Network Working Group C. Adams Request for Comments: 4210 University of Ottawa Obsoletes: 2510 S. Farrell Category: Standards Track Trinity College Dublin T. Kause SSH T. Mononen SafeNet September 2005 Internet X.509 Public Key Infrastructure Certificate Management Protocol (CMP) Status of This Memo This document specifies an Internet standards track protocol for the Internet community, and requests discussion and suggestions for improvements. Please refer to the current edition of the "Internet Official Protocol Standards" (STD 1) for the standardization state and status of this protocol. Distribution of this memo is unlimited. Copyright Notice Copyright (C) The Internet Society (2005). Abstract This document describes the Internet X.509 Public Key Infrastructure (PKI) Certificate Management Protocol (CMP). Protocol messages are defined for X.509v3 certificate creation and management. CMP provides on-line interactions between PKI components, including an exchange between a Certification Authority (CA) and a client system. Table of Contents 1. Introduction ....................................................5 2. Requirements ....................................................5 3. PKI Management Overview .........................................5 3.1. PKI Management Model .......................................6 3.1.1. Definitions of PKI Entities .........................6 3.1.1.1. Subjects and End Entities ..................6 3.1.1.2. Certification Authority ....................7 3.1.1.3. Registration Authority .....................7 3.1.2. PKI Management Requirements .........................8 3.1.3. PKI Management Operations ..........................10 4. Assumptions and Restrictions ...................................14 4.1. End Entity Initialization .................................14 Adams, et al. Standards Track [Page 1] RFC 4210 CMP September 2005 4.2. Initial Registration/Certification ........................14 4.2.1. Criteria Used ......................................15 4.2.1.1. Initiation of Registration/Certification ..15 4.2.1.2. End Entity Message Origin Authentication ..15 4.2.1.3. Location of Key Generation ................15 4.2.1.4. Confirmation of Successful Certification ..16 4.2.2. Mandatory Schemes ..................................16 4.2.2.1. Centralized Scheme ........................16 4.2.2.2. Basic Authenticated Scheme ................17 4.3. Proof-of-Possession (POP) of Private Key ..................17 4.3.1. Signature Keys .....................................18 4.3.2. Encryption Keys ....................................18 4.3.3. Key Agreement Keys .................................19 4.4. Root CA Key Update ........................................19 4.4.1. CA Operator Actions ................................20 4.4.2. Verifying Certificates .............................21 4.4.2.1. Verification in Cases 1, 4, 5, and 8 ......22 4.4.2.2. Verification in Case 2 ....................22 4.4.2.3. Verification in Case 3 ....................23 4.4.2.4. Failure of Verification in Case 6 .........23 4.4.2.5. Failure of Verification in Case 7 .........23 4.4.3. Revocation - Change of CA Key ......................23 5. Data Structures ................................................24 5.1. Overall PKI Message .......................................24 5.1.1. PKI Message Header .................................24 5.1.1.1. ImplicitConfirm ...........................27 5.1.1.2. ConfirmWaitTime ...........................27 5.1.2. PKI Message Body ...................................27 5.1.3. PKI Message Protection .............................28 5.1.3.1. Shared Secret Information .................29 5.1.3.2. DH Key Pairs ..............................30 5.1.3.3. Signature .................................30 5.1.3.4. Multiple Protection .......................30 5.2. Common Data Structures ....................................31 5.2.1. Requested Certificate Contents .....................31 5.2.2. Encrypted Values ...................................31 5.2.3. Status codes and Failure Information for PKI Messages .......................................32 5.2.4. Certificate Identification .........................33 5.2.5. Out-of-band root CA Public Key .....................33 5.2.6. Archive Options ....................................34 5.2.7. Publication Information ............................34 5.2.8. Proof-of-Possession Structures .....................34 5.2.8.1. Inclusion of the Private Key ..............35 5.2.8.2. Indirect Method ...........................35 5.2.8.3. Challenge-Response Protocol ...............35 5.2.8.4. Summary of PoP Options ....................37 Adams, et al. Standards Track [Page 2] RFC 4210 CMP September 2005 5.3. Operation-Specific Data Structures ........................38 5.3.1. Initialization Request .............................38 5.3.2. Initialization Response ............................39 5.3.3. Certification Request ..............................39 5.3.4. Certification Response .............................39 5.3.5. Key Update Request Content .........................40 5.3.6. Key Update Response Content ........................41 5.3.7. Key Recovery Request Content .......................41 5.3.8. Key Recovery Response Content ......................41 5.3.9. Revocation Request Content .........................41 5.3.10. Revocation Response Content .......................42 5.3.11. Cross Certification Request Content ...............42 5.3.12. Cross Certification Response Content ..............42 5.3.13. CA Key Update Announcement Content ................42 5.3.14. Certificate Announcement ..........................43 5.3.15. Revocation Announcement ...........................43 5.3.16. CRL Announcement ..................................43 5.3.17. PKI Confirmation Content ..........................43 5.3.18. Certificate Confirmation Content ..................44 5.3.19. PKI General Message Content .......................44 5.3.19.1. CA Protocol Encryption Certificate .......44 5.3.19.2. Signing Key Pair Types ...................45 5.3.19.3. Encryption/Key Agreement Key Pair Types ..45 5.3.19.4. Preferred Symmetric Algorithm ............45 5.3.19.5. Updated CA Key Pair ......................45 5.3.19.6. CRL ......................................46 5.3.19.7. Unsupported Object Identifiers ...........46 5.3.19.8. Key Pair Parameters ......................46 5.3.19.9. Revocation Passphrase ....................46 5.3.19.10. ImplicitConfirm .........................46 5.3.19.11. ConfirmWaitTime .........................47 5.3.19.12. Original PKIMessage .....................47 5.3.19.13. Supported Language Tags .................47 5.3.20. PKI General Response Content ......................47 5.3.21. Error Message Content .............................47 5.3.22. Polling Request and Response ......................48 6. Mandatory PKI Management Functions .............................51 6.1. Root CA Initialization ....................................51 6.2. Root CA Key Update ........................................51 6.3. Subordinate CA Initialization .............................51 6.4. CRL production ............................................52 6.5. PKI Information Request ...................................52 6.6. Cross Certification .......................................52 6.6.1. One-Way Request-Response Scheme: ...................52 6.7. End Entity Initialization .................................54 6.7.1. Acquisition of PKI Information .....................54 6.7.2. Out-of-Band Verification of Root-CA Key ............55 6.8. Certificate Request .......................................55 Adams, et al. Standards Track [Page 3] RFC 4210 CMP September 2005 6.9. Key Update ................................................55 7. Version Negotiation ............................................56 7.1. Supporting RFC 2510 Implementations .......................56 7.1.1. Clients Talking to RFC 2510 Servers ................56 7.1.2. Servers Receiving Version cmp1999 PKIMessages ......57 8. Security Considerations ........................................57 8.1. Proof-Of-Possession with a Decryption Key .................57 8.2. Proof-Of-Possession by Exposing the Private Key ...........57 8.3. Attack Against Diffie-Hellman Key Exchange ................57 9. IANA Considerations ............................................58 Normative References ..............................................58 Informative References ............................................59 A. Reasons for the Presence of RAs ................................61 B. The Use of Revocation Passphrase ...............................61 C. Request Message Behavioral Clarifications ......................63 D. PKI Management Message Profiles (REQUIRED) .....................65 D.1. General Rules for Interpretation of These Profiles ........65 D.2. Algorithm Use Profile .....................................66 D.3. Proof-of-Possession Profile ...............................68 D.4. Initial Registration/Certification (Basic Authenticated Scheme) .....................................68 D.5. Certificate Request .......................................74 D.6. Key Update Request ........................................75 E. PKI Management Message Profiles (OPTIONAL) .....................75 E.1. General Rules for Interpretation of These Profiles ........76 E.2. Algorithm Use Profile .....................................76 E.3. Self-Signed Certificates ..................................76 E.4. Root CA Key Update ........................................77 E.5. PKI Information Request/Response ..........................77 E.6. Cross Certification Request/Response (1-way) ..............79 E.7. In-Band Initialization Using External Identity Certificate ..............................................82 F. Compilable ASN.1 Definitions ...................................83 G. Acknowledgements ...............................................93 Adams, et al. Standards Track [Page 4] RFC 4210 CMP September 2005 1. Introduction This document describes the Internet X.509 Public Key Infrastructure (PKI) Certificate Management Protocol (CMP). Protocol messages are defined for certificate creation and management. The term "certificate" in this document refers to an X.509v3 Certificate as defined in [X509]. This specification obsoletes RFC 2510. This specification differs from RFC 2510 in the following areas: The PKI management message profile section is split to two appendices: the required profile and the optional profile. Some of the formerly mandatory functionality is moved to the optional profile. The message confirmation mechanism has changed substantially. A new polling mechanism is introduced, deprecating the old polling method at the CMP transport level. The CMP transport protocol issues are handled in a separate document [CMPtrans], thus the Transports section is removed. A new implicit confirmation method is introduced to reduce the number of protocol messages exchanged in a transaction. The new specification contains some less prominent protocol enhancements and improved explanatory text on several issues. 2. Requirements The key words "MUST", "MUST NOT", "REQUIRED", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document (in uppercase, as shown) are to be interpreted as described in [RFC2119]. 3. PKI Management Overview The PKI must be structured to be consistent with the types of individuals who must administer it. Providing such administrators with unbounded choices not only complicates the software required, but also increases the chances that a subtle mistake by an administrator or software developer will result in broader compromise. Similarly, restricting administrators with cumbersome mechanisms will cause them not to use the PKI. Adams, et al. Standards Track [Page 5] RFC 4210 CMP September 2005 Management protocols are REQUIRED to support on-line interactions between Public Key Infrastructure (PKI) components. For example, a management protocol might be used between a Certification Authority (CA) and a client system with which a key pair is associated, or between two CAs that issue cross-certificates for each other. 3.1. PKI Management Model Before specifying particular message formats and procedures, we first define the entities involved in PKI management and their interactions (in terms of the PKI management functions required). We then group these functions in order to accommodate different identifiable types of end entities. 3.1.1. Definitions of PKI Entities The entities involved in PKI management include the end entity (i.e., the entity to whom the certificate is issued) and the certification authority (i.e., the entity that issues the certificate). A registration authority MAY also be involved in PKI management. 3.1.1.1. Subjects and End Entities The term "subject" is used here to refer to the entity to whom the certificate is issued, typically named in the subject or subjectAltName field of a certificate. When we wish to distinguish the tools and/or software used by the subject (e.g., a local certificate management module), we will use the term "subject equipment". In general, the term "end entity" (EE), rather than "subject", is preferred in order to avoid confusion with the field name. It is important to note that the end entities here will include not only human users of applications, but also applications themselves (e.g., for IP security). This factor influences the protocols that the PKI management operations use; for example, application software is far more likely to know exactly which certificate extensions are required than are human users. PKI management entities are also end entities in the sense that they are sometimes named in the subject or subjectAltName field of a certificate or cross-certificate. Where appropriate, the term "end- entity" will be used to refer to end entities who are not PKI management entities. All end entities require secure local access to some information -- at a minimum, their own name and private key, the name of a CA that is directly trusted by this entity, and that CA's public key (or a fingerprint of the public key where a self-certified version is available elsewhere). Implementations MAY use secure local storage for more than this minimum (e.g., the end entity's own certificate or Adams, et al. Standards Track [Page 6] RFC 4210 CMP September 2005 application-specific information). The form of storage will also vary -- from files to tamper-resistant cryptographic tokens. The information stored in such local, trusted storage is referred to here as the end entity's Personal Security Environment (PSE). Though PSE formats are beyond the scope of this document (they are very dependent on equipment, et cetera), a generic interchange format for PSEs is defined here: a certification response message MAY be used. 3.1.1.2. Certification Authority The certification authority (CA) may or may not actually be a real "third party" from the end entity's point of view. Quite often, the CA will actually belong to the same organization as the end entities it supports. Again, we use the term "CA" to refer to the entity named in the issuer field of a certificate. When it is necessary to distinguish the software or hardware tools used by the CA, we use the term "CA equipment". The CA equipment will often include both an "off-line" component and an "on-line" component, with the CA private key only available to the "off-line" component. This is, however, a matter for implementers (though it is also relevant as a policy issue). We use the term "root CA" to indicate a CA that is directly trusted by an end entity; that is, securely acquiring the value of a root CA public key requires some out-of-band step(s). This term is not meant to imply that a root CA is necessarily at the top of any hierarchy, simply that the CA in question is trusted directly. A "subordinate CA" is one that is not a root CA for the end entity in question. Often, a subordinate CA will not be a root CA for any entity, but this is not mandatory. 3.1.1.3. Registration Authority In addition to end-entities and CAs, many environments call for the existence of a Registration Authority (RA) separate from the Certification Authority. The functions that the registration authority may carry out will vary from case to case but MAY include personal authentication, token distribution, revocation reporting, name assignment, key generation, archival of key pairs, et cetera. Adams, et al. Standards Track [Page 7] RFC 4210 CMP September 2005 This document views the RA as an OPTIONAL component: when it is not present, the CA is assumed to be able to carry out the RA's functions so that the PKI management protocols are the same from the end- entity's point of view. Again, we distinguish, where necessary, between the RA and the tools used (the "RA equipment"). Note that an RA is itself an end entity. We further assume that all RAs are in fact certified end entities and that RAs have private keys that are usable for signing. How a particular CA equipment identifies some end entities as RAs is an implementation issue (i.e., this document specifies no special RA certification operation). We do not mandate that the RA is certified by the CA with which it is interacting at the moment (so one RA may work with more than one CA whilst only being certified once). In some circumstances, end entities will communicate directly with a CA even where an RA is present. For example, for initial registration and/or certification, the subject may use its RA, but communicate directly with the CA in order to refresh its certificate. 3.1.2. PKI Management Requirements The protocols given here meet the following requirements on PKI management 1. PKI management must conform to the ISO/IEC 9594-8/ITU-T X.509 standards. 2. It must be possible to regularly update any key pair without affecting any other key pair. 3. The use of confidentiality in PKI management protocols must be kept to a minimum in order to ease acceptance in environments where strong confidentiality might cause regulatory problems. 4. PKI management protocols must allow the use of different industry-standard cryptographic algorithms (specifically including RSA, DSA, MD5, and SHA-1). This means that any given CA, RA, or end entity may, in principle, use whichever algorithms suit it for its own key pair(s). 5. PKI management protocols must not preclude the generation of key pairs by the end-entity concerned, by an RA, or by a CA. Key generation may also occur elsewhere, but for the purposes of PKI management we can regard key generation as occurring wherever the key is first present at an end entity, RA, or CA. Adams, et al. Standards Track [Page 8] RFC 4210 CMP September 2005 6. PKI management protocols must support the publication of certificates by the end-entity concerned, by an RA, or by a CA. Different implementations and different environments may choose any of the above approaches. 7. PKI management protocols must support the production of Certificate Revocation Lists (CRLs) by allowing certified end entities to make requests for the revocation of certificates. This must be done in such a way that the denial-of-service attacks, which are possible, are not made simpler. 8. PKI management protocols must be usable over a variety of "transport" mechanisms, specifically including mail, http, TCP/IP and ftp. 9. Final authority for certification creation rests with the CA. No RA or end-entity equipment can assume that any certificate issued by a CA will contain what was requested; a CA may alter certificate field values or may add, delete, or alter extensions according to its operating policy. In other words, all PKI entities (end-entities, RAs, and CAs) must be capable of handling responses to requests for certificates in which the actual certificate issued is different from that requested (for example, a CA may shorten the validity period requested). Note that policy may dictate that the CA must not publish or otherwise distribute the certificate until the requesting entity has reviewed and accepted the newly-created certificate (typically through use of the certConf message). 10. A graceful, scheduled change-over from one non-compromised CA key pair to the next (CA key update) must be supported (note that if the CA key is compromised, re-initialization must be performed for all entities in the domain of that CA). An end entity whose PSE contains the new CA public key (following a CA key update) must also be able to verify certificates verifiable using the old public key. End entities who directly trust the old CA key pair must also be able to verify certificates signed using the new CA private key (required for situations where the old CA public key is "hardwired" into the end entity's cryptographic equipment). 11. The functions of an RA may, in some implementations or environments, be carried out by the CA itself. The protocols must be designed so that end entities will use the same protocol regardless of whether the communication is with an RA or CA. Naturally, the end entity must use the correct RA of CA public key to protect the communication. Adams, et al. Standards Track [Page 9] RFC 4210 CMP September 2005 12. Where an end entity requests a certificate containing a given public key value, the end entity must be ready to demonstrate possession of the corresponding private key value. This may be accomplished in various ways, depending on the type of certification request. See Section 4.3 for details of the in- band methods defined for the PKIX-CMP (i.e., Certificate Management Protocol) messages. 3.1.3. PKI Management Operations The following diagram shows the relationship between the entities defined above in terms of the PKI management operations. The letters in the diagram indicate "protocols" in the sense that a defined set of PKI management messages can be sent along each of the lettered lines. Adams, et al. Standards Track [Page 10] RFC 4210 CMP September 2005 +---+ cert. publish +------------+ j | | <--------------------- | End Entity | <------- | C | g +------------+ "out-of-band" | e | | ^ loading | r | | | initial | t | a | | b registration/ | | | | certification | / | | | key pair recovery | | | | key pair update | C | | | certificate update | R | PKI "USERS" V | revocation request | L | -------------------+-+-----+-+------+-+------------------- | | PKI MANAGEMENT | ^ | ^ | | ENTITIES a | | b a | | b | R | V | | | | e | g +------+ d | | | p | <------------ | RA | <-----+ | | | o | cert. | | ----+ | | | | s | publish +------+ c | | | | | i | | | | | | t | V | V | | o | g +------------+ i | r | <------------------------| CA |-------> | y | h +------------+ "out-of-band" | | cert. publish | ^ publication | | CRL publish | | +---+ | | cross-certification e | | f cross-certificate | | update | | V | +------+ | CA-2 | +------+ Figure 1 - PKI Entities At a high level, the set of operations for which management messages are defined can be grouped as follows. 1. CA establishment: When establishing a new CA, certain steps are required (e.g., production of initial CRLs, export of CA public key). 2. End entity initialization: this includes importing a root CA public key and requesting information about the options supported by a PKI management entity. Adams, et al. Standards Track [Page 11] RFC 4210 CMP September 2005 3. Certification: various operations result in the creation of new certificates: 1. initial registration/certification: This is the process whereby an end entity first makes itself known to a CA or RA, prior to the CA issuing a certificate or certificates for that end entity. The end result of this process (when it is successful) is that a CA issues a certificate for an end entity's public key, and returns that certificate to the end entity and/or posts that certificate in a public repository. This process may, and typically will, involve multiple "steps", possibly including an initialization of the end entity's equipment. For example, the end entity's equipment must be securely initialized with the public key of a CA, to be used in validating certificate paths. Furthermore, an end entity typically needs to be initialized with its own key pair(s). 2. key pair update: Every key pair needs to be updated regularly (i.e., replaced with a new key pair), and a new certificate needs to be issued. 3. certificate update: As certificates expire, they may be "refreshed" if nothing relevant in the environment has changed. 4. CA key pair update: As with end entities, CA key pairs need to be updated regularly; however, different mechanisms are required. 5. cross-certification request: One CA requests issuance of a cross-certificate from another CA. For the purposes of this standard, the following terms are defined. A "cross- certificate" is a certificate in which the subject CA and the issuer CA are distinct and SubjectPublicKeyInfo contains a verification key (i.e., the certificate has been issued for the subject CA's signing key pair). When it is necessary to distinguish more finely, the following terms may be used: a cross-certificate is called an "inter-domain cross- certificate" if the subject and issuer CAs belong to different administrative domains; it is called an "intra- domain cross-certificate" otherwise. 1. Note 1. The above definition of "cross-certificate" aligns with the defined term "CA-certificate" in X.509. Note that this term is not to be confused with the X.500 "cACertificate" attribute type, which is unrelated. Adams, et al. Standards Track [Page 12] RFC 4210 CMP September 2005 2. Note 2. In many environments, the term "cross- certificate", unless further qualified, will be understood to be synonymous with "inter-domain cross- certificate" as defined above. 3. Note 3. Issuance of cross-certificates may be, but is not necessarily, mutual; that is, two CAs may issue cross-certificates for each other. 6. cross-certificate update: Similar to a normal certificate update, but involving a cross-certificate. 4. Certificate/CRL discovery operations: some PKI management operations result in the publication of certificates or CRLs: 1. certificate publication: Having gone to the trouble of producing a certificate, some means for publishing it is needed. The "means" defined in PKIX MAY involve the messages specified in Sections 5.3.13 to 5.3.16, or MAY involve other methods (LDAP, for example) as described in [RFC2559], [RFC2585] (the "Operational Protocols" documents of the PKIX series of specifications). 2. CRL publication: As for certificate publication. 5. Recovery operations: some PKI management operations are used when an end entity has "lost" its PSE: 1. key pair recovery: As an option, user client key materials (e.g., a user's private key used for decryption purposes) MAY be backed up by a CA, an RA, or a key backup system associated with a CA or RA. If an entity needs to recover these backed up key materials (e.g., as a result of a forgotten password or a lost key chain file), a protocol exchange may be needed to support such recovery. 6. Revocation operations: some PKI operations result in the creation of new CRL entries and/or new CRLs: 1. revocation request: An authorized person advises a CA of an abnormal situation requiring certificate revocation. 7. PSE operations: whilst the definition of PSE operations (e.g., moving a PSE, changing a PIN, etc.) are beyond the scope of this specification, we do define a PKIMessage (CertRepMessage) that can form the basis of such operations. Adams, et al. Standards Track [Page 13] RFC 4210 CMP September 2005 Note that on-line protocols are not the only way of implementing the above operations. For all operations, there are off-line methods of achieving the same result, and this specification does not mandate use of on-line protocols. For example, when hardware tokens are used, many of the operations MAY be achieved as part of the physical token delivery. Later sections define a set of standard messages supporting the above operations. Transport protocols for conveying these exchanges in different environments (file-based, on-line, E-mail, and WWW) are beyond the scope of this document and are specified separately. 4. Assumptions and Restrictions 4.1. End Entity Initialization The first step for an end entity in dealing with PKI management entities is to request information about the PKI functions supported and to securely acquire a copy of the relevant root CA public key(s). 4.2. Initial Registration/Certification There are many schemes that can be used to achieve initial registration and certification of end entities. No one method is suitable for all situations due to the range of policies that a CA may implement and the variation in the types of end entity which can occur. However, we can classify the initial registration/certification schemes that are supported by this specification. Note that the word "initial", above, is crucial: we are dealing with the situation where the end entity in question has had no previous contact with the PKI. Where the end entity already possesses certified keys, then some simplifications/alternatives are possible. Having classified the schemes that are supported by this specification we can then specify some as mandatory and some as optional. The goal is that the mandatory schemes cover a sufficient number of the cases that will arise in real use, whilst the optional schemes are available for special cases that arise less frequently. In this way, we achieve a balance between flexibility and ease of implementation. We will now describe the classification of initial registration/certification schemes. Adams, et al. Standards Track [Page 14] RFC 4210 CMP September 2005 4.2.1. Criteria Used 4.2.1.1. Initiation of Registration/Certification In terms of the PKI messages that are produced, we can regard the initiation of the initial registration/certification exchanges as occurring wherever the first PKI message relating to the end entity is produced. Note that the real-world initiation of the registration/certification procedure may occur elsewhere (e.g., a personnel department may telephone an RA operator). The possible locations are at the end entity, an RA, or a CA. 4.2.1.2. End Entity Message Origin Authentication The on-line messages produced by the end entity that requires a certificate may be authenticated or not. The requirement here is to authenticate the origin of any messages from the end entity to the PKI (CA/RA). In this specification, such authentication is achieved by the PKI (CA/RA) issuing the end entity with a secret value (initial authentication key) and reference value (used to identify the secret value) via some out-of-band means. The initial authentication key can then be used to protect relevant PKI messages. Thus, we can classify the initial registration/certification scheme according to whether or not the on-line end entity -> PKI messages are authenticated or not. Note 1: We do not discuss the authentication of the PKI -> end entity messages here, as this is always REQUIRED. In any case, it can be achieved simply once the root-CA public key has been installed at the end entity's equipment or it can be based on the initial authentication key. Note 2: An initial registration/certification procedure can be secure where the messages from the end entity are authenticated via some out-of-band means (e.g., a subsequent visit). 4.2.1.3. Location of Key Generation In this specification, "key generation" is regarded as occurring wherever either the public or private component of a key pair first occurs in a PKIMessage. Note that this does not preclude a centralized key generation service; the actual key pair MAY have been Adams, et al. Standards Track [Page 15] RFC 4210 CMP September 2005 generated elsewhere and transported to the end entity, RA, or CA using a (proprietary or standardized) key generation request/response protocol (outside the scope of this specification). Thus, there are three possibilities for the location of "key generation": the end entity, an RA, or a CA. 4.2.1.4. Confirmation of Successful Certification Following the creation of an initial certificate for an end entity, additional assurance can be gained by having the end entity explicitly confirm successful receipt of the message containing (or indicating the creation of) the certificate. Naturally, this confirmation message must be protected (based on the initial authentication key or other means). This gives two further possibilities: confirmed or not. 4.2.2. Mandatory Schemes The criteria above allow for a large number of initial registration/certification schemes. This specification mandates that conforming CA equipment, RA equipment, and EE equipment MUST support the second scheme listed below (Section 4.2.2.2). Any entity MAY additionally support other schemes, if desired. 4.2.2.1. Centralized Scheme In terms of the classification above, this scheme is, in some ways, the simplest possible, where: o initiation occurs at the certifying CA; o no on-line message authentication is required; o "key generation" occurs at the certifying CA (see Section 4.2.1.3); o no confirmation message is required. In terms of message flow, this scheme means that the only message required is sent from the CA to the end entity. The message must contain the entire PSE for the end entity. Some out-of-band means must be provided to allow the end entity to authenticate the message received and to decrypt any encrypted values. Adams, et al. Standards Track [Page 16] RFC 4210 CMP September 2005 4.2.2.2. Basic Authenticated Scheme In terms of the classification above, this scheme is where: o initiation occurs at the end entity; o message authentication is REQUIRED; o "key generation" occurs at the end entity (see Section 4.2.1.3); o a confirmation message is REQUIRED. In terms of message flow, the basic authenticated scheme is as follows: End entity RA/CA ========== ============= out-of-band distribution of Initial Authentication Key (IAK) and reference value (RA/CA -> EE) Key generation Creation of certification request Protect request with IAK -->>-- certification request -->>-- verify request process request create response --<<-- certification response --<<-- handle response create confirmation -->>-- cert conf message -->>-- verify confirmation create response --<<-- conf ack (optional) --<<-- handle response (Where verification of the cert confirmation message fails, the RA/CA MUST revoke the newly issued certificate if it has been published or otherwise made available.) 4.3. Proof-of-Possession (POP) of Private Key In order to prevent certain attacks and to allow a CA/RA to properly check the validity of the binding between an end entity and a key pair, the PKI management operations specified here make it possible for an end entity to prove that it has possession of (i.e., is able to use) the private key corresponding to the public key for which a certificate is requested. A given CA/RA is free to choose how to enforce POP (e.g., out-of-band procedural means versus PKIX-CMP Adams, et al. Standards Track [Page 17] RFC 4210 CMP September 2005 in-band messages) in its certification exchanges (i.e., this may be a policy issue). However, it is REQUIRED that CAs/RAs MUST enforce POP by some means because there are currently many non-PKIX operational protocols in use (various electronic mail protocols are one example) that do not explicitly check the binding between the end entity and the private key. Until operational protocols that do verify the binding (for signature, encryption, and key agreement key pairs) exist, and are ubiquitous, this binding can only be assumed to have been verified by the CA/RA. Therefore, if the binding is not verified by the CA/RA, certificates in the Internet Public-Key Infrastructure end up being somewhat less meaningful. POP is accomplished in different ways depending upon the type of key for which a certificate is requested. If a key can be used for multiple purposes (e.g., an RSA key) then any appropriate method MAY be used (e.g., a key that may be used for signing, as well as other purposes, SHOULD NOT be sent to the CA/RA in order to prove possession). This specification explicitly allows for cases where an end entity supplies the relevant proof to an RA and the RA subsequently attests to the CA that the required proof has been received (and validated!). For example, an end entity wishing to have a signing key certified could send the appropriate signature to the RA, which then simply notifies the relevant CA that the end entity has supplied the required proof. Of course, such a situation may be disallowed by some policies (e.g., CAs may be the only entities permitted to verify POP during certification). 4.3.1. Signature Keys For signature keys, the end entity can sign a value to prove possession of the private key. 4.3.2. Encryption Keys For encryption keys, the end entity can provide the private key to the CA/RA, or can be required to decrypt a value in order to prove possession of the private key (see Section 5.2.8). Decrypting a value can be achieved either directly or indirectly. The direct method is for the RA/CA to issue a random challenge to which an immediate response by the EE is required. Adams, et al. Standards Track [Page 18] RFC 4210 CMP September 2005 The indirect method is to issue a certificate that is encrypted for the end entity (and have the end entity demonstrate its ability to decrypt this certificate in the confirmation message). This allows a CA to issue a certificate in a form that can only be used by the intended end entity. This specification encourages use of the indirect method because it requires no extra messages to be sent (i.e., the proof can be demonstrated using the {request, response, confirmation} triple of messages). 4.3.3. Key Agreement Keys For key agreement keys, the end entity and the PKI management entity (i.e., CA or RA) must establish a shared secret key in order to prove that the end entity has possession of the private key. Note that this need not impose any restrictions on the keys that can be certified by a given CA. In particular, for Diffie-Hellman keys the end entity may freely choose its algorithm parameters provided that the CA can generate a short-term (or one-time) key pair with the appropriate parameters when necessary. 4.4. Root CA Key Update This discussion only applies to CAs that are directly trusted by some end entities. Self-signed CAs SHALL be considered as directly trusted CAs. Recognizing whether a non-self-signed CA is supposed to be directly trusted for some end entities is a matter of CA policy and is thus beyond the scope of this document. The basis of the procedure described here is that the CA protects its new public key using its previous private key and vice versa. Thus, when a CA updates its key pair it must generate two extra cACertificate attribute values if certificates are made available using an X.500 directory (for a total of four: OldWithOld, OldWithNew, NewWithOld, and NewWithNew). When a CA changes its key pair, those entities who have acquired the old CA public key via "out-of-band" means are most affected. It is these end entities who will need access to the new CA public key protected with the old CA private key. However, they will only require this for a limited period (until they have acquired the new CA public key via the "out-of-band" mechanism). This will typically be easily achieved when these end entities' certificates expire. Adams, et al. Standards Track [Page 19] RFC 4210 CMP September 2005 The data structure used to protect the new and old CA public keys is a standard certificate (which may also contain extensions). There are no new data structures required. Note 1. This scheme does not make use of any of the X.509 v3 extensions as it must be able to work even for version 1 certificates. The presence of the KeyIdentifier extension would make for efficiency improvements. Note 2. While the scheme could be generalized to cover cases where the CA updates its key pair more than once during the validity period of one of its end entities' certificates, this generalization seems of dubious value. Not having this generalization simply means that the validity periods of certificates issued with the old CA key pair cannot exceed the end of the OldWithNew validity period. Note 3. This scheme ensures that end entities will acquire the new CA public key, at the latest by the expiry of the last certificate they owned that was signed with the old CA private key (via the "out-of-band" means). Certificate and/or key update operations occurring at other times do not necessarily require this (depending on the end entity's equipment). 4.4.1. CA Operator Actions To change the key of the CA, the CA operator does the following: 1. Generate a new key pair; 2. Create a certificate containing the old CA public key signed with the new private key (the "old with new" certificate); 3. Create a certificate containing the new CA public key signed with the old private key (the "new with old" certificate); 4. Create a certificate containing the new CA public key signed with the new private key (the "new with new" certificate); 5. Publish these new certificates via the repository and/or other means (perhaps using a CAKeyUpdAnn message); 6. Export the new CA public key so that end entities may acquire it using the "out-of-band" mechanism (if required). The old CA private key is then no longer required. However, the old CA public key will remain in use for some time. The old CA public key is no longer required (other than for non-repudiation) when all end entities of this CA have securely acquired the new CA public key. Adams, et al. Standards Track [Page 20] RFC 4210 CMP September 2005 The "old with new" certificate must have a validity period starting at the generation time of the old key pair and ending at the expiry date of the old public key. The "new with old" certificate must have a validity period starting at the generation time of the new key pair and ending at the time by which all end entities of this CA will securely possess the new CA public key (at the latest, the expiry date of the old public key). The "new with new" certificate must have a validity period starting at the generation time of the new key pair and ending at or before the time by which the CA will next update its key pair. 4.4.2. Verifying Certificates Normally when verifying a signature, the verifier verifies (among other things) the certificate containing the public key of the signer. However, once a CA is allowed to update its key there are a range of new possibilities. These are shown in the table below. Repository contains NEW Repository contains only OLD and OLD public keys public key (due to, e.g., delay in publication) PSE PSE Contains PSE Contains PSE Contains Contains OLD public NEW public OLD public NEW public key key key key Signer's Case 1: Case 3: Case 5: Case 7: certifi- This is In this case Although the In this case cate is the the verifier CA operator the CA protected standard must access has not operator has using NEW case where the updated the not updated public the repository in repository the the repository key verifier order to get verifier can and so the can the value of verify the verification directly the NEW certificate will FAIL verify the public key directly - certificate this is thus without the same as using the case 1. repository Adams, et al. Standards Track [Page 21] RFC 4210 CMP September 2005 Signer's Case 2: Case 4: Case 6: Case 8: certifi- In this In this case The verifier Although the cate is case the the verifier thinks this CA operator protected verifier can directly is the has not using OLD must verify the situation of updated the public access the certificate case 2 and repository the key repository without will access verifier can in order using the the verify the to get the repository repository; certificate value of however, the directly - the OLD verification this is thus public key will FAIL the same as case 4. 4.4.2.1. Verification in Cases 1, 4, 5, and 8 In these cases, the verifier has a local copy of the CA public key that can be used to verify the certificate directly. This is the same as the situation where no key change has occurred. Note that case 8 may arise between the time when the CA operator has generated the new key pair and the time when the CA operator stores the updated attributes in the repository. Case 5 can only arise if the CA operator has issued both the signer's and verifier's certificates during this "gap" (the CA operator SHOULD avoid this as it leads to the failure cases described below) 4.4.2.2. Verification in Case 2 In case 2, the verifier must get access to the old public key of the CA. The verifier does the following: 1. Look up the caCertificate attribute in the repository and pick the OldWithNew certificate (determined based on validity periods; note that the subject and issuer fields must match); 2. Verify that this is correct using the new CA key (which the verifier has locally); 3. If correct, check the signer's certificate using the old CA key. Case 2 will arise when the CA operator has issued the signer's certificate, then changed the key, and then issued the verifier's certificate; so it is quite a typical case. Adams, et al. Standards Track [Page 22] RFC 4210 CMP September 2005 4.4.2.3. Verification in Case 3 In case 3, the verifier must get access to the new public key of the CA. The verifier does the following: 1. Look up the CACertificate attribute in the repository and pick the NewWithOld certificate (determined based on validity periods; note that the subject and issuer fields must match); 2. Verify that this is correct using the old CA key (which the verifier has stored locally); 3. If correct, check the signer's certificate using the new CA key. Case 3 will arise when the CA operator has issued the verifier's certificate, then changed the key, and then issued the signer's certificate; so it is also quite a typical case. 4.4.2.4. Failure of Verification in Case 6 In this case, the CA has issued the verifier's PSE, which contains the new key, without updating the repository attributes. This means that the verifier has no means to get a trustworthy version of the CA's old key and so verification fails. Note that the failure is the CA operator's fault. 4.4.2.5. Failure of Verification in Case 7 In this case, the CA has issued the signer's certificate protected with the new key without updating the repository attributes. This means that the verifier has no means to get a trustworthy version of the CA's new key and so verification fails. Note that the failure is again the CA operator's fault. 4.4.3. Revocation - Change of CA Key As we saw above, the verification of a certificate becomes more complex once the CA is allowed to change its key. This is also true for revocation checks as the CA may have signed the CRL using a newer private key than the one within the user's PSE. The analysis of the alternatives is the same as for certificate verification. Adams, et al. Standards Track [Page 23] RFC 4210 CMP September 2005 5. Data Structures This section contains descriptions of the data structures required for PKI management messages. Section 6 describes constraints on their values and the sequence of events for each of the various PKI management operations. 5.1. Overall PKI Message All of the messages used in this specification for the purposes of PKI management use the following structure: PKIMessage ::= SEQUENCE { header PKIHeader, body PKIBody, protection [0] PKIProtection OPTIONAL, extraCerts [1] SEQUENCE SIZE (1..MAX) OF CMPCertificate OPTIONAL } PKIMessages ::= SEQUENCE SIZE (1..MAX) OF PKIMessage The PKIHeader contains information that is common to many PKI messages. The PKIBody contains message-specific information. The PKIProtection, when used, contains bits that protect the PKI message. The extraCerts field can contain certificates that may be useful to the recipient. For example, this can be used by a CA or RA to present an end entity with certificates that it needs to verify its own new certificate (if, for example, the CA that issued the end entity's certificate is not a root CA for the end entity). Note that this field does not necessarily contain a certification path; the recipient may have to sort, select from, or otherwise process the extra certificates in order to use them. 5.1.1. PKI Message Header All PKI messages require some header information for addressing and transaction identification. Some of this information will also be present in a transport-specific envelope. However, if the PKI message is protected, then this information is also protected (i.e., we make no assumption about secure transport). Adams, et al. Standards Track [Page 24] RFC 4210 CMP September 2005 The following data structure is used to contain this information: PKIHeader ::= SEQUENCE { pvno INTEGER { cmp1999(1), cmp2000(2) }, sender GeneralName, recipient GeneralName, messageTime [0] GeneralizedTime OPTIONAL, protectionAlg [1] AlgorithmIdentifier OPTIONAL, senderKID [2] KeyIdentifier OPTIONAL, recipKID [3] KeyIdentifier OPTIONAL, transactionID [4] OCTET STRING OPTIONAL, senderNonce [5] OCTET STRING OPTIONAL, recipNonce [6] OCTET STRING OPTIONAL, freeText [7] PKIFreeText OPTIONAL, generalInfo [8] SEQUENCE SIZE (1..MAX) OF InfoTypeAndValue OPTIONAL } PKIFreeText ::= SEQUENCE SIZE (1..MAX) OF UTF8String The pvno field is fixed (at 2) for this version of this specification. The sender field contains the name of the sender of the PKIMessage. This name (in conjunction with senderKID, if supplied) should be sufficient to indicate the key to use to verify the protection on the message. If nothing about the sender is known to the sending entity (e.g., in the init. req. message, where the end entity may not know its own Distinguished Name (DN), e-mail name, IP address, etc.), then the "sender" field MUST contain a "NULL" value; that is, the SEQUENCE OF relative distinguished names is of zero length. In such a case, the senderKID field MUST hold an identifier (i.e., a reference number) that indicates to the receiver the appropriate shared secret information to use to verify the message. The recipient field contains the name of the recipient of the PKIMessage. This name (in conjunction with recipKID, if supplied) should be usable to verify the protection on the message. The protectionAlg field specifies the algorithm used to protect the message. If no protection bits are supplied (note that PKIProtection is OPTIONAL) then this field MUST be omitted; if protection bits are supplied, then this field MUST be supplied. senderKID and recipKID are usable to indicate which keys have been used to protect the message (recipKID will normally only be required where protection of the message uses Diffie-Hellman (DH) keys). Adams, et al. Standards Track [Page 25] RFC 4210 CMP September 2005 These fields MUST be used if required to uniquely identify a key (e.g., if more than one key is associated with a given sender name) and SHOULD be omitted otherwise. The transactionID field within the message header is to be used to allow the recipient of a message to correlate this with an ongoing transaction. This is needed for all transactions that consist of more than just a single request/response pair. For transactions that consist of a single request/response pair, the rules are as follows. A client MAY populate the transactionID field of the request. If a server receives such a request that has the transactionID field set, then it MUST set the transactionID field of the response to the same value. If a server receives such request with a missing transactionID field, then it MAY set transactionID field of the response. For transactions that consist of more than just a single request/response pair, the rules are as follows. Clients SHOULD generate a transactionID for the first request. If a server receives such a request that has the transactionID field set, then it MUST set the transactionID field of the response to the same value. If a server receives such request with a missing transactionID field, then it MUST populate the transactionID field of the response with a server-generated ID. Subsequent requests and responses MUST all set the transactionID field to the thus established value. In all cases where a transactionID is being used, a given client MUST NOT have more than one transaction with the same transactionID in progress at any time (to a given server). Servers are free to require uniqueness of the transactionID or not, as long as they are able to correctly associate messages with the corresponding transaction. Typically, this means that a server will require the {client, transactionID} tuple to be unique, or even the transactionID alone to be unique, if it cannot distinguish clients based on transport-level information. A server receiving the first message of a transaction (which requires more than a single request/response pair) that contains a transactionID that does not allow it to meet the above constraints (typically because the transactionID is already in use) MUST send back an ErrorMsgContent with a PKIFailureInfo of transactionIdInUse. It is RECOMMENDED that the clients fill the transactionID field with 128 bits of (pseudo-) random data for the start of a transaction to reduce the probability of having the transactionID in use at the server. The senderNonce and recipNonce fields protect the PKIMessage against replay attacks. The senderNonce will typically be 128 bits of (pseudo-) random data generated by the sender, whereas the recipNonce is copied from the senderNonce of the previous message in the transaction. Adams, et al. Standards Track [Page 26] RFC 4210 CMP September 2005 The messageTime field contains the time at which the sender created the message. This may be useful to allow end entities to correct/check their local time for consistency with the time on a central system. The freeText field may be used to send a human-readable message to the recipient (in any number of languages). The first language used in this sequence indicates the desired language for replies. The generalInfo field may be used to send machine-processable additional data to the recipient. The following generalInfo extensions are defined and MAY be supported. 5.1.1.1. ImplicitConfirm This is used by the EE to inform the CA that it does not wish to send a certificate confirmation for issued certificates. implicitConfirm OBJECT IDENTIFIER ::= {id-it 13} ImplicitConfirmValue ::= NULL If the CA grants the request to the EE, it MUST put the same extension in the PKIHeader of the response. If the EE does not find the extension in the response, it MUST send the certificate confirmation. 5.1.1.2. ConfirmWaitTime This is used by the CA to inform the EE how long it intends to wait for the certificate confirmation before revoking the certificate and deleting the transaction. confirmWaitTime OBJECT IDENTIFIER ::= {id-it 14} ConfirmWaitTimeValue ::= GeneralizedTime 5.1.2. PKI Message Body PKIBody ::= CHOICE { ir [0] CertReqMessages, --Initialization Req ip [1] CertRepMessage, --Initialization Resp cr [2] CertReqMessages, --Certification Req cp [3] CertRepMessage, --Certification Resp p10cr [4] CertificationRequest, --PKCS #10 Cert. Req. popdecc [5] POPODecKeyChallContent --pop Challenge popdecr [6] POPODecKeyRespContent, --pop Response kur [7] CertReqMessages, --Key Update Request kup [8] CertRepMessage, --Key Update Response krr [9] CertReqMessages, --Key Recovery Req Adams, et al. Standards Track [Page 27] RFC 4210 CMP September 2005 krp [10] KeyRecRepContent, --Key Recovery Resp rr [11] RevReqContent, --Revocation Request rp [12] RevRepContent, --Revocation Response ccr [13] CertReqMessages, --Cross-Cert. Request ccp [14] CertRepMessage, --Cross-Cert. Resp ckuann [15] CAKeyUpdAnnContent, --CA Key Update Ann. cann [16] CertAnnContent, --Certificate Ann. rann [17] RevAnnContent, --Revocation Ann. crlann [18] CRLAnnContent, --CRL Announcement pkiconf [19] PKIConfirmContent, --Confirmation nested [20] NestedMessageContent, --Nested Message genm [21] GenMsgContent, --General Message genp [22] GenRepContent, --General Response error [23] ErrorMsgContent, --Error Message certConf [24] CertConfirmContent, --Certificate confirm pollReq [25] PollReqContent, --Polling request pollRep [26] PollRepContent --Polling response } The specific types are described in Section 5.3 below. 5.1.3. PKI Message Protection Some PKI messages will be protected for integrity. (Note that if an asymmetric algorithm is used to protect a message and the relevant public component has been certified already, then the origin of the message can also be authenticated. On the other hand, if the public component is uncertified, then the message origin cannot be automatically authenticated, but may be authenticated via out-of-band means.) When protection is applied, the following structure is used: PKIProtection ::= BIT STRING The input to the calculation of PKIProtection is the DER encoding of the following data structure: ProtectedPart ::= SEQUENCE { header PKIHeader, body PKIBody } There MAY be cases in which the PKIProtection BIT STRING is deliberately not used to protect a message (i.e., this OPTIONAL field is omitted) because other protection, external to PKIX, will be applied instead. Such a choice is explicitly allowed in this specification. Examples of such external protection include PKCS #7 Adams, et al. Standards Track [Page 28] RFC 4210 CMP September 2005 [PKCS7] and Security Multiparts [RFC1847] encapsulation of the PKIMessage (or simply the PKIBody (omitting the CHOICE tag), if the relevant PKIHeader information is securely carried in the external mechanism). It is noted, however, that many such external mechanisms require that the end entity already possesses a public-key certificate, and/or a unique Distinguished Name, and/or other such infrastructure-related information. Thus, they may not be appropriate for initial registration, key-recovery, or any other process with "boot-strapping" characteristics. For those cases it may be necessary that the PKIProtection parameter be used. In the future, if/when external mechanisms are modified to accommodate boot-strapping scenarios, the use of PKIProtection may become rare or non-existent. Depending on the circumstances, the PKIProtection bits may contain a Message Authentication Code (MAC) or signature. Only the following cases can occur: 5.1.3.1. Shared Secret Information In this case, the sender and recipient share secret information (established via out-of-band means or from a previous PKI management operation). PKIProtection will contain a MAC value and the protectionAlg will be the following (see also Appendix D.2): id-PasswordBasedMac OBJECT IDENTIFIER ::= {1 2 840 113533 7 66 13} PBMParameter ::= SEQUENCE { salt OCTET STRING, owf AlgorithmIdentifier, iterationCount INTEGER, mac AlgorithmIdentifier } In the above protectionAlg, the salt value is appended to the shared secret input. The OWF is then applied iterationCount times, where the salted secret is the input to the first iteration and, for each successive iteration, the input is set to be the output of the previous iteration. The output of the final iteration (called "BASEKEY" for ease of reference, with a size of "H") is what is used to form the symmetric key. If the MAC algorithm requires a K-bit key and K <= H, then the most significant K bits of BASEKEY are used. If K > H, then all of BASEKEY is used for the most significant H bits of the key, OWF("1" || BASEKEY) is used for the next most significant H bits of the key, OWF("2" || BASEKEY) is used for the next most significant H bits of the key, and so on, until all K bits have been derived. [Here "N" is the ASCII byte encoding the number N and "||" represents concatenation.] Adams, et al. Standards Track [Page 29] RFC 4210 CMP September 2005 Note: it is RECOMMENDED that the fields of PBMParameter remain constant throughout the messages of a single transaction (e.g., ir/ip/certConf/pkiConf) in order to reduce the overhead associated with PasswordBasedMac computation). 5.1.3.2. DH Key Pairs Where the sender and receiver possess Diffie-Hellman certificates with compatible DH parameters, in order to protect the message the end entity must generate a symmetric key based on its private DH key value and the DH public key of the recipient of the PKI message. PKIProtection will contain a MAC value keyed with this derived symmetric key and the protectionAlg will be the following: id-DHBasedMac OBJECT IDENTIFIER ::= {1 2 840 113533 7 66 30} DHBMParameter ::= SEQUENCE { owf AlgorithmIdentifier, -- AlgId for a One-Way Function (SHA-1 recommended) mac AlgorithmIdentifier -- the MAC AlgId (e.g., DES-MAC, Triple-DES-MAC [PKCS11], } -- or HMAC [RFC2104, RFC2202]) In the above protectionAlg, OWF is applied to the result of the Diffie-Hellman computation. The OWF output (called "BASEKEY" for ease of reference, with a size of "H") is what is used to form the symmetric key. If the MAC algorithm requires a K-bit key and K <= H, then the most significant K bits of BASEKEY are used. If K > H, then all of BASEKEY is used for the most significant H bits of the key, OWF("1" || BASEKEY) is used for the next most significant H bits of the key, OWF("2" || BASEKEY) is used for the next most significant H bits of the key, and so on, until all K bits have been derived. [Here "N" is the ASCII byte encoding the number N and "||" represents concatenation.] 5.1.3.3. Signature In this case, the sender possesses a signature key pair and simply signs the PKI message. PKIProtection will contain the signature value and the protectionAlg will be an AlgorithmIdentifier for a digital signature (e.g., md5WithRSAEncryption or dsaWithSha-1). 5.1.3.4. Multiple Protection In cases where an end entity sends a protected PKI message to an RA, the RA MAY forward that message to a CA, attaching its own protection (which MAY be a MAC or a signature, depending on the information and certificates shared between the RA and the CA). This is accomplished Adams, et al. Standards Track [Page 30] RFC 4210 CMP September 2005 by nesting the entire message sent by the end entity within a new PKI message. The structure used is as follows. NestedMessageContent ::= PKIMessages (The use of PKIMessages, a SEQUENCE OF PKIMessage, lets the RA batch the requests of several EEs in a single new message. For simplicity, all messages in the batch MUST be of the same type (e.g., ir).) If the RA wishes to modify the message(s) in some way (e.g., add particular field values or new extensions), then it MAY create its own desired PKIBody. The original PKIMessage from the EE MAY be included in the generalInfo field of PKIHeader (to accommodate, for example, cases in which the CA wishes to check POP or other information on the original EE message). The infoType to be used in this situation is {id-it 15} (see Section 5.3.19 for the value of id-it) and the infoValue is PKIMessages (contents MUST be in the same order as the requests in PKIBody). 5.2. Common Data Structures Before specifying the specific types that may be placed in a PKIBody, we define some data structures that are used in more than one case. 5.2.1. Requested Certificate Contents Various PKI management messages require that the originator of the message indicate some of the fields that are required to be present in a certificate. The CertTemplate structure allows an end entity or RA to specify as much as it wishes about the certificate it requires. CertTemplate is identical to a Certificate, but with all fields optional. Note that even if the originator completely specifies the contents of a certificate it requires, a CA is free to modify fields within the certificate actually issued. If the modified certificate is unacceptable to the requester, the requester MUST send back a certConf message that either does not include this certificate (via a CertHash), or does include this certificate (via a CertHash) along with a status of "rejected". See Section 5.3.18 for the definition and use of CertHash and the certConf message. See Appendix C and [CRMF] for CertTemplate syntax. 5.2.2. Encrypted Values Where encrypted values (restricted, in this specification, to be either private keys or certificates) are sent in PKI messages, the EncryptedValue data structure is used. Adams, et al. Standards Track [Page 31] RFC 4210 CMP September 2005 See [CRMF] for EncryptedValue syntax. Use of this data structure requires that the creator and intended recipient be able to encrypt and decrypt, respectively. Typically, this will mean that the sender and recipient have, or are able to generate, a shared secret key. If the recipient of the PKIMessage already possesses a private key usable for decryption, then the encSymmKey field MAY contain a session key encrypted using the recipient's public key. 5.2.3. Status codes and Failure Information for PKI Messages All response messages will include some status information. The following values are defined. PKIStatus ::= INTEGER { accepted (0), grantedWithMods (1), rejection (2), waiting (3), revocationWarning (4), revocationNotification (5), keyUpdateWarning (6) } Responders may use the following syntax to provide more information about failure cases. PKIFailureInfo ::= BIT STRING { badAlg (0), badMessageCheck (1), badRequest (2), badTime (3), badCertId (4), badDataFormat (5), wrongAuthority (6), incorrectData (7), missingTimeStamp (8), badPOP (9), certRevoked (10), certConfirmed (11), wrongIntegrity (12), badRecipientNonce (13), timeNotAvailable (14), unacceptedPolicy (15), unacceptedExtension (16), addInfoNotAvailable (17), Adams, et al. Standards Track [Page 32] RFC 4210 CMP September 2005 badSenderNonce (18), badCertTemplate (19), signerNotTrusted (20), transactionIdInUse (21), unsupportedVersion (22), notAuthorized (23), systemUnavail (24), systemFailure (25), duplicateCertReq (26) } PKIStatusInfo ::= SEQUENCE { status PKIStatus, statusString PKIFreeText OPTIONAL, failInfo PKIFailureInfo OPTIONAL } 5.2.4. Certificate Identification In order to identify particular certificates, the CertId data structure is used. See [CRMF] for CertId syntax. 5.2.5. Out-of-band root CA Public Key Each root CA must be able to publish its current public key via some "out-of-band" means. While such mechanisms are beyond the scope of this document, we define data structures that can support such mechanisms. There are generally two methods available: either the CA directly publishes its self-signed certificate, or this information is available via the Directory (or equivalent) and the CA publishes a hash of this value to allow verification of its integrity before use. OOBCert ::= Certificate The fields within this certificate are restricted as follows: o The certificate MUST be self-signed (i.e., the signature must be verifiable using the SubjectPublicKeyInfo field); o The subject and issuer fields MUST be identical; o If the subject field is NULL, then both subjectAltNames and issuerAltNames extensions MUST be present and have exactly the same value; Adams, et al. Standards Track [Page 33] RFC 4210 CMP September 2005 o The values of all other extensions must be suitable for a self- signed certificate (e.g., key identifiers for subject and issuer must be the same). OOBCertHash ::= SEQUENCE { hashAlg [0] AlgorithmIdentifier OPTIONAL, certId [1] CertId OPTIONAL, hashVal BIT STRING } The intention of the hash value is that anyone who has securely received the hash value (via the out-of-band means) can verify a self-signed certificate for that CA. 5.2.6. Archive Options Requesters may indicate that they wish the PKI to archive a private key value using the PKIArchiveOptions structure. See [CRMF] for PKIArchiveOptions syntax. 5.2.7. Publication Information Requesters may indicate that they wish the PKI to publish a certificate using the PKIPublicationInfo structure. See [CRMF] for PKIPublicationInfo syntax. 5.2.8. Proof-of-Possession Structures If the certification request is for a signing key pair (i.e., a request for a verification certificate), then the proof-of-possession of the private signing key is demonstrated through use of the POPOSigningKey structure. See Appendix C and [CRMF] for POPOSigningKey syntax, but note that POPOSigningKeyInput has the following semantic stipulations in this specification. POPOSigningKeyInput ::= SEQUENCE { authInfo CHOICE { sender [0] GeneralName, publicKeyMAC PKMACValue }, publicKey SubjectPublicKeyInfo } Adams, et al. Standards Track [Page 34] RFC 4210 CMP September 2005 On the other hand, if the certification request is for an encryption key pair (i.e., a request for an encryption certificate), then the proof-of-possession of the private decryption key may be demonstrated in one of three ways. 5.2.8.1. Inclusion of the Private Key By the inclusion of the private key (encrypted) in the CertRequest (in the thisMessage field of POPOPrivKey (see Appendix C) or in the PKIArchiveOptions control structure, depending upon whether or not archival of the private key is also desired). 5.2.8.2. Indirect Method By having the CA return not the certificate, but an encrypted certificate (i.e., the certificate encrypted under a randomly- generated symmetric key, and the symmetric key encrypted under the public key for which the certification request is being made) -- this is the "indirect" method mentioned previously in Section 4.3.2. The end entity proves knowledge of the private decryption key to the CA by providing the correct CertHash for this certificate in the certConf message. This demonstrates POP because the EE can only compute the correct CertHash if it is able to recover the certificate, and it can only recover the certificate if it is able to decrypt the symmetric key using the required private key. Clearly, for this to work, the CA MUST NOT publish the certificate until the certConf message arrives (when certHash is to be used to demonstrate POP). See Section 5.3.18 for further details. 5.2.8.3. Challenge-Response Protocol By having the end entity engage in a challenge-response protocol (using the messages POPODecKeyChall and POPODecKeyResp; see below) between CertReqMessages and CertRepMessage -- this is the "direct" method mentioned previously in Section 4.3.2. (This method would typically be used in an environment in which an RA verifies POP and then makes a certification request to the CA on behalf of the end entity. In such a scenario, the CA trusts the RA to have done POP correctly before the RA requests a certificate for the end entity.) The complete protocol then looks as follows (note that req' does not necessarily encapsulate req as a nested message): Adams, et al. Standards Track [Page 35] RFC 4210 CMP September 2005 EE RA CA ---- req ----> <--- chall --- ---- resp ---> ---- req' ---> <--- rep ----- ---- conf ---> <--- ack ----- <--- rep ----- ---- conf ---> <--- ack ----- This protocol is obviously much longer than the 3-way exchange given in choice (2) above, but allows a local Registration Authority to be involved and has the property that the certificate itself is not actually created until the proof-of-possession is complete. In some environments, a different order of the above messages may be required, such as the following (this may be determined by policy): EE RA CA ---- req ----> <--- chall --- ---- resp ---> ---- req' ---> <--- rep ----- <--- rep ----- ---- conf ---> ---- conf ---> <--- ack ----- <--- ack ----- If the cert. request is for a key agreement key (KAK) pair, then the POP can use any of the 3 ways described above for enc. key pairs, with the following changes: (1) the parenthetical text of bullet 2) is replaced with "(i.e., the certificate encrypted under the symmetric key derived from the CA's private KAK and the public key for which the certification request is being made)"; (2) the first parenthetical text of the challenge field of "Challenge" below is replaced with "(using PreferredSymmAlg (see Section 5.3.19.4 and Appendix E.5) and a symmetric key derived from the CA's private KAK and the public key for which the certification request is being made)". Alternatively, the POP can use the POPOSigningKey structure given in [CRMF] (where the alg field is DHBasedMAC and the signature field is the MAC) as a fourth alternative for demonstrating POP if the CA already has a D-H certificate that is known to the EE. Adams, et al. Standards Track [Page 36] RFC 4210 CMP September 2005 The challenge-response messages for proof-of-possession of a private decryption key are specified as follows (see [MvOV97], p.404 for details). Note that this challenge-response exchange is associated with the preceding cert. request message (and subsequent cert. response and confirmation messages) by the transactionID used in the PKIHeader and by the protection (MACing or signing) applied to the PKIMessage. POPODecKeyChallContent ::= SEQUENCE OF Challenge Challenge ::= SEQUENCE { owf AlgorithmIdentifier OPTIONAL, witness OCTET STRING, challenge OCTET STRING } Note that the size of Rand needs to be appropriate for encryption under the public key of the requester. Given that "int" will typically not be longer than 64 bits, this leaves well over 100 bytes of room for the "sender" field when the modulus is 1024 bits. If, in some environment, names are so long that they cannot fit (e.g., very long DNs), then whatever portion will fit should be used (as long as it includes at least the common name, and as long as the receiver is able to deal meaningfully with the abbreviation). POPODecKeyRespContent ::= SEQUENCE OF INTEGER 5.2.8.4. Summary of PoP Options The text in this section provides several options with respect to POP techniques. Using "SK" for "signing key", "EK" for "encryption key", and "KAK" for "key agreement key", the techniques may be summarized as follows: RAVerified; SKPOP; EKPOPThisMessage; KAKPOPThisMessage; KAKPOPThisMessageDHMAC; EKPOPEncryptedCert; KAKPOPEncryptedCert; EKPOPChallengeResp; and KAKPOPChallengeResp. Given this array of options, it is natural to ask how an end entity can know what is supported by the CA/RA (i.e., which options it may use when requesting certificates). The following guidelines should clarify this situation for EE implementers. Adams, et al. Standards Track [Page 37] RFC 4210 CMP September 2005 RAVerified. This is not an EE decision; the RA uses this if and only if it has verified POP before forwarding the request on to the CA, so it is not possible for the EE to choose this technique. SKPOP. If the EE has a signing key pair, this is the only POP method specified for use in the request for a corresponding certificate. EKPOPThisMessage and KAKPOPThisMessage. Whether or not to give up its private key to the CA/RA is an EE decision. If the EE decides to reveal its key, then these are the only POP methods available in this specification to achieve this (and the key pair type will determine which of these two methods to use). KAKPOPThisMessageDHMAC. The EE can only use this method if (1) the CA has a DH certificate available for this purpose, and (2) the EE already has a copy of this certificate. If both these conditions hold, then this technique is clearly supported and may be used by the EE, if desired. EKPOPEncryptedCert, KAKPOPEncryptedCert, EKPOPChallengeResp, KAKPOPChallengeResp. The EE picks one of these (in the subsequentMessage field) in the request message, depending upon preference and key pair type. The EE is not doing POP at this point; it is simply indicating which method it wants to use. Therefore, if the CA/RA replies with a "badPOP" error, the EE can re-request using the other POP method chosen in subsequentMessage. Note, however, that this specification encourages the use of the EncryptedCert choice and, furthermore, says that the challenge-response would typically be used when an RA is involved and doing POP verification. Thus, the EE should be able to make an intelligent decision regarding which of these POP methods to choose in the request message. 5.3. Operation-Specific Data Structures 5.3.1. Initialization Request An Initialization request message contains as the PKIBody a CertReqMessages data structure, which specifies the requested certificate(s). Typically, SubjectPublicKeyInfo, KeyId, and Validity are the template fields which may be supplied for each certificate requested (see Appendix D profiles for further information). This message is intended to be used for entities when first initializing into the PKI. See Appendix C and [CRMF] for CertReqMessages syntax. Adams, et al. Standards Track [Page 38] RFC 4210 CMP September 2005 5.3.2. Initialization Response An Initialization response message contains as the PKIBody an CertRepMessage data structure, which has for each certificate requested a PKIStatusInfo field, a subject certificate, and possibly a private key (normally encrypted with a session key, which is itself encrypted with the protocolEncrKey). See Section 5.3.4 for CertRepMessage syntax. Note that if the PKI Message Protection is "shared secret information" (see Section 5.1.3), then any certificate transported in the caPubs field may be directly trusted as a root CA certificate by the initiator. 5.3.3. Certification Request A Certification request message contains as the PKIBody a CertReqMessages data structure, which specifies the requested certificates. This message is intended to be used for existing PKI entities who wish to obtain additional certificates. See Appendix C and [CRMF] for CertReqMessages syntax. Alternatively, the PKIBody MAY be a CertificationRequest (this structure is fully specified by the ASN.1 structure CertificationRequest given in [PKCS10]). This structure may be required for certificate requests for signing key pairs when interoperation with legacy systems is desired, but its use is strongly discouraged whenever not absolutely necessary. 5.3.4. Certification Response A Certification response message contains as the PKIBody a CertRepMessage data structure, which has a status value for each certificate requested, and optionally has a CA public key, failure information, a subject certificate, and an encrypted private key. CertRepMessage ::= SEQUENCE { caPubs [1] SEQUENCE SIZE (1..MAX) OF Certificate OPTIONAL, response SEQUENCE OF CertResponse } CertResponse ::= SEQUENCE { certReqId INTEGER, status PKIStatusInfo, certifiedKeyPair CertifiedKeyPair OPTIONAL, rspInfo OCTET STRING OPTIONAL -- analogous to the id-regInfo-utf8Pairs string defined Adams, et al. Standards Track [Page 39] RFC 4210 CMP September 2005 -- for regInfo in CertReqMsg [CRMF] } CertifiedKeyPair ::= SEQUENCE { certOrEncCert CertOrEncCert, privateKey [0] EncryptedValue OPTIONAL, -- see [CRMF] for comment on encoding publicationInfo [1] PKIPublicationInfo OPTIONAL } CertOrEncCert ::= CHOICE { certificate [0] Certificate, encryptedCert [1] EncryptedValue } Only one of the failInfo (in PKIStatusInfo) and certificate (in CertifiedKeyPair) fields can be present in each CertResponse (depending on the status). For some status values (e.g., waiting), neither of the optional fields will be present. Given an EncryptedCert and the relevant decryption key, the certificate may be obtained. The purpose of this is to allow a CA to return the value of a certificate, but with the constraint that only the intended recipient can obtain the actual certificate. The benefit of this approach is that a CA may reply with a certificate even in the absence of a proof that the requester is the end entity that can use the relevant private key (note that the proof is not obtained until the certConf message is received by the CA). Thus, the CA will not have to revoke that certificate in the event that something goes wrong with the proof-of-possession (but MAY do so anyway, depending upon policy). 5.3.5. Key Update Request Content For key update requests the CertReqMessages syntax is used. Typically, SubjectPublicKeyInfo, KeyId, and Validity are the template fields that may be supplied for each key to be updated. This message is intended to be used to request updates to existing (non-revoked and non-expired) certificates (therefore, it is sometimes referred to as a "Certificate Update" operation). An update is a replacement certificate containing either a new subject public key or the current subject public key (although the latter practice may not be appropriate for some environments). See Appendix C and [CRMF] for CertReqMessages syntax. Adams, et al. Standards Track [Page 40] RFC 4210 CMP September 2005 5.3.6. Key Update Response Content For key update responses, the CertRepMessage syntax is used. The response is identical to the initialization response. See Section 5.3.4 for CertRepMessage syntax. 5.3.7. Key Recovery Request Content For key recovery requests the syntax used is identical to the initialization request CertReqMessages. Typically, SubjectPublicKeyInfo and KeyId are the template fields that may be used to supply a signature public key for which a certificate is required (see Appendix D profiles for further information). See Appendix C and [CRMF] for CertReqMessages syntax. Note that if a key history is required, the requester must supply a Protocol Encryption Key control in the request message. 5.3.8. Key Recovery Response Content For key recovery responses, the following syntax is used. For some status values (e.g., waiting) none of the optional fields will be present. KeyRecRepContent ::= SEQUENCE { status PKIStatusInfo, newSigCert [0] Certificate OPTIONAL, caCerts [1] SEQUENCE SIZE (1..MAX) OF Certificate OPTIONAL, keyPairHist [2] SEQUENCE SIZE (1..MAX) OF CertifiedKeyPair OPTIONAL } 5.3.9. Revocation Request Content When requesting revocation of a certificate (or several certificates), the following data structure is used. The name of the requester is present in the PKIHeader structure. RevReqContent ::= SEQUENCE OF RevDetails RevDetails ::= SEQUENCE { certDetails CertTemplate, crlEntryDetails Extensions OPTIONAL } Adams, et al. Standards Track [Page 41] RFC 4210 CMP September 2005 5.3.10. Revocation Response Content The revocation response is the response to the above message. If produced, this is sent to the requester of the revocation. (A separate revocation announcement message MAY be sent to the subject of the certificate for which revocation was requested.) RevRepContent ::= SEQUENCE { status SEQUENCE SIZE (1..MAX) OF PKIStatusInfo, revCerts [0] SEQUENCE SIZE (1..MAX) OF CertId OPTIONAL, crls [1] SEQUENCE SIZE (1..MAX) OF CertificateList OPTIONAL } 5.3.11. Cross Certification Request Content Cross certification requests use the same syntax (CertReqMessages) as normal certification requests, with the restriction that the key pair MUST have been generated by the requesting CA and the private key MUST NOT be sent to the responding CA. This request MAY also be used by subordinate CAs to get their certificates signed by the parent CA. See Appendix C and [CRMF] for CertReqMessages syntax. 5.3.12. Cross Certification Response Content Cross certification responses use the same syntax (CertRepMessage) as normal certification responses, with the restriction that no encrypted private key can be sent. See Section 5.3.4 for CertRepMessage syntax. 5.3.13. CA Key Update Announcement Content When a CA updates its own key pair, the following data structure MAY be used to announce this event. CAKeyUpdAnnContent ::= SEQUENCE { oldWithNew Certificate, newWithOld Certificate, newWithNew Certificate } Adams, et al. Standards Track [Page 42] RFC 4210 CMP September 2005 5.3.14. Certificate Announcement This structure MAY be used to announce the existence of certificates. Note that this message is intended to be used for those cases (if any) where there is no pre-existing method for publication of certificates; it is not intended to be used where, for example, X.500 is the method for publication of certificates. CertAnnContent ::= Certificate 5.3.15. Revocation Announcement When a CA has revoked, or is about to revoke, a particular certificate, it MAY issue an announcement of this (possibly upcoming) event. RevAnnContent ::= SEQUENCE { status PKIStatus, certId CertId, willBeRevokedAt GeneralizedTime, badSinceDate GeneralizedTime, crlDetails Extensions OPTIONAL } A CA MAY use such an announcement to warn (or notify) a subject that its certificate is about to be (or has been) revoked. This would typically be used where the request for revocation did not come from the subject concerned. The willBeRevokedAt field contains the time at which a new entry will be added to the relevant CRLs. 5.3.16. CRL Announcement When a CA issues a new CRL (or set of CRLs) the following data structure MAY be used to announce this event. CRLAnnContent ::= SEQUENCE OF CertificateList 5.3.17. PKI Confirmation Content This data structure is used in the protocol exchange as the final PKIMessage. Its content is the same in all cases -- actually there is no content since the PKIHeader carries all the required information. PKIConfirmContent ::= NULL Adams, et al. Standards Track [Page 43] RFC 4210 CMP September 2005 Use of this message for certificate confirmation is NOT RECOMMENDED; certConf SHOULD be used instead. Upon receiving a PKIConfirm for a certificate response, the recipient MAY treat it as a certConf with all certificates being accepted. 5.3.18. Certificate Confirmation Content This data structure is used by the client to send a confirmation to the CA/RA to accept or reject certificates. CertConfirmContent ::= SEQUENCE OF CertStatus CertStatus ::= SEQUENCE { certHash OCTET STRING, certReqId INTEGER, statusInfo PKIStatusInfo OPTIONAL } For any particular CertStatus, omission of the statusInfo field indicates ACCEPTANCE of the specified certificate. Alternatively, explicit status details (with respect to acceptance or rejection) MAY be provided in the statusInfo field, perhaps for auditing purposes at the CA/RA. Within CertConfirmContent, omission of a CertStatus structure corresponding to a certificate supplied in the previous response message indicates REJECTION of the certificate. Thus, an empty CertConfirmContent (a zero-length SEQUENCE) MAY be used to indicate rejection of all supplied certificates. See Section 5.2.8, item (2), for a discussion of the certHash field with respect to proof-of- possession. 5.3.19. PKI General Message Content InfoTypeAndValue ::= SEQUENCE { infoType OBJECT IDENTIFIER, infoValue ANY DEFINED BY infoType OPTIONAL } -- where {id-it} = {id-pkix 4} = {1 3 6 1 5 5 7 4} GenMsgContent ::= SEQUENCE OF InfoTypeAndValue 5.3.19.1. CA Protocol Encryption Certificate This MAY be used by the EE to get a certificate from the CA to use to protect sensitive information during the protocol. Adams, et al. Standards Track [Page 44] RFC 4210 CMP September 2005 GenMsg: {id-it 1}, < absent > GenRep: {id-it 1}, Certificate | < absent > EEs MUST ensure that the correct certificate is used for this purpose. 5.3.19.2. Signing Key Pair Types This MAY be used by the EE to get the list of signature algorithms (e.g., RSA, DSA) whose subject public key values the CA is willing to certify. Note that for the purposes of this exchange, rsaEncryption and rsaWithSHA1, for example, are considered to be equivalent; the question being asked is, "Is the CA willing to certify an RSA public key?" GenMsg: {id-it 2}, < absent > GenRep: {id-it 2}, SEQUENCE SIZE (1..MAX) OF AlgorithmIdentifier 5.3.19.3. Encryption/Key Agreement Key Pair Types This MAY be used by the client to get the list of encryption/key agreement algorithms whose subject public key values the CA is willing to certify. GenMsg: {id-it 3}, < absent > GenRep: {id-it 3}, SEQUENCE SIZE (1..MAX) OF AlgorithmIdentifier 5.3.19.4. Preferred Symmetric Algorithm This MAY be used by the client to get the CA-preferred symmetric encryption algorithm for any confidential information that needs to be exchanged between the EE and the CA (for example, if the EE wants to send its private decryption key to the CA for archival purposes). GenMsg: {id-it 4}, < absent > GenRep: {id-it 4}, AlgorithmIdentifier 5.3.19.5. Updated CA Key Pair This MAY be used by the CA to announce a CA key update event. GenMsg: {id-it 5}, CAKeyUpdAnnContent Adams, et al. Standards Track [Page 45] RFC 4210 CMP September 2005 5.3.19.6. CRL This MAY be used by the client to get a copy of the latest CRL. GenMsg: {id-it 6}, < absent > GenRep: {id-it 6}, CertificateList 5.3.19.7. Unsupported Object Identifiers This is used by the server to return a list of object identifiers that it does not recognize or support from the list submitted by the client. GenRep: {id-it 7}, SEQUENCE SIZE (1..MAX) OF OBJECT IDENTIFIER 5.3.19.8. Key Pair Parameters This MAY be used by the EE to request the domain parameters to use for generating the key pair for certain public-key algorithms. It can be used, for example, to request the appropriate P, Q, and G to generate the DH/DSA key, or to request a set of well-known elliptic curves. GenMsg: {id-it 10}, OBJECT IDENTIFIER -- (Algorithm object-id) GenRep: {id-it 11}, AlgorithmIdentifier | < absent > An absent infoValue in the GenRep indicates that the algorithm specified in GenMsg is not supported. EEs MUST ensure that the parameters are acceptable to it and that the GenRep message is authenticated (to avoid substitution attacks). 5.3.19.9. Revocation Passphrase This MAY be used by the EE to send a passphrase to a CA/RA for the purpose of authenticating a later revocation request (in the case that the appropriate signing private key is no longer available to authenticate the request). See Appendix B for further details on the use of this mechanism. GenMsg: {id-it 12}, EncryptedValue GenRep: {id-it 12}, < absent > 5.3.19.10. ImplicitConfirm See Section 5.1.1.1 for the definition and use of {id-it 13}. Adams, et al. Standards Track [Page 46] RFC 4210 CMP September 2005 5.3.19.11. ConfirmWaitTime See Section 5.1.1.2 for the definition and use of {id-it 14}. 5.3.19.12 Original PKIMessage See Section 5.1.3 for the definition and use of {id-it 15}. 5.3.19.13. Supported Language Tags This MAY be used to determine the appropriate language tag to use in subsequent messages. The sender sends its list of supported languages (in order, most preferred to least); the receiver returns the one it wishes to use. (Note: each UTF8String MUST include a language tag.) If none of the offered tags are supported, an error MUST be returned. GenMsg: {id-it 16}, SEQUENCE SIZE (1..MAX) OF UTF8String GenRep: {id-it 16}, SEQUENCE SIZE (1) OF UTF8String 5.3.20. PKI General Response Content GenRepContent ::= SEQUENCE OF InfoTypeAndValue Examples of GenReps that MAY be supported include those listed in the subsections of Section 5.3.19. 5.3.21. Error Message Content This data structure MAY be used by EE, CA, or RA to convey error info. ErrorMsgContent ::= SEQUENCE { pKIStatusInfo PKIStatusInfo, errorCode INTEGER OPTIONAL, errorDetails PKIFreeText OPTIONAL } This message MAY be generated at any time during a PKI transaction. If the client sends this request, the server MUST respond with a PKIConfirm response, or another ErrorMsg if any part of the header is not valid. Both sides MUST treat this message as the end of the transaction (if a transaction is in progress). If protection is desired on the message, the client MUST protect it using the same technique (i.e., signature or MAC) as the starting message of the transaction. The CA MUST always sign it with a signature key. Adams, et al. Standards Track [Page 47] RFC 4210 CMP September 2005 5.3.22. Polling Request and Response This pair of messages is intended to handle scenarios in which the client needs to poll the server in order to determine the status of an outstanding ir, cr, or kur transaction (i.e., when the "waiting" PKIStatus has been received). PollReqContent ::= SEQUENCE OF SEQUENCE { certReqId INTEGER } PollRepContent ::= SEQUENCE OF SEQUENCE { certReqId INTEGER, checkAfter INTEGER, -- time in seconds reason PKIFreeText OPTIONAL } The following clauses describe when polling messages are used, and how they are used. It is assumed that multiple certConf messages can be sent during transactions. There will be one sent in response to each ip, cp, or kup that contains a CertStatus for an issued certificate. 1. In response to an ip, cp, or kup message, an EE will send a certConf for all issued certificates and, following the ack, a pollReq for all pending certificates. 2. In response to a pollReq, a CA/RA will return an ip, cp, or kup if one or more of the pending certificates is ready; otherwise, it will return a pollRep. 3. If the EE receives a pollRep, it will wait for at least as long as the checkAfter value before sending another pollReq. 4. If an ip, cp, or kup is received in response to a pollReq, then it will be treated in the same way as the initial response. Adams, et al. Standards Track [Page 48] RFC 4210 CMP September 2005 START | v Send ir | ip v Check status of returned <------------------------+ certs | | | +------------------------>|<------------------+ | | | | | | (issued) v (waiting) | | Add to <----------- Check CertResponse ------> Add to | conf list for each certificate pending list | / | / | (conf list) / (empty conf list) | / ip | / +----------------+ (empty pending list) / | pRep END <---- Send certConf Send pReq------------>Wait | ^ ^ | | | | | +-----------------+ +---------------+ (pending list) Adams, et al. Standards Track [Page 49] RFC 4210 CMP September 2005 In the following exchange, the end entity is enrolling for two certificates in one request. Step End Entity PKI -------------------------------------------------------------------- 1 Format ir 2 -> ir -> 3 Handle ir 4 Manual intervention is required for both certs. 5 <- ip <- 6 Process ip 7 Format pReq 8 -> pReq -> 9 Check status of cert requests 10 Certificates not ready 11 Format pRep 12 <- pRep <- 13 Wait 14 Format pReq 15 -> pReq -> 16 Check status of cert requests 17 One certificate is ready 18 Format ip 19 <- ip <- 20 Handle ip 21 Format certConf 22 -> certConf -> 23 Handle certConf 24 Format ack 25 <- pkiConf <- 26 Format pReq 27 -> pReq -> 28 Check status of certificate 29 Certificate is ready 30 Format ip 31 <- ip <- 31 Handle ip 32 Format certConf 33 -> certConf -> 34 Handle certConf 35 Format ack 36 <- pkiConf <- Adams, et al. Standards Track [Page 50] RFC 4210 CMP September 2005 6. Mandatory PKI Management Functions Some of the PKI management functions outlined in Section 3.1 above are described in this section. This section deals with functions that are "mandatory" in the sense that all end entity and CA/RA implementations MUST be able to provide the functionality described. This part is effectively the profile of the PKI management functionality that MUST be supported. Note, however, that the management functions described in this section do not need to be accomplished using the PKI messages defined in Section 5 if alternate means are suitable for a given environment (see Appendix D for profiles of the PKIMessages that MUST be supported). 6.1. Root CA Initialization [See Section 3.1.1.2 for this document's definition of "root CA".] A newly created root CA must produce a "self-certificate", which is a Certificate structure with the profile defined for the "newWithNew" certificate issued following a root CA key update. In order to make the CA's self certificate useful to end entities that do not acquire the self certificate via "out-of-band" means, the CA must also produce a fingerprint for its certificate. End entities that acquire this fingerprint securely via some "out-of-band" means can then verify the CA's self-certificate and, hence, the other attributes contained therein. The data structure used to carry the fingerprint is the OOBCertHash. 6.2. Root CA Key Update CA keys (as all other keys) have a finite lifetime and will have to be updated on a periodic basis. The certificates NewWithNew, NewWithOld, and OldWithNew (see Section 4.4.1) MAY be issued by the CA to aid existing end entities who hold the current self-signed CA certificate (OldWithOld) to transition securely to the new self- signed CA certificate (NewWithNew), and to aid new end entities who will hold NewWithNew to acquire OldWithOld securely for verification of existing data. 6.3. Subordinate CA Initialization [See Section 3.1.1.2 for this document's definition of "subordinate CA".] Adams, et al. Standards Track [Page 51] RFC 4210 CMP September 2005 From the perspective of PKI management protocols, the initialization of a subordinate CA is the same as the initialization of an end entity. The only difference is that the subordinate CA must also produce an initial revocation list. 6.4. CRL production Before issuing any certificates, a newly established CA (which issues CRLs) must produce "empty" versions of each CRL which are to be periodically produced. 6.5. PKI Information Request When a PKI entity (CA, RA, or EE) wishes to acquire information about the current status of a CA, it MAY send that CA a request for such information. The CA MUST respond to the request by providing (at least) all of the information requested by the requester. If some of the information cannot be provided, then an error must be conveyed to the requester. If PKIMessages are used to request and supply this PKI information, then the request MUST be the GenMsg message, the response MUST be the GenRep message, and the error MUST be the Error message. These messages are protected using a MAC based on shared secret information (i.e., PasswordBasedMAC) or using any other authenticated means (if the end entity has an existing certificate). 6.6. Cross Certification The requester CA is the CA that will become the subject of the cross-certificate; the responder CA will become the issuer of the cross-certificate. The requester CA must be "up and running" before initiating the cross-certification operation. 6.6.1. One-Way Request-Response Scheme: The cross-certification scheme is essentially a one way operation; that is, when successful, this operation results in the creation of one new cross-certificate. If the requirement is that cross- certificates be created in "both directions", then each CA, in turn, must initiate a cross-certification operation (or use another scheme). Adams, et al. Standards Track [Page 52] RFC 4210 CMP September 2005 This scheme is suitable where the two CAs in question can already verify each other's signatures (they have some common points of trust) or where there is an out-of-band verification of the origin of the certification request. Detailed Description: Cross certification is initiated at one CA known as the responder. The CA administrator for the responder identifies the CA it wants to cross certify and the responder CA equipment generates an authorization code. The responder CA administrator passes this authorization code by out-of-band means to the requester CA administrator. The requester CA administrator enters the authorization code at the requester CA in order to initiate the on- line exchange. The authorization code is used for authentication and integrity purposes. This is done by generating a symmetric key based on the authorization code and using the symmetric key for generating Message Authentication Codes (MACs) on all messages exchanged. (Authentication may alternatively be done using signatures instead of MACs, if the CAs are able to retrieve and validate the required public keys by some means, such as an out-of-band hash comparison.) The requester CA initiates the exchange by generating a cross- certification request (ccr) with a fresh random number (requester random number). The requester CA then sends the ccr message to the responder CA. The fields in this message are protected from modification with a MAC based on the authorization code. Upon receipt of the ccr message, the responder CA validates the message and the MAC, saves the requester random number, and generates its own random number (responder random number). It then generates (and archives, if desired) a new requester certificate that contains the requester CA public key and is signed with the responder CA signature private key. The responder CA responds with the cross certification response (ccp) message. The fields in this message are protected from modification with a MAC based on the authorization code. Upon receipt of the ccp message, the requester CA validates the message (including the received random numbers) and the MAC. The requester CA responds with the certConf message. The fields in this message are protected from modification with a MAC based on the authorization code. The requester CA MAY write the requester certificate to the Repository as an aid to later certificate path construction. Adams, et al. Standards Track [Page 53] RFC 4210 CMP September 2005 Upon receipt of the certConf message, the responder CA validates the message and the MAC, and sends back an acknowledgement using the PKIConfirm message. It MAY also publish the requester certificate as an aid to later path construction. Notes: 1. The ccr message must contain a "complete" certification request; that is, all fields except the serial number (including, e.g., a BasicConstraints extension) must be specified by the requester CA. 2. The ccp message SHOULD contain the verification certificate of the responder CA; if present, the requester CA must then verify this certificate (for example, via the "out-of-band" mechanism). (A simpler, non-interactive model of cross-certification may also be envisioned, in which the issuing CA acquires the subject CA's public key from some repository, verifies it via some out-of-band mechanism, and creates and publishes the cross-certificate without the subject CA's explicit involvement. This model may be perfectly legitimate for many environments, but since it does not require any protocol message exchanges, its detailed description is outside the scope of this specification.) 6.7. End Entity Initialization As with CAs, end entities must be initialized. Initialization of end entities requires at least two steps: o acquisition of PKI information o out-of-band verification of one root-CA public key (other possible steps include the retrieval of trust condition information and/or out-of-band verification of other CA public keys). 6.7.1. Acquisition of PKI Information The information REQUIRED is: o the current root-CA public key o (if the certifying CA is not a root-CA) the certification path from the root CA to the certifying CA together with appropriate revocation lists Adams, et al. Standards Track [Page 54] RFC 4210 CMP September 2005 o the algorithms and algorithm parameters that the certifying CA supports for each relevant usage Additional information could be required (e.g., supported extensions or CA policy information) in order to produce a certification request that will be successful. However, for simplicity we do not mandate that the end entity acquires this information via the PKI messages. The end result is simply that some certification requests may fail (e.g., if the end entity wants to generate its own encryption key, but the CA doesn't allow that). The required information MAY be acquired as described in Section 6.5. 6.7.2. Out-of-Band Verification of Root-CA Key An end entity must securely possess the public key of its root CA. One method to achieve this is to provide the end entity with the CA's self-certificate fingerprint via some secure "out-of-band" means. The end entity can then securely use the CA's self-certificate. See Section 6.1 for further details. 6.8. Certificate Request An initialized end entity MAY request an additional certificate at any time (for any purpose). This request will be made using the certification request (cr) message. If the end entity already possesses a signing key pair (with a corresponding verification certificate), then this cr message will typically be protected by the entity's digital signature. The CA returns the new certificate (if the request is successful) in a CertRepMessage. 6.9. Key Update When a key pair is due to expire, the relevant end entity MAY request a key update; that is, it MAY request that the CA issue a new certificate for a new key pair (or, in certain circumstances, a new certificate for the same key pair). The request is made using a key update request (kur) message (referred to, in some environments, as a "Certificate Update" operation). If the end entity already possesses a signing key pair (with a corresponding verification certificate), then this message will typically be protected by the entity's digital signature. The CA returns the new certificate (if the request is successful) in a key update response (kup) message, which is syntactically identical to a CertRepMessage. Adams, et al. Standards Track [Page 55] RFC 4210 CMP September 2005 7. Version Negotiation This section defines the version negotiation used to support older protocols between client and servers. If a client knows the protocol version(s) supported by the server (e.g., from a previous PKIMessage exchange or via some out-of-band means), then it MUST send a PKIMessage with the highest version supported by both it and the server. If a client does not know what version(s) the server supports, then it MUST send a PKIMessage using the highest version it supports. If a server receives a message with a version that it supports, then the version of the response message MUST be the same as the received version. If a server receives a message with a version higher or lower than it supports, then it MUST send back an ErrorMsg with the unsupportedVersion bit set (in the failureInfo field of the pKIStatusInfo). If the received version is higher than the highest supported version, then the version in the error message MUST be the highest version the server supports; if the received version is lower than the lowest supported version then the version in the error message MUST be the lowest version the server supports. If a client gets back an ErrorMsgContent with the unsupportedVersion bit set and a version it supports, then it MAY retry the request with that version. 7.1. Supporting RFC 2510 Implementations RFC 2510 did not specify the behaviour of implementations receiving versions they did not understand since there was only one version in existence. With the introduction of the present revision of the specification, the following versioning behaviour is recommended. 7.1.1. Clients Talking to RFC 2510 Servers If, after sending a cmp2000 message, a client receives an ErrorMsgContent with a version of cmp1999, then it MUST abort the current transaction. It MAY subsequently retry the transaction using version cmp1999 messages. If a client receives a non-error PKIMessage with a version of cmp1999, then it MAY decide to continue the transaction (if the transaction hasn't finished) using RFC 2510 semantics. If it does not choose to do so and the transaction is not finished, then it MUST abort the transaction and send an ErrorMsgContent with a version of cmp1999. Adams, et al. Standards Track [Page 56] RFC 4210 CMP September 2005 7.1.2. Servers Receiving Version cmp1999 PKIMessages If a server receives a version cmp1999 message it MAY revert to RFC 2510 behaviour and respond with version cmp1999 messages. If it does not choose to do so, then it MUST send back an ErrorMsgContent as described above in Section 7. 8. Security Considerations 8.1. Proof-Of-Possession with a Decryption Key Some cryptographic considerations are worth explicitly spelling out. In the protocols specified above, when an end entity is required to prove possession of a decryption key, it is effectively challenged to decrypt something (its own certificate). This scheme (and many others!) could be vulnerable to an attack if the possessor of the decryption key in question could be fooled into decrypting an arbitrary challenge and returning the cleartext to an attacker. Although in this specification a number of other failures in security are required in order for this attack to succeed, it is conceivable that some future services (e.g., notary, trusted time) could potentially be vulnerable to such attacks. For this reason, we re- iterate the general rule that implementations should be very careful about decrypting arbitrary "ciphertext" and revealing recovered "plaintext" since such a practice can lead to serious security vulnerabilities. 8.2. Proof-Of-Possession by Exposing the Private Key Note also that exposing a private key to the CA/RA as a proof-of- possession technique can carry some security risks (depending upon whether or not the CA/RA can be trusted to handle such material appropriately). Implementers are advised to: Exercise caution i