Page 9]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
The Content-Type for MOSS shall be an acceptable MIME content type
describing the cryptographic processing applied. (e.g.
multipart/signed). The content type of the inner content is described
in the content type line corresponding to that inner content, and for
HTTP messages shall be 'message/http'.
This header line is intended to convey information about a key which
has been arranged outside of the internal cryptographic format. One
use of this is to permit in-band communication of session keys for
return encryption in the case where one of the parties does not have
a key pair. However, this should also be useful in the event that the
parties choose to use some other mechanism, for instance, a one-time
key list.
This specification defines two methods for exchanging named keys,
Inband, Outband. Inband indicates that the session key was exchanged
previously, using a Key-Assign header of the corresponding method.
Outband arrangements imply that agents have external access to key
materials corresponding to a given name, presumably via database
access or perhaps supplied immediately by a user from keyboard input.
The syntax for the header line is:
Prearranged-Key-Info =
"Prearranged-Key-Info" ":" Hdr-Cipher "," CoveredDEK "," CoverKey-ID
CoverKey-ID = method ":" key-name
CoveredDEK = *HEX
method = "inband" | "outband"
While chaining ciphers require an Initialization Vector (IV) [FIPS-
81] to start off the chaining, that information is not carried by
this field. Rather, it should be passed internal to the cryptographic
format being used. Likewise, the bulk cipher used is specified in
this fashion.
<Hdr-Cipher> should be the name of the block cipher used to encrypt
the session key (see section 3.2.4.7)
<CoveredDEK> is the protected Data Encryption Key (a.k.a. transaction
key) under which the encapsulated message was encrypted. It should be
appropriately (randomly) generated by the sending agent, then
encrypted under the cover of the negotiated key (a.k.a. session key)
using the indicated header cipher, and then converted into hex.
Rescorla & Schiffman Experimental [Page 10]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
In order to avoid name collisions, cover key namespaces must be
maintained separately by host and port.
Note that some Content-Privacy-Domains, notably likely future
revisions of MOSS and CMS may have support for symmetric key
management.
The Prearranged-Key-Info field need not be used in such
circumstances. Rather, the native syntax is preferred. Keys
exchanged with Key-Assign, however, may be used in this situation.
This header is used to supply a Message Authenticity Check, providing
both message authentication and integrity, computed from the message
text, the time (optional -- to prevent replay attack), and a shared
secret between client and server. The MAC should be computed over the
encapsulated content of the S-HTTP message. S-HTTP/1.1 defined that
MACs should be computed using the following algorithm ('||' means
concatenation):
MAC = hex(H(Message||[<time>]||<shared key>))
The time should be represented as an unsigned 32 bit quantity
representing seconds since 00:00:00 GMT January 1, 1970 (the UNIX
epoch), in network byte order. The shared key format is a local
matter.
Recent research [VANO95] has demonstrated some weaknesses in this
approach, and this memo introduces a new construction, derived from
[RFC-2104]. In the name of backwards compatibility, we retain the
previous constructions with the same names as before. However, we
also introduce a new series of names (See Section 3.2.4.8 for the
names) that obey a different (hopefully stronger) construction. (^
means bitwise XOR)
HMAC = hex(H(K' ^ pad2 || H(K' ^ pad1 ||[<time>]|| Message)))
pad1 = the byte 0x36 repeated enough times to fill out a
hash input block. (I.e. 64 times for both MD5 and SHA-1)
pad2 = the byte 0x5c repeated enough times to fill out a
hash input block.
K' = H(<shared key>)
The original HMAC construction is for the use of a key with length
equal to the length of the hash output. Although it is considered
safe to use a key of a different length (Note that strength cannot be
increased past the length of the hash function itself, but can be
reduced by using a shorter key.) [KRAW96b] we hash the original key
Rescorla & Schiffman Experimental [Page 11]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
to permit the use of shared keys (e.g. passphrases) longer than the
length of the hash. It is noteworthy (though obvious) that this
technique does not increase the strength of short keys.
The format of the MAC-Info line is:
MAC-Info =
"MAC-Info" ":" [hex-time],
hash-alg, hex-hash-data, key-spec
hex-time = <unsigned seconds since Unix epoch represented as HEX>
hash-alg = <hash algorithms from section 3.2.4.8>
hex-hash-data = <computation as described above represented as HEX>
Key-Spec = "null" | "dek" | Key-ID
Key-Ids can refer either to keys bound using the Key-Assign header
line or those bound in the same fashion as the Outband method
described later. The use of a 'Null' key-spec implies that a zero
length key was used, and therefore that the MAC merely represents a
hash of the message text and (optionally) the time. The special
key-spec 'DEK' refers to the Data Exchange Key used to encrypt the
following message body (it is an error to use the DEK key-spec in
situations where the following message body is unencrypted).
If the time is omitted from the MAC-Info line, it should simply not
be included in the hash.
Note that this header line can be used to provide a more advanced
equivalent of the original HTTP Basic authentication mode in that the
user can be asked to provide a username and password. However, the
password remains private and message integrity can be assured.
Moreover, this can be accomplished without encryption of any kind.
In addition, MAC-Info permits fast message integrity verification (at
the loss of non-repudiability) for messages, provided that the
participants share a key (possibly passed using Key-Assign in a
previous message).
Note that some Content-Privacy-Domains, notably likely future
revisions of MOSS and CMS may have support for symmetric integrity
protection The MAC-Info field need not be used in such circumstances.
Rather, the native syntax is preferred. Keys exchanged with Key-
Assign, however, may be used in this situation.
The content of the message is largely dependent upon the values of
the Content-Privacy-Domain and Content-Transfer-Encoding fields.
Rescorla & Schiffman Experimental [Page 12]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
For a CMS message, with '8BIT' Content-Transfer-Encoding, the content
should simply be the CMS message itself.
If the Content-Privacy-Domain is MOSS, the content should consist of
a MOSS Security Multipart as described in RFC1847.
It is expected that once the privacy enhancements have been removed,
the resulting (possibly protected) contents will be a normal HTTP
request. Alternately, the content may be another Secure-HTTP message,
in which case privacy enhancements should be unwrapped until clear
content is obtained or privacy enhancements can no longer be removed.
(This permits embedding of enhancements, such as sequential Signed
and Enveloped enhancements.) Provided that all enhancements can be
removed, the final de-enhanced content should be a valid HTTP request
(or response) unless otherwise specified by the Content-Type line.
Note that this recursive encapsulation of messages potentially
permits security enhancements to be applied (or removed) for the
benefit of intermediaries who may be a party to the transaction
between a client and server (e.g., a proxy requiring client
authentication). How such intermediaries should indicate such
processing is described in Section 7.2.1.
Content-Privacy-Domain 'CMS' follows the form of the CMS standard
(see Appendix).
Message protection may proceed on two orthogonal axes: signature and
encryption. Any message may be either signed, encrypted, both, or
neither. Note that the 'auth' protection mode of S-HTTP is provided
independently of CMS coding via the MAC-Info header of section 2.3.6
since CMS does not support a 'KeyDigestedData' type, although it does
support a 'DigestedData' type.
This enhancement uses the 'SignedData' type of CMS. When digital
signatures are used, an appropriate certificate may either be
attached to the message (possibly along with a certificate chain) as
specified in CMS or the sender may expect the recipient to obtain its
certificate (and/or chain) independently. Note that an explicitly
allowed instance of this is a certificate signed with the private
component corresponding to the public component being attested to.
This shall be referred to as a self-signed certificate. What, if any,
weight to give to such a certificate is a purely local matter. In
Rescorla & Schiffman Experimental [Page 13]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
either case, a purely signed message is precisely CMS compliant.
This enhancement is performed precisely as enveloping (using either '
EnvelopedData' types) under CMS. A message encrypted in this fashion,
signed or otherwise, is CMS compliant. To have a message which is
both signed and encrypted, one simply creates the CMS SignedData
production and encapsulates it in EnvelopedData as described in CMS.
This uses the 'EncryptedData' type of CMS. In this mode, we encrypt
the content using a DEK encrypted under cover of a prearranged
session key (how this key may be exchanged is discussed later), with
key identification information specified on one of the header lines.
The IV is in the EncryptedContentInfo type of the EncryptedData
element. To have a message which is both signed and encrypted, one
simply creates the CMS SignedData production and encapsulates it in
EncryptedData as described in CMS.
The body of the message should be a MIME compliant message with
content type that matches the Content-Type line in the S-HTTP
headers. Encrypted messages should use multipart/encrypted. Signed
messages should use multipart/signed. However, since multipart/signed
does not convey keying material, is is acceptable to use
multipart/mixed where the first part is application/mosskey-data and
the second part is multipart/mixed in order to convey certificates
for use in verifying the signature.
Implementation Note: When both encryption and signature are applied
by the same agent, signature should in general be applied before
encryption.
In general, HTTP [RFC-2616] headers should appear in the inner
content (i.e. the message/http) of an S-HTTP message but should not
appear in the S-HTTP message wrapper for security reasons. However,
certain headers need to be visible to agents which do not have access
to the encapsulated data. These headers may appear in the S-HTTP
headers as well.
Rescorla & Schiffman Experimental [Page 14]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
Please note that although brief descriptions of the general purposes
of these headers are provided for clarity, the definitive reference
is [RFC-2616].
The host header specificies the internet host and port number of the
resource being requested. This header should be used to disambiguate
among multiple potential security contexts within which this message
could be interpreted. Note that the unwrapped HTTP message will have
it's own Host field (assuming it's an HTTP/1.1 message). If these
fields do not match, the server should respond with a 400 status
code.
The Connection field has precisely the same semantics in S-HTTP
headers as it does in HTTP headers. This permits persistent
connections to be used with S-HTTP.
As described in Section 1.3.2, every S-HTTP request is (at least
conceptually) preconditioned by the negotiation options provided by
the potential receiver. The two primary locations for these options
are
1. In the headers of an HTTP Request/Response.
2. In the HTML which contains the anchor being dereferenced.
There are two kinds of cryptographic options which may be provided:
Negotiation options, as discussed in Section 3.2 convey a potential
message recipient's cryptographic preferences. Keying options, as
discussed in Section 3.3 provide keying material (or pointers to
keying material) which may be of use to the sender when enhancing a
message.
Binding cryptographic options to anchors using HTML extensions is the
topic of the companion document [SHTML] and will not be treated here.
Rescorla & Schiffman Experimental [Page 15]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
Both parties are able to express their requirements and preferences
regarding what cryptographic enhancements they will permit/require
the other party to provide. The appropriate option choices depend on
implementation capabilities and the requirements of particular
applications.
A negotiation header is a sequence of specifications each conforming
to a four-part schema detailing:
Property -- the option being negotiated, such as bulk encryption
algorithm.
Value -- the value being discussed for the property, such as
DES-CBC
Direction -- the direction which is to be affected, namely:
during reception or origination (from the perspective of the
originator).
Strength -- strength of preference, namely: required, optional,
refused
As an example, the header line:
SHTTP-Symmetric-Content-Algorithms: recv-optional=DES-CBC,RC2
could be thought to say: "You are free to use DES-CBC or RC2 for bulk
encryption for encrypting messages to me."
We define new headers (to be used in the encapsulated HTTP header,
not in the S-HTTP header) to permit negotiation of these matters.
The general format for negotiation options is:
Option = Field ":" Key-val ";" *(Key-val)
Key-val = Key "=" Value *("," Value)
Key = Mode"-"Action ; This is represented as one
; token without whitespace
Mode = "orig" | "recv"
Action = "optional" | "required" | "refused"
Rescorla & Schiffman Experimental [Page 16]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
The <Mode> value indicates whether this <Key-val> refers to what the
agent's actions are upon sending privacy enhanced messages as opposed
to upon receiving them. For any given mode-action pair, the
interpretation to be placed on the enhancements (<Value>s) listed is:
'recv-optional:' The agent will process the enhancement if the
other party uses it, but will also gladly process messages
without the enhancement.
'recv-required:' The agent will not process messages without
this enhancement.
'recv-refused:' The agent will not process messages with this
enhancement.
'orig-optional:' When encountering an agent which refuses this
enhancement, the agent will not provide it, and when
encountering an agent which requires it, this agent will provide
it.
'orig-required:' The agent will always generate the enhancement.
'orig-refused:' The agent will never generate the enhancement.
The behavior of agents which discover that they are communicating
with an incompatible agent is at the discretion of the agents. It is
inappropriate to blindly persist in a behavior that is known to be
unacceptable to the other party. Plausible responses include simply
terminating the connection, or, in the case of a server response,
returning 'Not implemented 501'.
Optional values are considered to be listed in decreasing order of
preference. Agents are free to choose any member of the intersection
of the optional lists (or none) however.
If any <Key-Val> is left undefined, it should be assumed to be set to
the default. Any key which is specified by an agent shall override
any appearance of that key in any <Key-Val> in the default for that
field.
Rescorla & Schiffman Experimental [Page 17]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
For ciphers with variable key lengths, values may be parametrized
using the syntax <cipher>'['<length>']'
For example, 'RSA[1024]' represents a 1024 bit key for RSA. Ranges
may be represented as
<cipher>'['<bound1>'-'<bound2>']'
For purposes of preferences, this notation should be treated as if it
read (assuming x and y are integers)
<cipher>[x], <cipher>[x+1],...<cipher>[y] (if x<y)
and
<cipher>[x], <cipher>[x-1],...<cipher>[y] (if x>y)
The special value 'inf' may be used to denote infinite length.
Using simply <cipher> for such a cipher shall be read as the maximum
range possible with the given cipher.
This header refers to the Content-Privacy-Domain type of section
2.3.1. Acceptable values are as listed there. For instance,
SHTTP-Privacy-Domains: orig-required=cms;
recv-optional=cms,MOSS
would indicate that the agent always generates CMS compliant
messages, but can read CMS or MOSS (or, unenhanced messages).
This header indicates which algorithms may be used for key exchange.
Defined values are 'DH', 'RSA', 'Outband' and 'Inband'. DH refers to
Diffie-Hellman X9.42 style enveloping. [DH] RSA refers to RSA
enveloping. Outband refers to some sort of external key agreement.
Rescorla & Schiffman Experimental [Page 18]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
Inband refers to section 3.3.3.1.
The expected common configuration of clients having no certificates
and servers having certificates would look like this (in a message
sent by the server):
SHTTP-Key-Exchange-Algorithms: orig-optional=Inband, DH;
recv-required=DH
This header indicates what Digital Signature algorithms may be used.
Defined values are 'RSA' [PKCS-1] and 'NIST-DSS' [FIPS-186] Since
NIST-DSS and RSA use variable length moduli the parametrization
syntax of section 3.2.3 should be used. Note that a key length
specification may interact with the acceptability of a given
certificate, since keys (and their lengths) are specified in public-
key certificates.
This indicates what message digest algorithms may be used.
Previously defined values are 'RSA-MD2' [RFC-1319], 'RSA-MD5' [RFC-
1321], 'NIST-SHS' [FIPS-180].
This header specifies the symmetric-key bulk cipher used to encrypt
message content. Defined values are:
DES-CBC -- DES in Cipher Block Chaining (CBC) mode [FIPS-81]
DES-EDE-CBC -- 2 Key 3DES using Encrypt-Decrypt-Encrypt in outer
CBC mode
DES-EDE3-CBC -- 3 Key 3DES using Encrypt-Decrypt-Encrypt in outer
CBC mode
DESX-CBC -- RSA's DESX in CBC mode
IDEA-CBC -- IDEA in CBC mode
RC2-CBC -- RSA's RC2 in CBC mode
CDMF-CBC -- IBM's CDMF (weakened key DES) [JOHN93] in CBC mode
Since RC2 keys are variable length, the syntax of section 3.2.3
should be used.
Rescorla & Schiffman Experimental [Page 19]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
This header specifies the symmetric-key cipher used to encrypt
message headers.
DES-ECB -- DES in Electronic Codebook (ECB) mode [FIPS-81]
DES-EDE-ECB -- 2 Key 3DES using Encrypt-Decrypt-Encrypt in ECB mode
DES-EDE3-ECB -- 3 Key 3DES using Encrypt-Decrypt-Encrypt in ECB mode
DESX-ECB -- RSA's DESX in ECB mode
IDEA-ECB -- IDEA
RC2-ECB -- RSA's RC2 in ECB mode
CDMF-ECB -- IBM's CDMF in ECB mode
Since RC2 is variable length, the syntax of section 3.2.3 should be
used.
This header indicates what algorithms are acceptable for use in
providing a symmetric key MAC. 'RSA-MD2', 'RSA-MD5' and 'NIST-SHS'
persist from S-HTTP/1.1 using the old MAC construction. The tokens '
RSA-MD2-HMAC', 'RSA-MD5-HMAC' and 'NIST-SHS-HMAC' indicate the new
HMAC construction of 2.3.6 with the MD2, MD5, and SHA-1 algorithms
respectively.
This header indicates security enhancements to apply. Possible
values are 'sign', 'encrypt' and 'auth' indicating whether messages
are signed, encrypted, or authenticated (i.e., provided with a MAC),
respectively.
This is a generalized pattern match syntax to describe identifiers
for a large number of types of keying material. The general syntax
is:
Your-Key-Pattern =
"Your-Key-Pattern" ":" key-use "," pattern-info
key-use = "cover-key" | "auth-key" | "signing-key"
Rescorla & Schiffman Experimental [Page 20]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
This header specifies desired values for key names used for
encryption of transaction keys using the Prearranged-Key-Info syntax
of section 2.3.5. The pattern-info syntax consists of a series of
comma separated regular expressions. Commas should be escaped with
backslashes if they appear in the regexps. The first pattern should
be assumed to be the most preferred.
Auth-key patterns specify name forms desired for use for MAC
authenticators. The pattern-info syntax consists of a series of
comma separated regular expressions. Commas should be escaped with
backslashes if they appear in the regexps. The first pattern should
be assumed to be the most preferred.
This parameter describes a pattern or patterns for what keys are
acceptable for signing for the digital signature enhancement. The
pattern-info syntax for signing-key is:
pattern-info = name-domain "," pattern-data
The only currently defined name-domain is 'DN-1779'. This parameter
specifies desired values for fields of Distinguished Names. DNs are
considered to be represented as specified in RFC1779, the order of
fields and whitespace between fields is not significant.
All RFC1779 values should use ',' as a separator rather than ';',
since ';' is used as a statement separator in S-HTTP.
Pattern-data is a modified RFC1779 string, with regular expressions
permitted as field values. Pattern match is performed field-wise,
unspecified fields match any value (and therefore leaving the DN-
Pattern entirely unspecified allows for any DN). Certificate chains
may be matched as well (to allow for certificates without name
subordination). DN chains are considered to be ordered left-to-right
with the issuer of a given certificate on its immediate right,
although issuers need not be specified. A trailing '.' indicates that
the sequence of DNs is absolute. I.e. that the one furthest to the
right is a root.
Rescorla & Schiffman Experimental [Page 21]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
The syntax for the pattern values is,
Value = DN-spec *("," Dn-spec)["."]
Dn-spec = "/" *(Field-spec) "/"
Field-spec := Attr = "Pattern"
Attr = "CN" | "L" | "ST" | "O" |
"OU" | "C" | <or as appropriate>
Pattern = <POSIX 1003.2 regular expressions>
For example, to request that the other agent sign with a key
certified by the RSA Persona CA (which uses name subordination) one
could use the expression below. Note the use of RFC1779 quoting to
protect the comma (an RFC1779 field separator) and the POSIX 1003.2
quoting to protect the dot (a regular expression metacharacter).
Your-Key-Pattern: signing-key, DN-1779,
/OU=Persona Certificate, O="RSA Data Security,
Inc\."/
Explicit negotiation parameters take precedence over default values.
For a given negotiation option type, defaults for a given mode-action
pair (such as 'orig-required') are implicitly merged unless
explicitly overridden.
The default values (these may be negotiated downward or upward) are:
SHTTP-Privacy-Domains: orig-optional=CMS;
recv-optional=CMS
SHTTP-Certificate-Types: orig-optional=X.509;
recv-optional=X.509
SHTTP-Key-Exchange-Algorithms: orig-optional=DH,Inband,Outband;
Rescorla & Schiffman Experimental [Page 22]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
recv-optional=DH,Inband,Outband
SHTTP-Signature-Algorithms: orig-optional=NIST-DSS;
recv-optional=NIST-DSS
SHTTP-Message-Digest-Algorithms: orig-optional=RSA-MD5;
recv-optional=RSA-MD5
SHTTP-Symmetric-Content-Algorithms: orig-optional=DES-CBC;
recv-optional=DES-CBC
SHTTP-Symmetric-Header-Algorithms: orig-optional=DES-ECB;
recv-optional=DES-ECB
SHTTP-Privacy-Enhancements: orig-optional=sign,encrypt, auth;
recv-required=encrypt;
recv-optional=sign, auth
This header identifies a potential principal for whom the message
described by these options could be encrypted; Note that this
explicitly permits return encryption under (say) public key without
the other agent signing first (or under a different key than that of
the signature). The syntax of the Encryption-Identity line is:
Encryption-Identity =
"Encryption Identity" ":" name-class,key-sel,name-arg
name-class = "DN-1779" | MOSS name forms
The name-class is an ASCII string representing the domain within
which the name is to be interpreted, in the spirit of MOSS. In
addition to the MOSS name forms of RFC1848, we add the DN-1779 name
form to represent a more convenient form of distinguished name.
In order to permit public key operations on DNs specified by
Encryption-Identity headers without explicit certificate fetches by
the receiver, the sender may include certification information in the
Certificate-Info option. The format of this option is:
Certificate-Info: <Cert-Fmt>','<Cert-Group>
<Cert-Fmt> should be the type of <Cert-Group> being presented.
Rescorla & Schiffman Experimental [Page 23]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
Defined values are 'PEM' and 'CMS'. CMS certificate groups are
provided as a base-64 encoded CMS SignedData message containing
sequences of certificates with or without the SignerInfo field. A PEM
format certificate group is a list of comma-separated base64-encoded
PEM certificates.
Multiple Certificate-Info lines may be defined.
This option serves to indicate that the agent wishes to bind a key to
a symbolic name for (presumably) later reference.
The general syntax of the key-assign header is:
Key-Assign =
"Key-Assign" ":" Method "," Key-Name ","
Lifetime "," Ciphers ";" Method-args
Key-name = string
Lifetime = "this" | "reply" | ""
Method ="inband"
Ciphers = "null" | Cipher+
Cipher" = <Header cipher from section 3.2.4.7>
kv = "4" | "5"
Key-Name is the symbolic name to which this key is to be bound.
Ciphers is a list of ciphers for which this key is potentially
applicable (see the list of header ciphers in section 3.2.4.7). The
keyword 'null' should be used to indicate that it is inappropriate
for use with ANY cipher. This is potentially useful for exchanging
keys for MAC computation.
Lifetime is a representation of the longest period of time during
which the recipient of this message can expect the sender to accept
that key. 'this' indicates that it is likely to be valid only for
reading this transmission. 'reply' indicates that it is useful for a
reply to this message. If a Key-Assign with the reply lifetime
appears in a CRYPTOPTS block, it indicates that it is good for at
least one (but perhaps only one) dereference of this anchor. An
unspecified lifetime implies that this key may be reused for an
indefinite number of transactions.
Method should be one of a number of key exchange methods. The only
currently defined value is 'inband' referring to Inband keys (i.e.,
direct assignment).
Rescorla & Schiffman Experimental [Page 24]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
This header line may appear either in an unencapsulated header or in
an encapsulated message, though when an uncovered key is being
directly assigned, it may only appear in an encrypted encapsulated
content. Assigning to a key that already exists causes that key to be
overwritten.
Keys defined by this header are referred to elsewhere in this
specification as Key-IDs, which have the syntax:
Key-ID = method ":" key-name
This refers to the direct assignment of an uncovered key to a
symbolic name. Method-args should be just the desired session key
encoded in hexidecimal as in:
Key-Assign: inband,akey,reply,DES-ECB;0123456789abcdef
Short keys should be derived from long keys by reading bits from left
to right.
Note that inband key assignment is especially important in order to
permit confidential spontaneous communication between agents where
one (but not both) of the agents have key pairs. However, this
mechanism is also useful to permit key changes without public key
computations. The key information is carried in this header line must
be in the inner secured HTTP request, therefore use in unencrypted
messages is not permitted.
Nonces are opaque, transient, session-oriented identifiers which may
be used to provide demonstrations of freshness. Nonce values are a
local matter, although they are might well be simply random numbers
generated by the originator. The value is supplied simply to be
returned by the recipient.
This header is used by an originator to specify what value is to be
returned in the reply. The field may be any value. Multiple nonces
may be supplied, each to be echoed independently.
The Nonce should be returned in a Nonce-Echo header line. See section
4.1.1.
Rescorla & Schiffman Experimental [Page 25]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
In order for servers to bind a group of headers to an HTML anchor, it
is possible to combine a number of headers on a single S-HTTP
Cryptopts header line. The names of the anchors to which these
headers apply is indicated with a 'scope' parameter.
This option provides a set of cryptopts and a list of references to
which it applies. (For HTML, these references would be named using
the NAME tag). The names are provided in the scope attribute as a
comma separated list and separated from the next header line by a
semicolon. The format for the SHTTP-Cryptopts line is:
SHTTP-Cryptopts =
"SHTTP-Cryptopts" ":" scope ";" cryptopt-list
scope = "scope="<tag-spec> ; This is all one token without whitespace
tag-spec = tag *("," tag) | ""
cryptopt-list = cryptopt *(";" cryptopt)
cryptopt = <S-HTTP cryptopt lines described below>
tag = <value used in HTML anchor NAME attribute>
For example:
SHTTP-Cryptopts: scope=tag1,tag2;
SHTTP-Privacy-Domains:
orig-required=cms; recv-optional=cms,MOSS
If a message contains both S-HTTP negotiation headers and headers
grouped on SHTTP-Cryptopts line(s), the other headers shall be taken
to apply to all anchors not bound on the SHTTP-Cryptopts line(s).
Note that this is an all-or-nothing proposition. That is, if a
SHTTP-Cryptopts header binds options to a reference, then none of
these global options apply, even if some of the options headers do
not appear in the bound options. Rather, the S-HTTP defaults found in
Section 3.2.4.11 apply.
All S-HTTP compliant agents must generate the Security-Scheme header
in the headers of all HTTP messages they generate. This header
permits other agents to detect that they are communicating with an
S-HTTP compliant agent and generate the appropriate cryptographic
Rescorla & Schiffman Experimental [Page 26]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
options headers.
For implementations compliant with this specification, the value must
be 'S-HTTP/1.4'.
The header is used to return the value provided in a previously
received Nonce: field. This has to go in the encapsulated headers so
that it an be cryptographically protected.
A server may respond to a client request with an error code that
indicates that the request has not completely failed but rather that
the client may possibly achieve satisfaction through another request.
HTTP already has this concept with the 3XX redirection codes.
In the case of S-HTTP, it is conceivable (and indeed likely) that the
server expects the client to retry his request using another set of
cryptographic options. E.g., the document which contains the anchor
that the client is dereferencing is old and did not require digital
signature for the request in question, but the server now has a
policy requiring signature for dereferencing this URL. These options
should be carried in the header of the encapsulated HTTP message,
precisely as client options are carried.
The general idea is that the client will perform the retry in the
manner indicated by the combination of the original request and the
precise nature of the error and the cryptographic enhancements
depending on the options carried in the server response.
The guiding principle in client response to these errors should be to
provide the user with the same sort of informed choice with regard to
dereference of these anchors as with normal anchor dereference. For
instance, in the case above, it would be inappropriate for the client
to sign the request without requesting permission for the action.
Rescorla & Schiffman Experimental [Page 27]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
The HTTP errors 'Unauthorized 401', 'PaymentRequired 402' represent
failures of HTTP style authentication and payment schemes. While S-
HTTP has no explicit support for these mechanisms, they can be
performed under S-HTTP while taking advantage of the privacy services
offered by S-HTTP. (There are other errors for S-HTTP specific
authentication errors.)
This server status reply is provided so that the server may inform
the client that although the current request is rejected, a retried
request with different cryptographic enhancements is worth
attempting. This header shall also be used in the case where an HTTP
request has been made but an S-HTTP request should have been made.
Obviously, this serves no useful purpose other than signalling an
error if the original request should have been encrypted, but in
other situations (e.g. access control) may be useful.
In the case of a request that was made as an SHTTP request, it
indicates that for some reason the cryptographic enhancements applied
to the request were unsatisfactory and that the request should be
repeated with the options found in the response header. Note that
this can be used as a way to force a new public key negotiation if
the session key in use has expired or to supply a unique nonce for
the purposes of ensuring request freshness.
If the 420 code is returned in response to an HTTP request, it
indicates that the request should be retried using S-HTTP and the
cryptographic options indicated in the response header.
This error code indicates that something about the S-HTTP request was
bad. The error code is to be followed by an appropriate explanation,
e.g.:
421 BogusHeader Content-Privacy-Domain must be specified
Rescorla & Schiffman Experimental [Page 28]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
This response is analagous to the 420 response except that the
options in the message refer to enhancements that the client must
perform in order to satisfy the proxy.
This response code is specifically for use with proxy-server
interaction where the proxy has placed the If-Modified-Since header
in the S-HTTP headers of its request. This response indicates that
the following S-HTTP message contains sufficient keying material for
the proxy to forward the cached document for the new requestor.
In general, this takes the form of an S-HTTP message where the actual
enhanced content is missing, but all the headers and keying material
are retained. (I.e. the optional content section of the CMS message
has been removed.) So, if the original response was encrypted, the
response contains the original DEK re-covered for the new recipient.
(Notice that the server performs the same processing as it would have
in the server side caching case of 7.1 except that the message body
is elided.)
These headers are again internal to HTTP, but may contain S-HTTP
negotiation options of significance to S-HTTP. The request should be
redirected in the sense of HTTP, with appropriate cryptographic
precautions being observed.
Permitting automatic client retry in response to this sort of server
response permits several forms of attack. Consider for the moment
the simple credit card case:
The user views a document which requires his credit card. The
user verifies that the DN of the intended recipient is acceptable
and that the request will be encrypted and dereferences the
anchor. The attacker intercepts the server's reply and responds
with a message encrypted under the client's public key containing
the Moved 301 header. If the client were to automatically perform
this redirect it would allow compromise of the user's credit
card.
Rescorla & Schiffman Experimental [Page 29]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
This shows one possible danger of automatic retries -- potential
compromise of encrypted information. While it is impossible to
consider all possible cases, clients should never automatically
reencrypt data unless the server requesting the retry proves that he
already has the data. So, situations in which it would be acceptable
to reencrypt would be if:
1. The retry response was returned encrypted under an inband key
freshly generated for the original request.
2. The retry response was signed by the intended recipient of the
original request.
3. The original request used an outband key and the response is
encrypted under that key.
This is not an exhaustive list, however the browser author would be
well advised to consider carefully before implementing automatic
reencryption in other cases. Note that an appropriate behavior in
cases where automatic reencryption is not appropriate is to query the
user for permission.
Since we discourage automatic (without user confirmation) signing in
even the usual case, and given the dangers described above, it is
prohibited to automatically retry signature enchancement.
Servers which receive requests in the clear which should be secured
should return 'SecurityRetry 420' with header lines set to indicate
the required privacy enhancements.
We define a new URL protocol designator, 'shttp'. Use of this
designator as part of an anchor URL implies that the target server is
S-HTTP capable, and that a dereference of this URL should undergo S-
HTTP processing.
Rescorla & Schiffman Experimental [Page 30]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
Note that S-HTTP oblivious agents should not be willing to
dereference a URL with an unknown protocol specifier, and hence
sensitive data will not be accidentally sent in the clear by users of
non-secure clients.
While preparing a secure message, the browser should provide a visual
indication of the security of the transaction, as well as an
indication of the party who will be able to read the message. While
reading a signed and/or enveloped message, the browser should
indicate this and (if applicable) the identity of the signer. Self-
signed certificates should be clearly differentiated from those
validated by a certification hierarchy.
Failure to authenticate or decrypt an S-HTTP message should be
presented differently from a failure to retrieve the document.
Compliant clients may at their option display unverifiable documents
but must clearly indicate that they were unverifiable in a way
clearly distinct from the manner in which they display documents
which possessed no digital signatures or documents with verifiable
signatures.
Clients shall provide a method for determining that HTTP requests are
to be signed and for determining which (assuming there are many)
certificate is to be used for signature. It is suggested that users
be presented with some sort of selection list from which they may
choose a default. No signing should be performed without some sort of
explicit user interface action, though such action may take the form
of a persistent setting via a user preferences mechanism (although
this is discouraged.)
Clients shall provide a method to display the DN and certificate
chain associated with a given anchor to be dereferenced so that users
may determine for whom their data is being encrypted. This should be
distinct from the method for displaying who has signed the document
containing the anchor since these are orthogonal pieces of encryption
information.
Rescorla & Schiffman Experimental [Page 31]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
While S-HTTP has always supported preenhanced documents, in previous
versions it was never made clear how to actually implement them.
This section describes two methods for doing so: preenhancing the
HTTP request/response and preenhancing the underlying data.
The two primary motivations for preenhanced documents are security
and performance. These advantages primarily accrue to signing but may
also under special circumstances apply to confidentiality or
repudiable (MAC-based) authentication.
Consider the case of a server which repeatedly serves the same
content to multiple clients. One such example would be a server which
serves catalogs or price lists. Clearly, customers would like to be
able to verify that these are actual prices. However, since the
prices are typically the same to all comers, confidentiality is not
an issue. (Note: see Section 7.1.5 below for how to deal with this
case as well).
Consequently, the server might wish to sign the document once and
simply send the cached signed document out when a client makes a new
request, avoiding the overhead of a private key operation each time.
Note that conceivably, the signed document might have been generated
by a third party and placed in the server's cache. The server might
not even have the signing key! This illustrates the security benefit
of presigning: Untrusted servers can serve authenticated data without
risk even if the server is compromised.
The obvious implementation is simply to take a single
request/response, cache it, and send it out in situations where a new
message would otherwise be generated.
It is also possible using S-HTTP to sign the underlying data and send
it as an S-HTTP messsage. In order to do this, one would take the
signed document (a CMS or MOSS message) and attach both S-HTTP
headers (e.g. the S-HTTP request/response line, the Content-Privacy-
Domain) and the necessary HTTP headers (including a Content-Type that
reflects the inner content).
Rescorla & Schiffman Experimental [Page 32]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
SECURE * Secure-HTTP/1.4
Content-Type: text/html
Content-Privacy-Domain: CMS
Random signed message here...
This message itself cannot be sent, but needs to be recursively
encapsulated, as described in the next section.
As required by Section 7.3, the result above needs to be itself
encapsulated to protect the HTTP headers. the obvious case [and the
one illustrated here] is when confidentiality is required, but the
auth enhancement or even the null transform might be applied instead.
That is, the message shown above can be used as the inner content of
a new S-HTTP message, like so:
SECURE * Secure-HTTP/1.4
Content-Type: application/s-http
Content-Privacy-Domain: CMS
Encrypted version of the message above...
To unfold this, the receiver would decode the outer S-HTTP message,
reenter the (S-)HTTP parsing loop to process the new message, see
that that too was S-HTTP, decode that, and recover the inner content.
Note that this approach can also be used to provide freshness of
server activity (though not of the document itself) while still
providing nonrepudiation of the document data if a NONCE is included
in the request.
Although preenhancement works best with signature, it can also be
used with encryption under certain conditions. Consider the situation
where the same confidential document is to be sent out repeatedly.
The time spent to encrypt can be saved by caching the ciphertext and
simply generating a new key exchange block for each recipient. [Note
that this is logically equivalent to a multi- recipient message as
defined in both MOSS and CMS and so care must be taken to use proper
PKCS-1 padding if RSA is being used since otherwise, one may be open
to a low encryption exponent attack [HAST96].
Rescorla & Schiffman Experimental [Page 33]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
The use of S-HTTP presents implementation issues to the use of HTTP
proxies. While simply having the proxy blindly forward responses is
straightforward, it would be preferable if S-HTTP aware proxies were
still able to cache responses in at least some circumstances. In
addition, S-HTTP services should be usable to protect client-proxy
authentication. This section describes how to achieve those goals
using the mechanisms described above.
When an S-HTTP aware proxy receives a request (HTTP or S-HTTP) that
(by whatever access control rules it uses) it requires to be S-HTTP
authenticated (and if it isn't already so), it should return the 422
response code (5.7.4).
When the client receives the 422 response code, it should read the
cryptographic options that the proxy sent and determine whether or
not it is willing to apply that enhancement to the message. If the
client is willing to meet these requirements, it should recursively
encapsulate the request it previously sent using the appropriate
options. (Note that since the enhancement is recursively applied,
even clients which are unwilling to send requests to servers in the
clear may be willing to send the already encrypted message to the
proxy without further encryption.) (See Section 7.1 for another
example of a recursively encapsulated message)
When the proxy receives such a message, it should strip the outer
encapsulation to recover the message which should be sent to the
server.
All S-HTTP agents must support the MD5 message digest and MAC
authentication. As of S-HTTP/1.4, all agents must also support the
RSA-MD5-HMAC construction.
All S-HTTP agents must support Outband, Inband, and DH key exchange.
All agents must support encryption using DES-CBC.
Agents must support signature generation and verification using
NIST-DSS.
Rescorla & Schiffman Experimental [Page 34]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
We provide here a contrived example of a series of S-HTTP requests
and replies. Rows of equal signs are used to set off the narrative
from sample message traces. Note that the actual encrypted or signed
message bodies would normally be binary garbage. In an attempt to
preserve readability while still using (mostly) genuine messages, the
bodies of the requests have been base64 encoded. To regenerate actual
S-HTTP messages, it is necessary to remove the base64 encoding from
the message body.
Alice, using an S-HTTP-capable client, begins by making an HTTP
request which yields the following response page:
============================================================
200 OK HTTP/1.0
Server-Name: Navaho-0.1.3.3alpha
Certificate-Info: CMS,MIAGCSqGSIb3DQEHAqCAMIACAQExADCABgkqh
kiG9w0BBwEAAKCAM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Rescorla & Schiffman Experimental [Page 36]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
Encryption-Identity: DN-1779, null, CN=Setec Astronomy, OU=Persona
Certificate,O="RSA Data Security, Inc.", C=US;
SHTTP-Privacy-Enhancements: recv-required=encrypt
<A name=tag1 HREF="shttp://www.setec.com/secret">
Don't read this. </A>
============================================================
An appropriate HTTP request to dereference this URL would be:
============================================================
GET /secret HTTP/1.0
Security-Scheme: S-HTTP/1.4
User-Agent: Web-O-Vision 1.2beta
Accept: *.*
Key-Assign: Inband,1,reply,des-ecb;7878787878787878
============================================================
The added Key-Assign line that would not have been in an ordinary
HTTP request permits Bob (the server) to encrypt his reply to Alice,
even though Alice does not have a public key, since they would share
a key after the request is received by Bob. This request has the
following S-HTTP encapsulation:
============================================================
Secure * Secure-HTTP/1.4
Content-Type: message/http
Content-Privacy-Domain: CMS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============================================================
The data between the delimiters is a CMS message, RSA enveloped for
Setec Astronomy.
Bob decrypts the request, finds the document in question, and is
ready to serve it back to Alice.
Rescorla & Schiffman Experimental [Page 37]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
An appropriate HTTP server response would be:
============================================================
HTTP/1.0 200 OK
Security-Scheme: S-HTTP/1.4
Content-Type: text/html
Congratulations, you've won.
<A href="/prize.html"
CRYPTOPTS="Key-Assign: Inband,alice1,reply,des-ecb;020406080a0c0e0f;
SHTTP-Privacy-Enhancements: recv-required=auth">Click here to
claim your prize</A>
============================================================
This HTTP response, encapsulated as an S-HTTP message becomes:
============================================================
Secure * Secure-HTTP/1.4
Content-Type: message/http
Prearranged-Key-Info: des-ecb,697fa820df8a6e53,inband:1
Content-Privacy-Domain: CMS
MIAGCSqGSIb3DQEHBqCAMIACAQAwgAYJKoZIhvcNAQcBMBEGBSsOAwIHBAifqtdy
x6uIMYCCARgvFzJtOZBn773DtmXlx037ck3giqnV0WC0QAx5f+fesAiGaxMqWcir
r9XvT0nT0LgSQ/8tiLCDBEKdyCNgdcJAduy3D0r2sb5sNTT0TyL9uydG3w55vTnW
aPbCPCWLudArI1UHDZbnoJICrVehxG/sYX069M8v6VO8PsJS7//hh1yM+0nekzQ5
l1p0j7uWKu4W0csrlGqhLvEJanj6dQAGSTNCOoH3jzEXGQXntgesk8poFPfHdtj0
5RH4MuJRajDmoEjlrNcnGl/BdHAd2JaCo6uZWGcnGAgVJ/TVfSVSwN5nlCK87tXl
nL7DJwaPRYwxb3mnPKNq7ATiJPf5u162MbwxrddmiE7e3sST7naSN+GS0ateY5X7
AAAAAAAAAAA=
============================================================
The data between the delimiters is a CMS message encrypted under a
randomly-chosen DEK which can be recovered by computing:
DES-DECRYPT(inband:1,697fa820df8a6e53)
where 'inband:1' is the key exchanged in the Key-Assign line in the
original request.
Rescorla & Schiffman Experimental [Page 38]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
There is a link on the HTML page that was just returned, which Alice
dereferences, creating the HTTP message:
============================================================
GET /prize.html HTTP/1.0
Security-Scheme: S-HTTP/1.4
User-Agent: Web-O-Vision 1.1beta
Accept: *.*
============================================================
Which, when encapsulated as an S-HTTP message, becomes:
============================================================
Secure * Secure-HTTP/1.4
Content-Type: message/http
MAC-Info:31ff8122,rsa-md5,b3ca4575b841b5fc7553e69b0896c416,inband:alice1
Content-Privacy-Domain: CMS
MIAGCSqGSIb3DQEHAaCABGNHRVQgL3ByaXplLmh0bWwgSFRUUC8xLjAKU2VjdXJp
dHktU2NoZW1lOiBTLUhUVFAvMS4xClVzZXItQWdlbnQ6IFdlYi1PLVZpc2lvbiAx
LjFiZXRhCkFjY2VwdDogKi4qCgoAAAAA
============================================================
The data between the delimiters is a CMS 'Data' representation of the
request.
Rescorla & Schiffman Experimental [Page 39]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
Appendix: A Review of CMS
CMS ("Cryptographic Message Syntax Standard") is a cryptographic
message encapsulation format, similar to PEM, based on RSA's PKCS-7
cryptographic messaging syntax.
CMS is only one of two encapsulation formats supported by S-HTTP, but
it is to be preferred since it permits the least restricted set of
negotiable options, and permits binary encoding. In the interest of
making this specification more self-contained, we summarize CMS here.
CMS is defined in terms of OSI's Abstract Syntax Notation (ASN.1,
defined in X.208), and is concretely represented using ASN.1's Basic
Encoding Rules (BER, defined in X.209). A CMS message is a sequence
of typed content parts. There are six content types, recursively
composable:
Data -- Some bytes, with no enhancement.
SignedData -- A content part, with zero or more signature
blocks, and associated keying materials. Keying materials
can be transported via the degenerate case of no signature
blocks and no data.
EnvelopedData -- One or more (per recipient) key exchange
blocks and an encrypted content part.
DigestedData -- A content part with a single digest block.
EncryptedData -- An encrypted content part, with key
materials externally provided.
Here we will dispense with convention for the sake of ASN.1-impaired
readers, and present a syntax for CMS in informal BNF (with much
gloss). In the actual encoding, most productions have explicit tag
and length fields.
Message = *Content
Content = Data | SignedData | EnvelopedData |
DigestedData | EncryptedData
Data = Bytes
SignedData = *DigestAlg Content *Certificates
*CRLs SignerInfo*
EnvelopedData = *RecipientInfo BulkCryptAlg
Encrypted(Content)
Rescorla & Schiffman Experimental [Page 40]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
DigestedData = DigestAlg Content DigestBytes
EncryptedData = BulkCryptAlg Encrypted(Bytes)
SignerInfo = CertID ... Encrypted(DigestBytes) ...
RecipientInfo = CertID KeyCryptAlg Encrypted(DEK)
Appendix: Internet Media Type message/s-http
In addition to defining the S-HTTP/1.4 protocol, this document serves
as the specification for the Internet media type "message/s-http".
The following is to be registered with IANA.
Media Type name: message
Media subtype name: s-http
Required parameters: none
Optional parameters: version, msgtype
version: The S-HTTP version number of the enclosed message
(e.g. "1.4"). If not present, the version can be
determined from the first line of the body.
msgtype: The message type -- "request" or "response".
If not present, the type can be determined from the
first line of the body.
Encoding considerations: only "7bit", "8bit", or "binary"
are permitted.
Security considerations: this is a security protocol.
Bibliography and References
[BELL96] Bellare, M., Canetti, R., Krawczyk, H., "Keying Hash
Functions for Message Authentication", Preprint.
[FIPS-46-1] Federal Information Processing Standards Publication
(FIPS PUB) 46-1, Data Encryption Standard, Reaffirmed
1988 January 22 (supersedes FIPS PUB 46, 1977 January
15).
[FIPS-81] Federal Information Processing Standards Publication
(FIPS PUB) 81, DES Modes of Operation, 1980 December 2.
[FIPS-180] Federal Information Processing Standards Publication
(FIPS PUB) 180-1, "Secure Hash Standard", 1995 April 17.
[FIPS-186] Federal Information Processing Standards Publication
(FIPS PUB) 186, Digital Signature Standard, 1994 May 19.
Rescorla & Schiffman Experimental [Page 41]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
[HAST86] Hastad, J., "On Using RSA With Low Exponents in a Public
Key Network," Advances in Cryptology-CRYPTO 95
Proceedings, Springer-Verlag, 1986.
[JOHN93] Johnson, D.B., Matyas, S.M., Le, A.V., Wilkins, J.D.,
"Design of the Commercial Data Masking Facility Data
Privacy Algorithm," Proceedings 1st ACM Conference on
Computer & Communications Security, November 1993,
Fairfax, VA., pp. 93-96.
[KRAW96b] Krawczyk, H. personal communication.
[LAI92] Lai, X. "On the Design and Security of Block Ciphers,"
ETH Series in Information Processing, v. 1, Konstanz:
Hartung-Gorre Verlag, 1992.
[PKCS-6] RSA Data Security, Inc. "Extended Certificate Syntax
Standard", PKCS-6, Nov 1, 1993.
[CMS] Housley, R., "Cryptographic Message Syntax", RFC 2630,
June 1999.
[RFC-822] Crocker, D., "Standard For The Format Of ARPA Internet
Text Messages", STD 11, RFC 822, August 1982.
[RFC-1319] Kaliski, B., "The MD2 Message-Digest Algorithm", RFC
1319, April 1992.
[RFC-1321] Rivest, R., "The MD5 Message-Digest Algorithm", RFC 1321,
April 1992.
[RFC-1421] Linn, J., "Privacy Enhancement for Internet Electronic
Mail: Part I: Message Encryption and Authentication
Procedures", RFC 1421, February 1993.
[RFC-1422] Kent, S., "Privacy Enhancement for Internet Electronic
Mail: Part II: Certificate-Based Key Management", RFC
1422, February 1993.
[RFC-1779] Kille, S., "A String Representation of Distinguished
Names", RFC 1779, March 1995.
[RFC-2045] Freed, N. and N. Borenstein, "Multipurpose Internet Mail
Extensions (MIME) Part One: Format of Internet Message
Bodies", RFC 2045, September 1993.
[RFC-1738] T. Berners-Lee, "Uniform Resource Locators (URLs)", RFC
1738, December 1994.
Rescorla & Schiffman Experimental [Page 42]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
[RFC-1847] Galvin, J., Murphy, S., Crocker, S., and N. Freed,
"Security Muliparts for MIME: Multipart/Signed and
Multipart/Encrypted", RFC 1847, October 1995.
[RFC-1848] Crocker, S., Freed, N., Galvin, J., and S. Murphy, "MIME
Object Security Services", RFC 1848, October 1995.
[RFC-1864] Myers, J. and M. Rose, "The Content-MD5 Header Field",
RFC 1864, October 1995.
[RFC-2616] Fielding, R., Gettys, J., Mogul, J., Frystyk, H.,
Masinter, L., Leach, P. and T. Berners-Lee, "Hypertext
Transfer Protocol -- HTTP/1.1" RFC 2616, June 1999.
[RFC-2617] Franks, J., Hallam-Baker, P., Hostetler, J., Leach, P.,
Luotonen, A. and L. Stewart, "HTTP Authentication: Basic
and Digest Access Authentication", RFC 2617, June 1999.
[RFC-2104] Krawczyk, H., Bellare, M. and R. Canetti, "HMAC: Keyed-
Hashing for Message Authentication", RFC 2104, February
1997.
[SHTML] Rescorla, E. and A. Schiffman, "Security Extensions For
HTML", RFC 2659, August 1999.
[VANO95] B. Prennel and P. van Oorschot, "On the security of two
MAC algorithms", to appear Eurocrypt'96.
[X509] CCITT Recommendation X.509 (1988), "The Directory -
Authentication Framework".
Security Considerations
This entire document is about security.
Rescorla & Schiffman Experimental [Page 43]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
Authors' Addresses
Eric Rescorla
RTFM, Inc.
30 Newell Road, #16
East Palo Alto, CA 94303
Phone: (650) 328-8631
EMail: ekr@rtfm.com
Allan M. Schiffman
SPYRUS/Terisa
5303 Betsy Ross Drive
Santa Clara, CA 95054
Phone: (408) 327-1901
EMail: ams@terisa.com
Rescorla & Schiffman Experimental [Page 44]
RFC 2660 The Secure HyperText Transfer Protocol August 1999
Copyright (C) The Internet Society (1999). All Rights Reserved.
This document and translations of it may be copied and furnished to
others, and derivative works that comment on or otherwise explain it
or assist in its implementation may be prepared, copied, published
and distributed, in whole or in part, without restriction of any
kind, provided that the above copyright notice and this paragraph are
included on all such copies and derivative works. However, this
document itself may not be modified in any way, such as by removing
the copyright notice or references to the Internet Society or other
Internet organizations, except as needed for the purpose of
developing Internet standards in which case the procedures for
copyrights defined in the Internet Standards process must be
followed, or as required to translate it into languages other than
English.
The limited permissions granted above are perpetual and will not be
revoked by the Internet Society or its successors or assigns.
This document and the information contained herein is provided on an
"AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING
TASK FORCE DISCLAIMS 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.
Acknowledgement
Funding for the RFC Editor function is currently provided by the
Internet Society.
Rescorla & Schiffman Experimental [Page 45]