| Network Working Group | J. Schaad |
| Internet-Draft | Soaring Hawk Consulting |
| Intended status: Standards Track | January 12, 2013 |
| Expires: July 16, 2013 |
Plasma Service Cryptographic Message Syntax (CMS) Processing
draft-schaad-plasma-cms-03
Secure MIME (S/MIME) defined a method of placing security labels on a Cryptographic Message Syntax (CMS) object. These labels are placed as part of the data signed and validated by the parties. This means that the message content is visible to the recipient prior to the label enforcement. A new model for enforcement of policy using a third party is described in RFC TBD [I.D-draft-freeman-plasma-requirements]. This is the Policy Augmented S/MIME (PLASMA) system. This document provides the details needed to implement the new Plasma model in the CMS infrastructure.
An additional benefit of using the Plasma module is that the server, based on policy, manages who has access to the message and how the keys are protected.
The document details how the client encryption and decryption processes are performed, defines how to construct the CMS recipient info structure, a new content to hold the data required for the Plasma server to store the keys and policy information. The document does not cover the protocol between the client and the Plasma policy enforcement server. One example of the client/server protocol can be found in RFC TBD [plasma-token].
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet- Drafts is at http:/⁠/⁠datatracker.ietf.org/⁠drafts/⁠current/⁠.
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."
This Internet-Draft will expire on July 16, 2013.
Copyright (c) 2013 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.
In the traditional S/MIME (Secure MIME) e-mail system, the sender of a message is responsible for determining the list of recipients for a message, obtaining a valid public or group key for each recipient, applying a security label to a message, and sending the message. The recipient of a message is responsible for the enforcement of any security labels on the message. While this system works in theory, in practice it has some difficulties that have led to problems in getting S/MIME mail widely deployed. This document is part of an effort to provide an alternative method of allocating the responsibilities above to different entities in an attempt to make the process work better.
In a Policy Augmented S/MIME (PLASMA) deployment of S/MIME, the sender of the message is still responsible for determining the list of recipients for the message and determining the security label to be applied to the message; however the Plasma server is now responsible for obtaining valid keys for recipients and checking the policy access for the recipients. The recipient is no longer responsible for enforcement of the policy as this is off-loaded to the Plasma server component. Doing this allows for the following changes in behavior of the system:
While this document describes the processes in terms of dealing with email system, a Plasma server can be used with any number of clients that need to protected content. Thus the same system could be used for protection of documents without having to specify in advance the individuals who should be able to open the document; it would just be allowed by the server based on the policy applied to the document.
This document is laid out as follows:
Some of the terms used in this document include:
When capitalized, 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].
Details on the model and the requirements for the Plasma system can be found in [I.D-draft-freeman-plasma-requirements].
In order for the Plasma system to function in CMS, a method needs to be chosen and described for how the CEK is to be protected and carried with the message, such that the recipient will be able to identified that this is a Plasma enabled message, know which Plasma server to contact and be able to get back the CEK needed to decrypt the message. Not all recipients of a message that has been encrypted using a Plasma server will need to contact the server in order to decrypt the message. There is nothing in what we are doing that prevents a message sender from building recipient info structures in a normal manner, except the possibly that the policy applied to the encrypted content could restrict it from happening. Additionally the Plasma server could return the standard recipient info structures to be added to the message for recipients, if it can pre-authorize them for access to the message and it can obtain the appropriate keying material.
There are two distinct methods that were considered for identifying a recipient info structure as being a Plasma enabled object. The first would be to define a new recipient info structure placed in the Other Recipient Info structure. The second option is to create a new key attribute placed in the KEK Recipient Info structure.
The use of a new recipient info structure would have been the easiest to document and implement, if most major CMS implementations had kept up with the latest versions. However it is known that several implementations stopped with RFC 2630 [RFC2630] and it was not until RFC 3369 [RFC3369] that the Other Recipient Info choice was introduced along with the language stating that implementations need to gracefully handle unimplemented alternatives in the recipient info choice. This means that if a new recipient info structure was defined and adopted, the mail message would fail decoding for many recipients, even for those recipients that had a key transfer or key agreement recipient info structure.
Given the current state of implementations, it was determined that the second method would be used as it will work with more implementations. After implementation it might be found that using the first method is the better way to go, in that case the decision can be re-visited.
The use of the KEKRecipientInfo type may seem to be a stretch at first, it was defined for those situations where a symmetric key had already been distributed and either a specific key or a specific transformation on the key was to be applied in order to decrypt the KEK value. However, the Plasma recipient info can be thought of as a strange way to do the transformation and thus it kind of fits into the model. It is in a structure that will be supported by the most basic CMS implementiation and it is easy for client implementations to make the determination of a Plasma recipient info by looking at the OID for the other key attribute structure.
A recipient info structure as defined in this document MUST be created by a Plasma server and MUST NOT be created by client software. A protocol such as the one in RFC TBD1 [plasma-token] is used to transport the recipient info structure between the client and the server.
For the convenience of the reader we include the KEKRecipientInfo structure pieces here (copied from [CMS-ASN]):
KEKRecipientInfo ::= SEQUENCE {
version CMSVersion, -- always set to 4
kekid KEKIdentifier,
keyEncryptionAlgorithm KeyEncryptionAlgorithmIdentifier,
encryptedKey EncryptedKey }
KEKIdentifier ::= SEQUENCE {
keyIdentifier OCTET STRING,
date GeneralizedTime OPTIONAL,
other OtherKeyAttribute OPTIONAL }
OtherKeyAttribute ::= SEQUENCE {
keyAttrId KEY-ATTRIBUTE.
&id({SupportedKeyAttributes}),
keyAttr KEY-ATTRIBUTE.
&Type({SupportedKeyAttributes}{@keyAttrId})}
For a Plasma KEKRecipientInfo structure, the fields are filled in as follows:
NOTE: As can be seen above, we have bent the model so much that it no longer really makes any sense to try and keep this as a KEK algorithm unless we completely re-design it. The fact that we now always work with CEK values rather than KEK values means that we no longer have a key value or a key encryption algorithm. This means that we need to step back and re-examine the entire question of how we do the encoding. At this point in time I am leading to converting to say this is a key transport algorithm that uses a random value for the SKI. Clients would then need to look at both the key identifier and the algorithm in order to determine if it understands things. The other option is to bite the bullet and move to an Other Key Info structure.
The PLASMA Other Key Attribute functions as the lock box for the KEK used in encrypting the CEK. In addition to the KEK, the lock box also contains the information that is needed by the Plasma Server to know the policy(s) applied to the encrypted data and possible parameters for the policy and for the client to validate that the lock box applies to the encrypted content.
The relevant section from the ASN.1 module which contains the content is:
--
-- New Other Key Attribute value for Plasma
-- This structure holds the encrypted KEK value for the server
-- and other signed attributes used by the client for checking
-- the structure applies in this case
--
keyatt-plasma-kek KEY-ATTRIBUTE ::= {
SignedData IDENTIFIED BY id-keyatt-plasma-token
}
id-keyatt-plasma-token OBJECT IDENTIFIER ::= {iso(1) member-body(2)
us(840) rsadsi(113549) pkcs(1) pkcs9(9) TBD2 }
We define a new KEY-ATTRIBUTE called keyatt-plasma-kek. This attribute is identified by the id-keyatt-plasma-token OID. The data structure that is associated with this key attribute is the CMS SignedData structure. The CMS SignedData structure is used directly without a CMS ContentInfo structure wrapping it.
The SignedData structure fields are filled as follows (some less significant fields are omitted):
When creating the recipient info structures for the AuthEnvelopedData structure, there will normally only need to be a single entry in the sequence as the only entity that needs to decrypt the PLASMA Lockbox is the Plasma Service. In the event that the service is distributed over multiple servers then multiple lock boxes may need to be created. One of the implications of the fact that the originator of the message is the only recipient is that, although the signing key needs to be contained in a certificate, there is no corresponding requirement that the encryption key needs to be in a certificate. Instead of using a certificate, a subject key identifier that is meaningful only to the Plasma Service can be used.
There are a number of circumstances that a Plasma server would apply multiple signatures to a single lockbox. These circumstances include during key rollover while a certificate is approaching expiration, esp. if there is going to be a change in the trust anchor being used. Another circumstance would be if a new signature algorithm is being rolled out, having the old and the new algorithm on the message during the rollout period increases the chances that the signature can be validated. In these circumstances, the multiple signature attribute defined in RFC 5752 [RFC5752] allows for a client to determine that a signature has been removed which might be attempted as part of an attack to use a more insecure algorithm.
The inner content type for a PLASMA Other Key Attribute is a PLASMA Lockbox. This content is contained in an encrypted CMS object which is encrypted by and for the Plasma server itself, as such the contents and structure is known just to the Plasma server.
The content type is designed so that the Plasma server does not need to keep any state dealing with a message on the server itself. This allows for minimal information to be kept on the server, it only needs the state of it's current transactions, and the message can be processed by any of a number of servers without needing to replicate state about the message between them.
The relevant section from the ASN.1 module which defines this content is:
--
-- PLASMA Content Type
--
ct-plasma-LockBox CONTENT-TYPE ::= {
TYPE PLASMA-LockBox
IDENTIFIED BY id-ct-plasma-LockBox
}
id-ct-plasma-LockBox OBJECT IDENTIFIER ::= {iso(1) member-body(2)
us(840) rsadsi(113549) pkcs(1) pkcs7(7) TBD1}
PLASMA-LockBox ::= SEQUENCE {
policy OCTET STRING,
keyList KeyList,
attrList AttributeList OPTIONAL
}
KeyList ::= SEQUENCE {
namedRecipients [0] SEQUENCE SIZE (1..MAX) OF
NamedRecipient OPTIONAL,
defaultRecipients [1] SEQUENCE SIZE (1..MAX) OF
OneCek OPTIONAL,
...
}
(WITH COMPONENTS {
...,
namedRecipients PRESENT
} |
WITH COMPONENTS {
...,
defaultRecipients PRESENT
})
NamedRecipient ::= SEQUENCE {
recipientName UTF8String, -- name of the recipient
keyPolicy [0] OCTET STRING OPTIONAL,
keyIdentifier OCTET STRING OPTIONAL,
keyValue RecipientInfo,
...
}
OneCek ::= SEQUENCE {
keyPolicy [0] OCTET STRING OPTIONAL,
keyIdentifier [1] OCTET STRING OPTIONAL,
keyValue OCTET STRING,
...
}
AttributeList ::= SEQUENCE SIZE (1..MAX) OF
SingleAttribute{{PlasmaLockboxAttributes}}
PlasmaLockboxAttributes ATTRIBUTE ::= { aa-plasma-AuditTrailIdentifier |
aa-plasma-SignerInfo | aa-plasma-Xacml-Attribute, ... }
PlasmaSignedAttributes ATTRIBUTE ::= {
aa-plasma-url | aa-plasma-econtent-hash
}
In the above ASN.1, the following items are defined:
This structure is tagged as extensible; this was done because there may be a need to add additional fields such as other name types in the future.
This structure is tagged as extensible; this was done because there may be a need to add additional fields such as other name types in the future.
The recipientName field of the NamedRecipient structure is designed so that a client can build a CMS recipient info structure targeted to a specific recipient. In order for the Plasma server to know which of these named recipient structure to return it requires that the sender identify the recipient for the CMS recipient info structure and that the recipient identify itself so that the Plasma server can find the correct structure. We are using Email names in the form of internationalized RFC 5321 [RFC5321] address names. There are a number of issues that are associated with the use of this name form for comparison purposes. As stated in Section 2.3.11 of RFC 5321, [RFC6530]. The server that holds the mailbox can have a consistent rule for normalization, but a Plasma server in separate domain may not know the appropriate rules to use.
Plasma servers SHOULD do the following when comparing the Email addresses found in the recipientName field:
It is required that the name of the Plasma Server be securely communicated to the message recipient. For this purpose a URL is used as this can communicate both a server name as well as additional parameters that can be used to identify a specific service on the server.
The relevant section from the ASN.1 module for this attribute is:
--
-- Define the Signed Attribute to carry the
-- Email Policy Server URL
--
-- This attribute is added to the SignedAttributSet set of
-- attributes in [CMS-ASN]
--
aa-plasma-url ATTRIBUTE ::= {
TYPE UTF8String IDENTIFIED BY id-aa-plasma-url
}
id-aa-plasma-url OBJECT IDENTIFIER ::= { iso(1) member-body(2)
us(840) rsadsi(113549) pkcs(1) pkcs9(9) TBD3}
From this we can see the following:
The attribute structure defined for signed attributes allows for multiple values to be carried in a single attribute. For this attribute there MUST be at least one value. If there is more than one value, each value MUST be unique. Multiple values are allowed as there can be multiple Plasma servers that can be used to evaluate the policy. Since the URLs will be sorted during encoding, the order of URLs does not indicate any order of priority, any of the values may be used.
This attribute is only included in a SignedData object by a Plasma Server. There are no processing rules for the sender of a message. The processing rules for a recipient can be found in Section 5.
For privacy reasons, it is highly desirable that the recipient of a message can validate that the Plasma lock box embedded in a message is associated with encrypted data it is attached to. For this reason, in addition to the requirement that a recipient validate the signature of the Plasma server over the lock box, a new attribute is defined which contains the hash of the encrypted content.
--
-- Define the Signed Attribute to carry the
-- hash of encrypted data
--
-- This attribute is added to the SignedAttributeSet set of
-- attributes in [CMS-ASN]
--
aa-plasma-econtent-hash ATTRIBUTE ::= {
TYPE HashValue IDENTIFIED BY id-aa-plasma-econtent-hash
}
id-aa-plasma-econtent-hash OBJECT IDENTIFIER ::= {iso(1) member-body(2)
us(840) rsadsi(113549) pkcs(1) pkcs9(9) TBD4}
HashValue ::= SEQUENCE {
hashAlgorithm DigestAlgorithmIdentifier,
hashValue OCTET STRING
}
The above ASN.1 fragment defines the following items:
The hash is computed over the encrypted content, without including any of the ASN.1 wrapping around the content. Thus this value does not cover the content type identifier, the encryption algorithm and parameters or any authenticated attributes for AEAD algorithms.
The Audit Trail Identifier attribute allows for a unique and persistent identifier to be carried as part of a Plasma Lockbox message. This identifier can then be used by Plasma servers when creating log entries in the audit trail to designate a single Plasma message. This use of a single identifier allows for better correlation to occur by auditors, however as the identifier is hidden from all viewers except the Plasma server, the message itself is not locatable from the log entries.
--
-- Attribute to hold an Audit Trail Identifier
--
aa-plasma-AuditTrailIdentifier ATTRIBUTE ::= {
TYPE OCTET STRING IDENTIFIED BY id-aa-plasma-Audit-Trail-Identifier
}
id-aa-plasma-Audit-Trail-Identifier OBJECT IDENTIFIER ::= {
iso(1) member-body(2)
us(840) rsadsi(113549) pkcs(1) pkcs9(9) TBD5}
The relevant section from the ASN.1 module which defines this attribute is:
In this ASN.1, the following items are defined:
The use of OCTET STRING for the content allows for the greatest flexibility for Plasma Servers in devising the value to use. The content of the Audit Tail Identifier is intended to be an opaque value to all entities, all Plasma servers MUST be able to ignore how the value is structured.
Some policies require that the inner content of an encrypted message be signed as well. However the encrypted data stream of the message is not provided to the Plasma server for it to verify that it was done successfully. The only places to check is in the audit trail for the message and/or to allow the client to do the check that the signature is present. This attribute provides a location for the Plasma server to place the provided CMS SignerInfo structure(s) provided by the client to be carried with the message. The server can then push the structure(s) to the client and the client can validate that the actual signatures on the message match the signatures provided by the server. All servers MUST be able to parse this attribute and convert it to an appropriate XACML attribute to return to clients.
--
-- Attribute to hold a SignerInfo structure
--
aa-plasma-SignerInfo ATTRIBUTE ::= {
TYPE SignerInfo IDENTIFIED BY id-aa-plasma-signerInfo
}
id-aa-plasma-signerInfo OBJECT IDENTIFIER ::= {iso(1) member-body(2)
us(840) rsadsi(113549) pkcs(1) pkcs9(9) TBD6}
The relevant section from the ASN.1 module which defines this attribute is:
In this ASN.1, the following items are defined:
There can be one or more attribute values in the attribute set. Each of the values is to be treated independently and returned to the client. The values may be returned in a single Attributes XML element.
Many times Plasma servers be in situation where they will need to return various values to the clients. These decisions will frequently be taken by the originating Plasma server, since it may be the only one that has the data to be returned. This attribute allows for any data to be carried in the form of an XACML attribute XML structure. Since the content is an XACML attribute, it can be pushed to the client without the server needing to understand or evaluate the content being presented. The Signer Info attribute presented in the previous section could have been implemented using this attribute rather than defining it's own attribute, however the space savings was deemed sufficient to justify the creation of the new attribute.
--
-- Attribute to hold an arbitrary XACML XML attribute
-- structure
--
aa-plasma-Xacml-Attribute ATTRIBUTE ::= {
TYPE OCTET STRING IDENTIFIED BY id-aa-plasma-Xacml-Attribute
}
id-aa-plasma-Xacml-Attribute OBJECT IDENTIFIER ::= {iso(1) member-body(2)
us(840) rsadsi(113549) pkcs(1) pkcs9(9) TBD7}
The relevant section from the ASN.1 module which defines this attribute is:
In this ASN.1, the following items are defined:
There can be one or more attributes values associated with the attribute set. Each of the values is to be treated independently and returned as separate items to the client.
The data type is an OCTET STRING to allow for alternate XML encodings to be used. All servers MUST be able to deal with a UTF8 string XML encoding of the policy in this location. See Appendix B for alternate encoding methods. If a server cannot understand the encoding presented, the server MUST fail processing of the lockbox.
This is the set of processing steps that a sender needs to do (the order of the steps is not normative):
The Sender Agent SHOULD validate all of the signatures if more than one signature exists.
When looking at the validation steps that are given here, the results need to be the same but the order that the steps are taken can be different. As an example, it can make sense to do the policy check in step [r_policy] before doing the signature validation as this would not require any network access.
This is the set of processing that the recipient needs to do:
In some circumstances a message recipient may be permitted to read a message sent under a certan policy, but it not permitted to send a message for that policy. In the event that a complex policy is used the recipient may be permitted to read under one policy, but not have any rights under a second policy. In both of these case a recipient of a message would be unable to either reply or forward a message using the same policy as they received it under. For this reason, the protocol used to communicate with the Plasma server will frequently return a single purpose policy that permits a recipient to reply to a message using the same policy as the original message.
The SMIMECapabilities attribute was defined by S/MIME in [SMIME-MSG] so that the abilities of a client can be advertised to the recipients of an S/MIME message. This information can be advertised either directly in an S/MIME message sent from a client to a recipient, or more indirectly by publishing the information in an LDAP directory [RFC4262].
A new S/MIME capability is defined by this document so that a client can advertise to others that it understands how to deal with Plasma recipient information. The ASN.1 for this attribute is:
--
-- Create an S/MIME capability for advertising that
-- a client can understand the PLASMA recipient info
-- structure information
--
cap-Plasma-RecipientInfo SMIME-CAPS ::= {
IDENTIFIED BY id-cap-plasma-recipientInfo
}
id-cap-plasma-recipientInfo OBJECT IDENTIFIER ::= {
id-cap TBD5
}
We define a new SMIME-CAPS object called cap-Plasma-RecipentInfo. This attribute is identified by the the OID id-cap-plasma-recipientInfo and has no data structure associated with it. When encoded as an S/MIME capability the parameters MUST to be absent and not NULL.
Servers MUST implement AES-GCC-128 ([RFC5084]) for the content encryption algorithm in section 3.1. Other authenticated encryption algorithms MAY be implemented.
Servers MUST implement RSA v1.5 as a key transport algorithm for lockboxes created in section 3.1 and for pre-authenticated lock boxes returned in step 8 of section 4.1. Servers SHOULD implement RSA OAEP as a key transport algorithm in the same locations. Other key transport algorithms MAY be implemented.
Servers MUST implement EC-DH as a key agreement algorithm for lockboxes created in section 3.1 and for pre-authenticated lock boxes returned in step 8 of section 4.1. Servers MAY implement other key agreement algorithms.
Servers MUST implement the RSA v1.5 signature algorithm with SHA-256 and SHA-512. Servers MUST implement the EC-DSA signature algorithm with SHA-256 and SHA-512 for producing signature on the Plasma lock box. Other signature algorithms MAY be implemented as well.
Clients MUST implement the mandatory algorithms defined for S/MIME [SMIME-MSG] for the lock boxes created in step 7 and transmitted to the server in step 8 of Section 4. Other algorithms MAY be implemented.
Clients MUST implement SHA-256 and SHA-512 for computation of the Plasma Encrypted Content Hash in section 3.4. Other algorithms MAY be implemented, but doing so can cause clients that do not implement this algorithm to not attempt to read the message.
When verifying signatures on the Plasma lock boxes, clients MUST be able to verify the RSA v1.5 signature algorithm with SHA-256 and SHA-512. Clients MUST also be able to verify the EC-DSA signature algorithm with SHA-256 and SHA-512 signature algorithm. Clients MAY be able to verify other signature algorithms.
Man in the middle attack on the protocol from the sending agent to the email policy server.
Man in the middle attack on the protocol from the receiving agent to the email policy server.
Still more expansion....
The hash computed for the Plasma Encrypted Content Hash attribute has different security concerns that a hash used for signature computation. This hash value is used to get a degree of assurance that the encrypted content is associated with Plasma lock box. In the event that a collision exists, then the client will go and talk to the server to get a content encryption key when that key will not successfully decrypt the content. However this does not affect the privacy of the client as the client's decision to talk to the server is based on the URL(s) of the server and the validation of the server's signature. This means that an attacker that substitutes an encrypted content needs not only to have the hash of the encrypted content be correct, but the decrypted content needs to make sense. In order for an attacker to have the client talk to it, it needs to attack the certificates or signature produced on the lock box and not the encrypted content itself.
All of the object identifiers defined by this document are done so under the existing S/MIME Object Identifier arc. No actions by IANA are required for this document.
| [RFC3369] | Housley, R., "Cryptographic Message Syntax (CMS)", RFC 3369, August 2002. |
| [RFC2630] | Housley, R., "Cryptographic Message Syntax", RFC 2630, June 1999. |
| [RFC4262] | Santesson, S., "X.509 Certificate Extension for Secure/Multipurpose Internet Mail Extensions (S/MIME) Capabilities", RFC 4262, December 2005. |
| [RFC5084] | Housley, R., "Using AES-CCM and AES-GCM Authenticated Encryption in the Cryptographic Message Syntax (CMS)", RFC 5084, November 2007. |
| [plasma-token] | Schaad, J., "Plasma Service Trust Processing", Work in progress draft-schaad-plasma-service, March 2012. |
| [XACML] | Rissanen, E, "eXtensible Access Control Markup Language (XACML) Version 3.0", OASIS Standard xacml-201008, August 2010. |
| [EXI] | Kamiya, T. and J. Schneider, "Efficient XML Interchange (EXI) Format 1.0", World Wide Web Consortium CR CR-exi-20091208, December 2009. |
PolicySMime -- TBD Get a module number --
DEFINITIONS IMPLICIT TAGS ::=
BEGIN
IMPORTS
-- Cryptographic Message Syntax (CMS) [RFC5652]
CONTENT-TYPE, RecipientInfo, KEY-ATTRIBUTE, SignedData,
DigestAlgorithmIdentifier, SignerInfo
FROM CryptographicMessageSyntax-2010
{ iso(1) member-body(2) us(840) rsadsi(113549)
pkcs(1) pkcs-9(9) smime(16) modules(0) id-mod-cms-2009(58) }
-- Common PKIX structures [RFC5912]
SMIME-CAPS
FROM AlgorithmInformation-2009
{ iso(1) identified-organization(3) dod(6) internet(1)
security(5) mechanisms(5) pkix(7) id-mod(0)
id-mod-algorithmInformation-02(58)}
ATTRIBUTE, SingleAttribute{}
FROM PKIX-CommonTypes-2009
{ iso(1) identified-organization(3) dod(6) internet(1)
security(5) mechanisms(5) pkix(7) id-mod(0)
id-mod-pkixCommon-02(57) }
ESSSecurityLabel
FROM ExtendedSecurityServices-2009
{ iso(1) member-body(2) us(840) rsadsi(113549) pkcs(1) pkcs-9(9)
smime(16) modules(0) id-mod-ess-2006-02(42) }
id-cap
FROM SecureMimeMessage
{ iso(1) member-body(2) us(840) rsadsi(113549) pkcs(1) pkcs-9(9)
smime(16) modules(0) id-mod-msg-v3dot1-02(39) }
;
--
-- PLASMA Content Type
--
ct-plasma-LockBox CONTENT-TYPE ::= {
TYPE PLASMA-LockBox
IDENTIFIED BY id-ct-plasma-LockBox
}
id-ct-plasma-LockBox OBJECT IDENTIFIER ::= {iso(1) member-body(2)
us(840) rsadsi(113549) pkcs(1) pkcs7(7) TBD1}
PLASMA-LockBox ::= SEQUENCE {
policy OCTET STRING,
keyList KeyList,
attrList AttributeList OPTIONAL
}
KeyList ::= SEQUENCE {
namedRecipients [0] SEQUENCE SIZE (1..MAX) OF
NamedRecipient OPTIONAL,
defaultRecipients [1] SEQUENCE SIZE (1..MAX) OF
OneCek OPTIONAL,
...
}
(WITH COMPONENTS {
...,
namedRecipients PRESENT
} |
WITH COMPONENTS {
...,
defaultRecipients PRESENT
})
NamedRecipient ::= SEQUENCE {
recipientName UTF8String, -- name of the recipient
keyPolicy [0] OCTET STRING OPTIONAL,
keyIdentifier OCTET STRING OPTIONAL,
keyValue RecipientInfo,
...
}
OneCek ::= SEQUENCE {
keyPolicy [0] OCTET STRING OPTIONAL,
keyIdentifier [1] OCTET STRING OPTIONAL,
keyValue OCTET STRING,
...
}
AttributeList ::= SEQUENCE SIZE (1..MAX) OF
SingleAttribute{{PlasmaLockboxAttributes}}
PlasmaLockboxAttributes ATTRIBUTE ::= { aa-plasma-AuditTrailIdentifier |
aa-plasma-SignerInfo | aa-plasma-Xacml-Attribute, ... }
PlasmaSignedAttributes ATTRIBUTE ::= {
aa-plasma-url | aa-plasma-econtent-hash
}
--
-- New Other Key Attribute value for Plasma
-- This structure holds the encrypted KEK value for the server
-- and other signed attributes used by the client for checking
-- the structure applies in this case
--
keyatt-plasma-kek KEY-ATTRIBUTE ::= {
SignedData IDENTIFIED BY id-keyatt-plasma-token
}
id-keyatt-plasma-token OBJECT IDENTIFIER ::= {iso(1) member-body(2)
us(840) rsadsi(113549) pkcs(1) pkcs9(9) TBD2 }
--
-- Define the Signed Attribute to carry the
-- Email Policy Server URL
--
-- This attribute is added to the SignedAttributSet set of
-- attributes in [CMS-ASN]
--
aa-plasma-url ATTRIBUTE ::= {
TYPE UTF8String IDENTIFIED BY id-aa-plasma-url
}
id-aa-plasma-url OBJECT IDENTIFIER ::= { iso(1) member-body(2)
us(840) rsadsi(113549) pkcs(1) pkcs9(9) TBD3}
--
-- Define the Signed Attribute to carry the
-- hash of encrypted data
--
-- This attribute is added to the SignedAttributeSet set of
-- attributes in [CMS-ASN]
--
aa-plasma-econtent-hash ATTRIBUTE ::= {
TYPE HashValue IDENTIFIED BY id-aa-plasma-econtent-hash
}
id-aa-plasma-econtent-hash OBJECT IDENTIFIER ::= {iso(1) member-body(2)
us(840) rsadsi(113549) pkcs(1) pkcs9(9) TBD4}
HashValue ::= SEQUENCE {
hashAlgorithm DigestAlgorithmIdentifier,
hashValue OCTET STRING
}
--
-- Create an S/MIME capability for advertising that
-- a client can understand the PLASMA recipient info
-- structure information
--
cap-Plasma-RecipientInfo SMIME-CAPS ::= {
IDENTIFIED BY id-cap-plasma-recipientInfo
}
id-cap-plasma-recipientInfo OBJECT IDENTIFIER ::= {
id-cap TBD5
}
--
-- Attribute to hold an Audit Trail Identifier
--
aa-plasma-AuditTrailIdentifier ATTRIBUTE ::= {
TYPE OCTET STRING IDENTIFIED BY id-aa-plasma-Audit-Trail-Identifier
}
id-aa-plasma-Audit-Trail-Identifier OBJECT IDENTIFIER ::= {
iso(1) member-body(2)
us(840) rsadsi(113549) pkcs(1) pkcs9(9) TBD5}
--
-- Attribute to hold a SignerInfo structure
--
aa-plasma-SignerInfo ATTRIBUTE ::= {
TYPE SignerInfo IDENTIFIED BY id-aa-plasma-signerInfo
}
id-aa-plasma-signerInfo OBJECT IDENTIFIER ::= {iso(1) member-body(2)
us(840) rsadsi(113549) pkcs(1) pkcs9(9) TBD6}
--
-- Attribute to hold an arbitrary XACML XML attribute
-- structure
--
aa-plasma-Xacml-Attribute ATTRIBUTE ::= {
TYPE OCTET STRING IDENTIFIED BY id-aa-plasma-Xacml-Attribute
}
id-aa-plasma-Xacml-Attribute OBJECT IDENTIFIER ::= {iso(1) member-body(2)
us(840) rsadsi(113549) pkcs(1) pkcs9(9) TBD7}
END
This appendix is informative.
The fields for encoding a policy expression is an ASN.1 OCTET STRING. This field type was chosen so that servers would have the widest choice of methods to encode the policy expressions. For stand alone servers, the only issue is that the server will be able to correctly extract and use the policy expression, as such it can be kept in XML or converted into a format that is more natural to the policy evaluation engine used by the server. When one wants to use multiple servers, then all of the servers involved need to be able to use the encoded format(s) and re-map them into the internal versions that are used locally. This is far more complicated when the servers are hosted by different organizations that might be using different local policy evaluation engines.
It is RECOMMENDED that what ever encoding method is used normally, a provision exist for the XML version of the policy string as presented in RFC XXX [plasma-token] exist without change. Doing so allows for a single common format to be shared among all Plasma servers independent of the organization providing the server and the one operating the server. The server will be able to determine the set of other servers that will be able to process the content, as the server must be configured with that information in order to create the appropriate lock boxes for the other servers to access the encrypted content.
There are two different methods that exist where the XML encoding can be compressed before placing it into the OCTET STRING. The first would be to use the Efficient XML Interchange (EXI) Format documented in [EXI]. A second method would be to use the standard DEFLATE algorithm either with or without a pre-defined library.
A possible method of encoding would to be use the first byte to identify the encoding technique, reserving 0x3C for vanilla XML strings. Following bytes could be used to determine which pre-defined table was used and then the compressed encoding.