Network Working Group                                            S. Kent
Request for Comments: 1422                                           BBN
Obsoletes: 1114                                  IAB IRTF PSRG, IETF PEM
                                                          February 1993


          Privacy Enhancement for Internet Electronic Mail:
              Part II: Certificate-Based Key Management

Status of this Memo

  This RFC specifies an IAB standards track protocol for the Internet
  community, and requests discussion and suggestions for improvements.
  Please refer to the current edition of the "IAB Official Protocol
  Standards" for the standardization state and status of this protocol.
  Distribution of this memo is unlimited.

Acknowledgements

  This memo is the outgrowth of a series of meetings of the Privacy and
  Security Research Group of the Internet Research Task Force (IRTF)
  and the Privacy-Enhanced Electronic Mail Working Group of the
  Internet Engineering Task Force (IETF).  I would like to thank the
  members of the PSRG and the PEM WG for their comments and
  contributions at the meetings which led to the preparation of this
  document.  I also would like to thank contributors to the PEM-DEV
  mailing list who have provided valuable input which is reflected in
  this memo.

1.  Executive Summary

  This is one of a series of documents defining privacy enhancement
  mechanisms for electronic mail transferred using Internet mail
  protocols.  RFC 1421 [6] prescribes protocol extensions and
  processing procedures for RFC-822 mail messages, given that suitable
  cryptographic keys are held by originators and recipients as a
  necessary precondition.  RFC 1423 [7] specifies algorithms, modes and
  associated identifiers for use in processing privacy-enhanced
  messages, as called for in RFC 1421 and this document.  This document
  defines a supporting key management architecture and infrastructure,
  based on public-key certificate techniques, to provide keying
  information to message originators and recipients.  RFC 1424 [8]
  provides additional specifications for services in conjunction with
  the key management infrastructure described herein.

  The key management architecture described in this document is
  compatible with the authentication framework described in CCITT 1988
  X.509 [2].  This document goes beyond X.509 by establishing



Kent                                                            [Page 1]

RFC 1422           Certificate-Based Key Management        February 1993


  procedures and conventions for a key management infrastructure for
  use with Privacy Enhanced Mail (PEM) and with other protocols, from
  both the TCP/IP and OSI suites, in the future.  There are several
  motivations for establishing these procedures and conventions (as
  opposed to relying only on the very general framework outlined in
  X.509):

      -It is important that a certificate management infrastructure
          for use in the Internet community accommodate a range of
          clearly-articulated certification policies for both users
          and   organizations in a well-architected fashion.
          Mechanisms must be provided to enable each user to be
          aware of the policies governing any certificate which the
          user may encounter.  This requires the introduction
          and standardization of procedures and conventions that are
          outside the scope of X.509.

      -The procedures for authenticating originators and recipient in
          the course of message submission and delivery should be
          simple, automated and uniform despite the existence of
          differing certificate management policies.  For example,
          users should not have to engage in careful examination of a
          complex set of certification relationships in order to
          evaluate the credibility of a claimed identity.

      -The authentication framework defined by X.509 is designed to
          operate in the X.500 directory server environment.  However
          X.500 directory servers are not expected to be ubiquitous
          in the Internet in the near future, so some conventions
          are adopted to facilitate operation of the key management
          infrastructure in the near term.

      -Public key cryptosystems are central to the authentication
          technology of X.509 and those which enjoy the most
          widespread use are patented in the U.S.  Although this
          certification management scheme is compatible with
          the use of different digital signature algorithms, it is
          anticipated that the RSA cryptosystem will be used as
          the primary signature algorithm in establishing the
          Internet certification hierarchy.  Special license
          arrangements have been made to facilitate the
          use of this algorithm in the U.S. portion of Internet
          environment.

  The infrastructure specified in this document establishes a single
  root for all certification within the Internet, the Internet Policy
  Registration Authority (IPRA).  The IPRA establishes global policies,
  described in this document, which apply to all certification effected



Kent                                                            [Page 2]

RFC 1422           Certificate-Based Key Management        February 1993


  under this hierarchy.  Beneath IPRA root are Policy Certification
  Authorities (PCAs), each of which establishes and publishes (in the
  form of an informational RFC) its policies for registration of users
  or organizations.  Each PCA is certified by the IPRA. (It is
  desirable that there be a relatively small number of PCAs, each with
  a substantively different policy, to facilitate user familiarity with
  the set of PCA policies.  However there is no explicit requirement
  that the set of PCAs be limited in this fashion.)  Below PCAs,
  Certification Authorities (CAs) will be established to certify users
  and subordinate organizational entities (e.g., departments, offices,
  subsidiaries, etc.).  Initially, we expect the majority of users will
  be registered via organizational affiliation, consistent with current
  practices for how most user mailboxes are provided.  In this sense
  the registration is analogous to the issuance of a university or
  company ID card.

  Some CAs are expected to provide certification for residential users
  in support of users who wish to register independent of any
  organizational affiliation.  Over time, we anticipate that civil
  government entities which  already provide analogous identification
  services in other contexts, e.g.,  driver's licenses, may provide
  this service.  For users who wish anonymity while taking advantage of
  PEM privacy facilities, one or more PCAs will be established with
  policies that allow for registration of users, under subordinate CAs,
  who do not wish to disclose their identities.

2.  Overview of Approach

  This document defines a key management architecture based on the use
  of public-key certificates, primarily in support of the message
  encipherment and authentication procedures defined in RFC 1421.  The
  concept of public-key certificates is defined in X.509 and this
  architecture is a compliant subset of that envisioned in X.509.

  Briefly, a (public-key) certificate is a data structure which
  contains the name of a user (the "subject"), the public component
  (This document adopts the terms "private component" and "public
  component" to refer to the quantities which are, respectively, kept
  secret and made publicly available in asymmetric cryptosystems.  This
  convention is adopted to avoid possible confusion arising from use of
  the term "secret key" to refer to either the former quantity or to a
  key in a symmetric cryptosystem.)  of that user, and the name of an
  entity (the "issuer") which vouches that the public component is
  bound to the named user.  This data, along with a time interval over
  which the binding is claimed to be valid, is cryptographically signed
  by the issuer using the issuer's private component.  The subject and
  issuer names in certificates are Distinguished Names (DNs) as defined
  in the directory system (X.500).



Kent                                                            [Page 3]

RFC 1422           Certificate-Based Key Management        February 1993


  Once signed, certificates can be stored in directory servers,
  transmitted via non-secure message exchanges, or distributed via any
  other means that make certificates easily accessible to message
  system users, without regard for the security of the transmission
  medium.  Certificates are used in PEM to provide the originator of a
  message with the (authenticated) public component of each recipient
  and to provide each recipient with the (authenticated) public
  component of the originator.  The following brief discussion
  illustrates the procedures for both originator and recipients.

  Prior to sending an encrypted message (using PEM), an originator must
  acquire a certificate for each recipient and must validate these
  certificates.  Briefly, validation is performed by checking the
  digital signature in the certificate, using the public component of
  the issuer whose private component was used to sign the certificate.
  The issuer's public component is made available via some out of band
  means (for the IPRA) or is itself distributed in a certificate to
  which this validation procedure is applied recursively.  In the
  latter case, the issuer of a user's certificate becomes the subject
  in a certificate issued by another certifying authority (or a PCA),
  thus giving rise to a certification hierarchy.  The validity interval
  for each certificate is checked and Certificate Revocation Lists
  (CRLs) are checked to ensure that none of the certificates employed
  in the validation process has been revoked by an issuer.

  Once a certificate for a recipient is validated, the public component
  contained in the certificate is extracted and used to encrypt the
  data encryption key (DEK), which, in turn, is used to encrypt the
  message itself.  The resulting encrypted DEK is incorporated into the
  Key-Info field of the message header.  Upon receipt of an encrypted
  message, a recipient employs his private component to decrypt this
  field, extracting the DEK, and then uses this DEK to decrypt the
  message.

  In order to provide message integrity and data origin authentication,
  the originator generates a message integrity code (MIC), signs
  (encrypts) the MIC using the private component of his public-key
  pair, and includes the resulting value in the message header in the
  MIC-Info field.  The certificate of the originator is (optionally)
  included in the header in the Certificate field as described in RFC
  1421.  This is done in order to facilitate validation in the absence
  of ubiquitous directory services.  Upon receipt of a privacy enhanced
  message, a recipient validates the originator's certificate (using
  the IPRA public component as the root of a certification path),
  checks to ensure that it has not been revoked, extracts the public
  component from the certificate, and uses that value to recover
  (decrypt) the MIC.  The recovered MIC is compared against the locally
  calculated MIC to verify the integrity and data origin authenticity



Kent                                                            [Page 4]

RFC 1422           Certificate-Based Key Management        February 1993


  of the message.

3.  Architecture

  3.1  Scope and Restrictions

  The architecture described below is intended to provide a basis for
  managing public-key cryptosystem values in support of privacy
  enhanced electronic mail in the Internet environment.  The
  architecture describes procedures for registering certification
  authorities and users, for generating and distributing certificates,
  and for generating and distributing CRLs.  RFC 1421 describes the
  syntax and semantics of header fields used to transfer certificates
  and to represent the DEK and MIC in this public-key context.
  Definitions of the algorithms, modes of use and associated
  identifiers are separated in RFC 1423 to facilitate the adoption of
  additional algorithms in the future.  This document focuses on the
  management aspects of certificate-based, public-key cryptography for
  privacy enhanced mail.

  The proposed architecture imposes conventions for the certification
  hierarchy which are not strictly required by the X.509 recommendation
  nor by the technology itself.  These conventions are motivated by
  several factors, primarily the need for authentication semantics
  compatible with automated validation and the automated determination
  of the policies under which certificates are issued.

  Specifically, the architecture proposes a system in which user (or
  mailing list) certificates represent the leaves in a certification
  hierarchy.  This certification hierarchy is largely isomorphic to the
  X.500 directory naming hierarchy, with two exceptions: the IPRA forms
  the root of the tree (the root of the X.500 DIT is not instantiated
  as a node), and a number of Policy Certification Authorities (PCAs)
  form the "roots" of subtrees, each of which represents a different
  certification policy.

  Not every level in the directory hierarchy need correspond to a
  certification authority.  For example, the appearance of geographic
  entities in a distinguished name (e.g., countries, states, provinces,
  localities) does not require that various governments become
  certifying authorities in order to instantiate this architecture.
  However, it is anticipated that, over time, a number of such points
  in the hierarchy will be instantiated as CAs in order to simplify
  later transition of management to appropriate governmental
  authorities.

  These conventions minimize the complexity of validating user
  certificates, e.g., by making explicit the relationship between a



Kent                                                            [Page 5]

RFC 1422           Certificate-Based Key Management        February 1993


  certificate issuer and the user (via the naming hierarchy). Note that
  in this architecture, only PCAs may be certified by the IPRA, and
  every CA's certification path can be traced to a PCA, through zero or
  more CAs.  If a CA is certified by more than one PCA, each
  certificate issued by a PCA for the CA must contain a distinct public
  component.  These conventions result in a certification hierarchy
  which is a compatible subset of that permitted under X.509, with
  respect to both syntax and semantics.

  Although the key management architecture described in this document
  has been designed primarily to support privacy enhanced mail, this
  infrastructure also may, in principle, be used to support X.400 mail
  security facilities (as per 1988 X.411) and X.500 directory
  authentication facilities.  Thus, establishment of this
  infrastructure paves the way for use of these and other OSI protocols
  in the Internet in the future.  In the future, these certificates
  also may be employed in the provision of security services in other
  protocols in the TCP/IP and OSI suites as well.

  3.2  Relation to X.509 Architecture

  CCITT 1988 Recommendation X.509, "The Directory - Authentication
  Framework", defines a framework for authentication of entities
  involved in a distributed directory service.  Strong authentication,
  as defined in X.509, is accomplished with the use of public-key
  cryptosystems.  Unforgeable certificates are generated by
  certification authorities; these authorities may be organized
  hierarchically, though such organization is not required by X.509.
  There is no implied mapping between a certification hierarchy and the
  naming hierarchy imposed by directory system naming attributes.

  This document interprets the X.509 certificate mechanism to serve the
  needs of PEM in the Internet environment.  The certification
  hierarchy proposed in this document in support of privacy enhanced
  mail is intentionally a subset of that allowed under X.509.  This
  certification hierarchy also embodies semantics which are not
  explicitly addressed by X.509, but which are consistent with X.509
  precepts.  An overview of the rationale for these semantics is
  provided in Section 1.

  3.3  Certificate Definition

  Certificates are central to the key management architecture for X.509
  and PEM.  This section provides an overview of the syntax and a
  description of the semantics of certificates.  Appendix A includes
  the ASN.1 syntax for certificates.   A certificate includes the
  following contents:




Kent                                                            [Page 6]

RFC 1422           Certificate-Based Key Management        February 1993


      1.  version

      2.  serial number

      3.  signature (algorithm ID and parameters)

      4.  issuer name

      5.  validity period

      6.  subject name

      7.  subject public key (and associated algorithm ID)

  3.3.1  Version Number

  The version number field is intended to facilitate orderly changes in
  certificate formats over time.  The initial version number for
  certificates used in PEM is the X.509 default which has a value of
  zero (0), indicating the 1988 version.  PEM implementations are
  encouraged to accept later versions as they are endorsed by
  CCITT/ISO.

  3.3.2  Serial Number

  The serial number field provides a short form, unique identifier for
  each certificate generated by an issuer.  An issuer must ensure that
  no two distinct certificates with the same issuer DN contain the same
  serial number.  (This requirement must be met even when the
  certification function is effected on a distributed basis and/or when
  the same issuer DN is certified under two different PCAs.  This is
  especially critical for residential CAs certified under different
  PCAs.) The serial number is used in CRLs to identify revoked
  certificates, as described in Section 3.4.3.4.  Although this
  attribute is an integer, PEM UA processing of this attribute need not
  involve any arithmetic operations.  All PEM UA implementations must
  be capable of processing serial numbers at least 128 bits in length,
  and size-independent support serial numbers is encouraged.

  3.3.3  Signature

  This field specifies the algorithm used by the issuer to sign the
  certificate, and any parameters associated with the algorithm. (The
  certificate signature is appended to the data structure, as defined
  by the signature macro in X.509.  This algorithm identification
  information is replicated with the signature.)  The signature is
  validated by the UA processing a certificate, in order to determine
  that the integrity of its contents have not been modified subsequent



Kent                                                            [Page 7]

RFC 1422           Certificate-Based Key Management        February 1993


  to signing by a CA (IPRA, or PCA).  In this context, a signature is
  effected through the use of a Certificate Integrity Check (CIC)
  algorithm and a public-key encryption algorithm.  RFC 1423 contains
  the definitions and algorithm IDs for signature algorithms employed
  in this architecture.

  3.3.4  Subject Name

  A certificate provides a representation of its subject's identity in
  the form of a Distinguished Name (DN).  The fundamental binding
  ensured by the key management architecture is that between the public
  component and the user's identity in this form.  A distinguished name
  is an X.500 directory system concept and if a user is already
  registered in an X.500 directory, his distinguished name is defined
  via that registration.  Users who are not registered in a directory
  should keep in mind likely directory naming structure (schema) when
  selecting a distinguished name for inclusion in a certificate.

  3.3.5  Issuer Name

  A certificate provides a representation of its issuer's identity, in
  the form of a Distinguished Name.  The issuer identification is used
  to select the appropriate issuer public component to employ in
  performing certificate validation.  (If an issuer (CA) is certified
  by multiple PCAs, then the issuer DN does not uniquely identify the
  public component used to sign the certificate.  In such circumstances
  it may be necessary to attempt certificate validation using multiple
  public components, from certificates held by the issuer under
  different PCAs.  If the 1992 version of a certificate is employed,
  the issuer may employ distinct issuer UIDs in the certificates it
  issues, to further facilitate selection of the right issuer public
  component.) The issuer is the certifying authority (IPRA, PCA or CA)
  who vouches for the binding between the subject identity and the
  public key contained in the certificate.

  3.3.6  Validity Period

  A certificate carries a pair of date and time indications, indicating
  the start and end of the time period over which a certificate is
  intended to be used.  The duration of the interval may be constant
  for all user certificates issued by a given CA or it might differ
  based on the nature of the user's affiliation.  For example, an
  organization might issue certificates with shorter intervals to
  temporary employees versus permanent employees.  It is recommended
  that the UTCT (Coordinated Universal Time) values recorded here
  specify granularity to no more than the minute, even though finer
  granularity can be expressed in the format.  (Implementors are warned
  that no DER is defined for UTCT in X.509, thus transformation between



Kent                                                            [Page 8]

RFC 1422           Certificate-Based Key Management        February 1993


  local and transfer syntax must be performed carefully, e.g., when
  computing the hash value for a certificate.  For example, a UTCT
  value which includes explict, zero values for seconds would not
  produce the same hash value as one in which the seconds were
  omitted.) It also recommended that all times be expressed as
  Greenwich Mean Time (Zulu), to simplify comparisons and avoid
  confusion relating to daylight savings time.  Note that UTCT
  expresses the value of a year modulo 100 (with no indication of
  century), hence comparisons involving dates in different centuries
  must be performed with care.

  The longer the interval, the greater the likelihood that compromise
  of a private component or name change will render it invalid and thus
  require that the certificate be revoked.  Once revoked, the
  certificate must remain on the issuer's CRL (see Section 3.4.3.4)
  until the validity interval expires.  PCAs may impose restrictions on
  the maximum validity interval that may be elected by CAs operating in
  their certification domain (see Appendix B).

  3.3.7  Subject Public Key

  A certificate carries the public component of its associated subject,
  as well as an indication of the algorithm, and any algorithm
  parameters, with which the public component is to be used.  This
  algorithm identifier is independent of that which is specified in the
  signature field described above.  RFC 1423 specifies the algorithm
  identifiers which may be used in this context.

  3.4  Roles and Responsibilities

  One way to explain the architecture proposed by this document is to
  examine the roles which are defined for various entities in the
  architecture and to describe what is required of each entity in order
  for the proposed system to work properly.  The following sections
  identify four types of entities within this architecture: users and
  user agents, the Internet Policy Registration Authority, Policy
  Certification Authorities, and other Certification Authorities.  For
  each type of entity, this document specifies the procedures which the
  entity must execute as part of the architecture and the
  responsibilities the entity assumes as a function of its role in the
  architecture.

  3.4.1  Users and User Agents

  The term User Agent (UA) is taken from CCITT X.400 Message Handling
  Systems (MHS) Recommendations, which define it as follows: "In the
  context of message handling, the functional object, a component of
  MHS, by means of which a single direct user engages in message



Kent                                                            [Page 9]

RFC 1422           Certificate-Based Key Management        February 1993


  handling."   In the Internet environment, programs such as rand mh
  and Gnu emacs rmail are UAs.  UAs exchange messages by calling on a
  supporting Message Transfer Service (MTS), e.g., the SMTP mail relays
  used in the Internet.

  3.4.1.1  Generating and Protecting Component Pairs

  A UA process supporting PEM must protect the private component of its
  associated entity (e.g., a human user or a mailing list) from
  disclosure, though the means by which this is effected is a local
  matter.  It is essential that the user take all available precautions
  to protect his private component as the secrecy of this value is
  central to the security offered by PEM to that user.   For example,
  the private component might be stored in encrypted form, protected
  with a locally managed symmetric encryption key (e.g., using DES).
  The user would supply a password or passphrase which would be
  employed as a symmetric key to decrypt the private component when
  required for PEM processing (either on a per message or per session
  basis).  Alternatively, the private component might be stored on a
  diskette which would be inserted by the user whenever he originated
  or received PEM messages.  Explicit zeroing of memory locations where
  this component transiently resides could provide further protection.
  Other precautions, based on local operating system security
  facilities, also should be employed.

  It is recommended that each user employ ancillary software (not
  otherwise associated with normal UA operation) or hardware to
  generate his personal public-key component pair.  Software for
  generating user component pairs will be available as part of the
  reference implementation of PEM distributed freely in the U.S.
  portion of the Internet.  It is critically important that the
  component pair generation procedure be effected in as secure a
  fashion as possible, to ensure that the resulting private component
  is unpredictable.  Introduction of adequate randomness into the
  component pair generation procedure is potentially the most difficult
  aspect of this process and the user is advised to pay particular
  attention to this aspect.  (Component pairs employed in public-key
  cryptosystems tend to be large integers which must be "randomly"
  selected subject to mathematical constraints imposed by the
  cryptosystem.  Input(s) used to seed the component pair generation
  process must be as unpredictable as possible.  An example of a poor
  random number selection technique is one in which a pseudo-random
  number generator is seeded solely with the current date and time.  An
  attacker who could determine approximately when a component pair was
  generated could easily regenerate candidate component pairs and
  compare the public component to the user's public component to detect
  when the corresponding private component had been found.)




Kent                                                           [Page 10]

RFC 1422           Certificate-Based Key Management        February 1993


  There is no requirement imposed by this architecture that anyone
  other than the user, including any certification authority, have
  access to the user's private component.  Thus a user may retain his
  component pair even if his certificate changes, e.g., due to rollover
  in the validity interval or because of a change of certifying
  authority.  Even if a user is issued a certificate in the context of
  his employment, there is generally no requirement that the employer
  have access to the user's private component.  The rationale is that
  any messages signed by the user are verifiable using his public
  component.   In the event that the corresponding private component
  becomes unavailable, any ENCRYPTED messages directed to the user
  would be indecipherable and would require retransmission.

  Note that if the user stores messages in ENCRYPTED form, these
  messages also would become indecipherable in the event that the
  private component is lost or changed.  To minimize the potential for
  loss of data in such circumstances messages can be transformed into
  MIC-ONLY or MIC-CLEAR form if cryptographically-enforced
  confidentiality is not required for the messages stored within the
  user's computer.  Alternatively, these transformed messages might be
  forwarded in ENCRYPTED form to a (trivial) distribution list which
  serves in a backup capacity and for which the user's employer holds
  the private component.

  A user may possess multiple certificates which may embody the same or
  different public components.  For example, these certificates might
  represent  a current and a former organizational user identity and a
  residential user identity.  It is recommended that a PEM UA be
  capable of supporting a user who possess multiple certificates,
  irrespective of whether the certificates associated with the user
  contain the same or different DNs or public components.

  3.4.1.2  User Registration

  Most details of user registration are a local matter, subject to
  policies established by the user's CA and the PCA under which that CA
  has been certified.  In general a user must provide, at a minimum,
  his public component and distinguished name to a CA, or a
  representative thereof, for inclusion in the user's certificate.
  (The user also might provide a  complete certificate, minus the
  signature, as described in RFC 1424.)  The CA will employ some means,
  specified by the CA in accordance with the policy of its PCA, to
  validate the user's claimed identity and to ensure that the public
  component provided is associated with the user whose distinguished
  name is to be bound into the certificate.  (In the case of PERSONA
  certificates, described below, the procedure is a bit different.) The
  certifying authority generates a certificate containing the user's
  distinguished name and public component, the authority's



Kent                                                           [Page 11]

RFC 1422           Certificate-Based Key Management        February 1993


  distinguished name and other information (see Section 3.3) and signs
  the result using the private component of the authority.

  3.4.1.3  CRL Management

  Mechanisms for managing a UA certificate cache are, in typical
  standards parlance, a local matter.  However, proper maintenance of
  such a cache is critical to the correct, secure operation of a PEM UA
  and provides a basis for improved performance.  Moreover, use of a
  cache permits a PEM UA to operate in the absence of directories (and
  in circumstances where directories are inaccessible).  The following
  discussion  provides a paradigm for one aspect of cache management,
  namely the processing of CRLs, the functional equivalent of which
  must be embodied in any PEM UA implementation compliant with this
  document.  The specifications for CRLs used with PEM are provided in
  Section 3.5.

  X.500 makes provision for the storage of CRLs as directory attributes
  associated with CA entries.  Thus, when X.500 directories become
  widely available, UAs can retrieve CRLs from directories as required.
  In the interim, the IPRA will coordinate with PCAs to provide a
  robust database facility which will contain CRLs issued by the IPRA,
  by PCAs, and by all CAs.  Access to this database will be provided
  through mailboxes maintained by each PCA.  Every PEM UA must provide
  a facility for requesting CRLs from this database using the
  mechanisms defined in RFC 1424.  Thus the UA must include a
  configuration parameter which specifies one or more mailbox addresses
  from which CRLs may be retrieved.  Access to the CRL database may be
  automated, e.g., as part of the certificate validation process (see
  Section 3.6) or may be user directed.  Responses to CRL requests will
  employ the PEM header format specified in RFC 1421 for CRL
  propagation.  As noted in RFC 1421, every PEM UA must be capable of
  processing CRLs distributed via such messages.  This message format
  also may be employed to support a "push" (versus a "pull") model of
  CRL distribution, i.e., to support unsolicited distribution of CRLs.

  CRLs received by a PEM UA must be validated (A CRL is validated in
  much the same manner as a certificate, i.e., the CIC (see RFC 1113)
  is calculated and compared against the decrypted signature value
  obtained from the CRL.  See Section 3.6 for additional details
  related to validation of certificates.) prior to being processed
  against any cached certificate information.  Any cache entries which
  match CRL entries should be marked as revoked, but it is not
  necessary to delete cache entries marked as revoked nor to delete
  subordinate entries.  In processing a CRL against the cache it is
  important to recall that certificate serial numbers are unique only
  for each issuer and that multiple, distinct CRLs may be issued under
  the same CA DN (signed using different private components), so care



Kent                                                           [Page 12]

RFC 1422           Certificate-Based Key Management        February 1993


  must be exercised in effecting this cache search.  (This situation
  may arise either because an organizational CA is certified by
  multiple PCAs, or because multiple residential CAs are certified
  under different PCAs.)

  This procedure applies to cache entries associated with PCAs and CAs,
  as well as user entries.  The UA also must retain each CRL to screen
  incoming messages to detect use of revoked certificates carried in
  PEM message headers.  Thus a UA must be capable of processing and
  retaining CRLs issued by the IPRA (which will list revoked PCA
  certificates), by any PCA (which will list revoked CA certificate
  issued by that PCA), and by any CA (which will list revoked user or
  subordinate CA certificates issued by that CA).

  3.4.1.4  Facilitating Interoperation

  In the absence of ubiquitous directory services or knowledge
  (acquired through out-of-band means) that a recipient already
  possesses the necessary issuer certificates, it is recommended that
  an originating (PEM) UA include sufficient certificates to permit
  validation of the user's public key.  To this end every PEM UA must
  be capable of including a full (originator) certification path, i.e.,
  including the user's certificate (using the "Originator-Certificate"
  field) and every superior (CA/PCA) certificate (using "Issuer-
  Certificate" fields) back to the IPRA, in a PEM message.  A PEM UA
  may send less than a full certification path, e.g., based on analysis
  of a recipient list, but a UA which provides this sort of
  optimization must also provide the user with a capability to force
  transmission of a full certification path.

  Optimization for the transmitted originator certification path may be
  effected by a UA as a side effect of the processing performed during
  message submission.  When an originator submits an ENCRYPTED message
  (as per RFC 1421, his UA must validate the certificates of the
  recipients (see Section 3.6).  In the course of performing this
  validation the UA can determine the minimum set of certificates which
  must be included to ensure that all recipients can process the
  received message.  Submission of a MIC-ONLY or MIC-CLEAR message (as
  per RFC 1421) does not entail validation of recipient certificates
  and thus it may not be possible for the originator's UA to determine
  the minimum certificate set as above.

  3.4.2  The Internet Policy Registration Authority (IPRA)

  The IPRA acts as the root of the certification hierarchy for the
  Internet community.  The public component of the IPRA forms the
  foundation for all certificate validation within this hierarchy.  The
  IPRA will be operated under the auspices of the Internet Society, an



Kent                                                           [Page 13]

RFC 1422           Certificate-Based Key Management        February 1993


  international, non-profit organization.  The IPRA certifies all PCAs,
  ensuring that they agree to abide by the Internet-wide policy
  established by the IPRA.  This policy, and the services provided by
  the IPRA, are detailed below.

  3.4.2.1  PCA Registration

  The IPRA certifies only PCAs, not CAs or users.  Each PCA must file
  with the IPRA a description of its proposed policy.  This document
  will be published as an informational RFC.  A copy of the document,
  signed by the IPRA (in the form of a PEM MIC-ONLY message) will be
  made available via electronic mail access by the IPRA.  This
  convention is adopted so that every Internet user has a reference
  point for determining the policies associated with the issuance of
  any certificate which he may encounter.  The existence of a digitally
  signed copy of the document ensures the immutability of the document.
  Authorization of a PCA to operate in the Internet hierarchy is
  signified by the publication of the policy document, and the issuance
  of a certificate to the PCA, signed by the IPRA.  An outline for PCA
  policy statements is contained in Section 3.4.3 of this document.

  As part of registration, each PCA will be required to execute a legal
  agreement with the IPRA, and to pay a fee to defray the costs of
  operating the IPRA.  Each a PCA must specify its distinguished name.
  The IPRA will take reasonable precautions to ensure that the
  distinguished name claimed by a PCA is legitimate, e.g., requiring
  the PCA to provide documentation supporting its claim to a DN.
  However, the certification of a PCA by the IPRA does not constitute a
  endorsement of the PCA's claim to this DN outside of the context of
  this certification system.

  3.4.2.2  Ensuring the Uniqueness of Distinguished Names

  A fundamental requirement of this certification scheme is that
  certificates are not issued to distinct entities under the same
  distinguished name.  This requirement is important to the success of
  distributed management for the certification hierarchy.  The IPRA
  will not certify two PCAs with the same distinguished name and no PCA
  may certify two CAs with the same DN.  However, since PCAs are
  expected to certify organizational CAs in widely disjoint portions of
  the directory namespace, and since X.500 directories are not
  ubiquitous, a facility is required for coordination among PCAs to
  ensure the uniqueness of CA DNs.  (This architecture allows multiple
  PCAs to certify residential CAs and thus multiple, distinct
  residential CAs with identical DNs may come into existence, at least
  until such time as civil authorities assume responsibilities for such
  certification.  Thus, on an interim basis, the architecture
  explicitly accommodates the potential for duplicate residential CA



Kent                                                           [Page 14]

RFC 1422           Certificate-Based Key Management        February 1993


  DNs.)

  In support of the uniqueness requirement, the IPRA will establish and
  maintain a database to detect potential, unintended duplicate
  certification of CA distinguished names.  This database will be made
  accessible to all PCAs via an email interface.  Each entry in this
  database will consist of a 4-tuple.  The first element in each entry
  is a hash value, computed on a canonical, ASN.1 encoded
  representation of a CA distinguished name.  The second element
  contains the subjectPublicKey that appears in the CA's certificate.
  The third element is the distinguished name of the PCA which
  registered the entry.  The fourth element consists of the date and
  time at which the entry was made, as established by the IPRA.  This
  database structure provides a degree of privacy for CAs registered by
  PCAs, while providing a facility for ensuring global uniqueness of CA
  DNs certified in this scheme.

  In order to avoid conflicts, a PCA should query the database using a
  CA DN hash value as a search key, prior to certifying a CA.  The
  database will return any entries which match the query, i.e., which
  have the same CA DN.  The PCA can use the information contained in
  any returned entries to determine if any PCAs should be contacted to
  resolve possible DN conflicts.  If no potential conflicts appear, a
  PCA can then submit a candidate entry, consisting of the first three
  element values, plus any entries returned by the query.  The database
  will register this entry, supplying the time and date stamp, only if
  two conditions are met: (1) the first two elements (the CA DN hash
  and the CA subjectPublicKey) of the candidate entry together must be
  unique and, (2) any other entries included in the submission must
  match what the current database would return if the query
  corresponding to the candidate entry were submitted.

  If the database detects a conflicting entry (failure of case 1
  above), or if the submission indicates that the PCA's perception of
  possible conflicting entries is not current (failure of case 2), the
  submission is rejected and the database will return the potential
  conflicting entry (entries).  If the submission is successful, the
  database will return the timestamped new entry.  The database does
  not, in itself, guarantee uniqueness of CA DNs as it allows for two
  DNs associated with different public components to be registered.
  Rather, it is the responsibility of PCAs to coordinate with one
  another whenever the database indicates a potential DN conflict and
  to resolve such conflicts prior to certification of CAs.  Details of
  the protocol used to access the database will be provided in another
  document.

  As noted earlier, a CA may be certified under more than one PCA,
  e.g., because the CA wants to issue certificates under two different



Kent                                                           [Page 15]

RFC 1422           Certificate-Based Key Management        February 1993


  policies.  If a CA is certified by multiple different PCAs, the CA
  must employ a different public key pair for each PCA.  In such
  circumstances the certificate issued to the CA by each PCA will
  contain a different subjectPublicKey and thus will represent a
  different entry in this database.  The same situation may arise if
  multiple, equivalent residential CAs are certified by different PCAs.

  To complete the strategy for ensuring uniqueness of DNs, there is a
  DN subordination requirement levied on CAs.  In general, CAs are
  expected to sign certificates only if the subject DN in the
  certificate is subordinate to the issuer (CA) DN.  This ensures that
  certificates issued by a CA are syntactically constrained to refer to
  subordinate entities in the X.500 directory information tree (DIT),
  and this further limits the possibility of duplicate DN registration.
  CAs may sign certificates which do not comply with this requirement
  if the certificates are "cross-certificates" or "reverse
  certificates" (see X.509) used with applications other than PEM.

  The IPRA also will establish and maintain a separate database to
  detect potential duplicate certification of (residential) user
  distinguished names.  Each entry in this database will consist of 4-
  tuple as above, but the first components is the hash of a residential
  user DN and the third component is the DN of the residential CA DN
  which registered the user.  This structure provides a degree of
  privacy for users registered by CAs which service residential users
  while providing a facility for ensuring global uniqueness of user DNs
  certified under this scheme.  The same database access facilities are
  provided as described above for the CA database.  Here it is the
  responsibility of the CAs to coordinate whenever the database
  indicates a potential conflict and to resolve the conflict prior to
  (residential) user certification.

  3.4.2.3  Accuracy of Distinguished Names

  As noted above, the IPRA will make a reasonable effort to ensure that
  PCA DNs are accurate.  The procedures employed to ensure the accuracy
  of a CA distinguished name, i.e., the confidence attached to the
  DN/public component binding implied by a certificate, will vary
  according to PCA policy.  However, it is expected that every PCA will
  make a good faith effort to ensure the legitimacy of each CA DN
  certified by the PCA.  Part of this effort should include a check
  that the purported CA DN is consistent with any applicable national
  standards for DN assignment, e.g., NADF recommendations within North
  America [5,9].







Kent                                                           [Page 16]

RFC 1422           Certificate-Based Key Management        February 1993


  3.4.2.4  Distinguished Name Conventions

  A few basic DN conventions are included in the IPRA policy.  The IPRA
  will certify PCAs, but not CAs nor users.  PCAs will certify CAs, but
  not users.  These conventions are required to allow simple
  certificate validation within PEM, as described later.  Certificates
  issued by CAs (for use with PEM) will be for users or for other CAs,
  either of which must have DNs subordinate to that of the issuing CA.

  The attributes employed in constructing DNs will be specified in a
  list maintained by the IANA, to provide a coordinated basis for
  attribute identification for all applications employing DNs.  This
  list will initially be populated with attributes taken from X.520.
  This document does not impose detailed restrictions on the attributes
  used to identify different entities to which certificates are issued,
  but PCAs may impose such restrictions as part of their policies.
  PCAs, CAs and users are urged to employ only those DN attributes
  which have printable representations, to facilitate display and
  entry.

  3.4.2.5  CRL Management

  Among the procedures articulated by each PCA in its policy statement
  are procedures for the maintenance and distribution of CRLs by the
  PCA itself and by its subordinate CAs.  The frequency of issue of
  CRLs may vary according to PCA-specific policy, but every PCA and CA
  must issue a CRL upon inception to provide a basis for uniform
  certificate validation procedures throughout the Internet hierarchy.
  The IPRA will maintain a CRL for all the PCAs it certifies and this
  CRL will be updated monthly.  Each PCA will maintain a CRL for all of
  the CAs which it certifies and these CRLs will be updated in
  accordance with each PCA's policy.   The format for these CRLs is
  that specified in Section 3.5.2 of the document.

  In the absence of ubiquitous X.500 directory services, the IPRA will
  require each PCA to provide, for its users, robust database access to
  CRLs for the Internet hierarchy, i.e., the IPRA CRL, PCA CRLs, and
  CRLs from all CAs.  The means by which this database is implemented
  is to be coordinated between the IPRA and PCAs.  This database will
  be accessible via email as specified in RFC 1424, both for retrieval
  of (current) CRLs by any user, and for submission of new CRLs by CAs,
  PCAs and the IPRA.  Individual PCAs also may elect to maintain CRL
  archives for their CAs, but this is not required by this policy.

  3.4.2.6  Public Key Algorithm Licensing Issues

  This certification hierarchy is architecturally independent of any
  specific digital signature (public key) algorithm.  Some algorithms,



Kent                                                           [Page 17]

RFC 1422           Certificate-Based Key Management        February 1993


  employed for signing certificates and validating certificate
  signatures, are patented in some countries.  The IPRA will not grant
  a license to any PCA for the use of any signature algorithm in
  conjunction with the management of this certification hierarchy.  The
  IPRA will acquire, for itself, any licenses needed for it to sign
  certificates and CRLs for PCAs, for all algorithms which the IPRA
  supports.  Every PCA will be required to represent to the IPRA that
  the PCA has obtained any licenses required to issue (sign)
  certificates and CRLs in the environment(s) which the PCA will serve.

  For example, the RSA cryptosystem is patented in the United States
  and thus any PCA operating in the U.S. and using RSA to sign
  certificates and CRLs must represent that it has a valid license to
  employ the RSA algorithm in this fashion.  In contrast, a PCA
  employing RSA and operating outside of the U.S. would represent that
  it is exempt from these licensing constraints.

  3.4.3  Policy Certification Authorities

  The policy statement submitted by a prospective PCA must address the
  topics in the following outline.  Additional policy information may
  be contained in the statement, but PCAs are requested not to use
  these statements as advertising vehicles.

  1. PCA Identity-  The DN of the PCA must be specified.  A postal
  address, an Internet mail address, and telephone (and optional fax)
  numbers must be provided for (human) contact with the PCA.  The date
  on which this statement is effective, and its scheduled duration must
  be specified.

  2. PCA Scope- Each PCA must describe the community which the PCA
  plans to serve.  A PCA should indicate if it will certify
  organizational, residential, and/or PERSONA CAs.   There is not a
  requirement that a single PCA serve only one type of CA, but if a PCA
  serves multiple types of CAs, the policy statement must specify
  clearly how a user can distinguish among these classes.  If the PCA
  will operate CAs to directly serve residential or PERSONA users, it
  must so state.

  3. PCA Security & Privacy- Each PCA must specify the technical and
  procedural security measures it will employ in the generation and
  protection of its component pair.  If any security requirements are
  imposed on CAs certified by the PCA these must be specified as well.
  A PCA also must specify what measures it will take to protect the
  privacy of any information collected in the course of certifying CAs.
  If the PCA operates one or more CAs directly, to serve residential or
  PERSONA users, then this statement on privacy measures applies to
  these CAs as well.



Kent                                                           [Page 18]

RFC 1422           Certificate-Based Key Management        February 1993


  4. Certification Policy-  Each PCA must specify the policy and
  procedures which govern its certification of CAs and how this policy
  applies transitively to entities (users or subordinate CAs) certified
  by these CAs.  For example, a PCA must state what procedure is
  employed to verify the claimed identity of a CA, and the CA's right
  to use a DN.  Similarly, if any requirements are imposed on CAs to
  validate the identity of users, these requirements must be specified.
  Since all PCAs are required to cooperate in the resolution of
  potential DN conflicts, each PCA is required to specify the procedure
  it will employ to resolve such conflicts.  If the PCA imposes a
  maximum validity interval for the CA certificates it issues, and/or
  for user (or subordinate CA) certificates issued by the CAs it
  certifies, then these restrictions must be specified.

  5. CRL Management-  Each PCA must specify the frequency with which it
  will issue scheduled CRLs.  It also must specify any constraints it
  imposes on the frequency of scheduled issue of CRLs by the CAs it
  certifies, and by subordinate CAs.  Both maximum and minimum
  constraints should be specified.  Since the IPRA policy calls for
  each CRL issued by a CA to be forwarded to the cognizant PCA, each
  PCA must specify a mailbox address to which CRLs are to be
  transmitted.  The PCA also must specify a mailbox address for CRL
  queries.  If the PCA offers any additional CRL management services,
  e.g., archiving of old CRLs, then procedures for invoking these
  services must be specified.  If the PCA requires CAs to provide any
  additional CRL management services, such services must be specified
  here.

  6. Naming Conventions- If the PCA imposes any conventions on DNs used
  by the CAs it certifies, or by entities certified by these CAs, these
  conventions must be specified.  If any semantics are associated with
  such conventions, these semantics must be specified.

  7. Business Issues- If a legal agreement must be executed between a
  PCA and the CAs it certifies, reference to that agreement must be
  noted, but the agreement itself ought not be a part of the policy
  statement.  Similarly, if any fees are charged by the PCA this should
  be noted, but the fee structure per se ought not be part of this
  policy statement.

  8. Other- Any other topics the PCA deems relevant to a statement of
  its policy can be included.  However, the PCA should be aware that a
  policy statement is considered to be an immutable, long lived
  document and thus considerable care should be exercised in deciding
  what material is to be included in the statement.






Kent                                                           [Page 19]

RFC 1422           Certificate-Based Key Management        February 1993


  3.4.4  Certification Authorities

  In X.509 the term "certification authority" is defined as "an
  authority trusted by one or more users to create and assign
  certificates".  X.509 imposes few constraints on CAs, but practical
  implementation of a worldwide certification system requires
  establishment of technical and procedural conventions by which all
  CAs are expected to abide.  Such conventions are established
  throughout this document.  All CAs are required to maintain a
  database of the DNs which they have certified and to take measures to
  ensure that they do not certify duplicate DNs, either for users or
  for subordinate CAs.

  It is critical that the private component of a CA be afforded a high
  level of security, otherwise the authenticity guarantee implied by
  certificates signed by the CA is voided.  Some PCAs may impose
  stringent requirements on CAs within their purview to ensure that a
  high level of security is afforded the certificate signing process,
  but not all PCAs are expected to impose such constraints.

  3.4.4.1  Organizational CAs

  Many of the CAs certified by PCAs are expected to represent
  organizations.  A wide range of organizations are encompassed by this
  model: commercial, governmental, educational, non-profit,
  professional societies, etc.  The common thread is that the entities
  certified by these CAs have some form of affiliation with the
  organization.  The object classes for organizations, organizational
  units, organizational persons, organizational roles, etc., as defined
  in X.521, form the models for entities certified by such CAs.  The
  affiliation implied by organizational certification motivates the DN
  subordination requirement cited in Section 3.4.2.4.

  As an example, an organizational user certificate might contain a
  subject DN of the form: C = "US" SP = "Massachusetts" L = "Cambridge"
  O = "Bolt Beranek and Newman" OU = "Communications Division" CN =
  "Steve Kent".  The issuer of this certificate might have a DN of the
  form: C = "US" SP = "Massachusetts" L = "Cambridge" O= "Bolt Beranek
  and Newman".  Note that the organizational unit attribute is omitted
  from the issuer DN, implying that there is no CA dedicated to the
  "Communications Division".

  3.4.4.2  Residential CAs

  Users may wish to obtain certificates which do not imply any
  organizational affiliation but which do purport to accurately and
  uniquely identify them.  Such users can be registered as residential
  persons and the DN of such a user should be consistent with the



Kent                                                           [Page 20]

RFC 1422           Certificate-Based Key Management        February 1993


  attributes of the corresponding X.521 object class.  Over time we
  anticipate that such users will be accommodated by civil government
  entities who will assume electronic certification responsibility at
  geographically designated points in the naming hierarchy.  Until
  civil authorities are prepared to issue certificates of this form,
  residential user CAs will accommodate such users.

  Because residential CAs may be operated under the auspices of
  multiple PCAs, there is a potential for the same residential CA DN to
  be assumed by several distinct entities.  This represents the one
  exception to the rule articulated throughout this document that no
  two entities may have the same DN.  This conflict is tolerated so as
  to allow residential CAs to be established offering different
  policies.  Two requirements are levied upon residential CAs as a
  result: (1) residential CAs must employ the residential DN conflict
  detection database maintained by the IPRA, and (2) residential CAs
  must coordinate to ensure that they do not assign duplicate
  certificate serial numbers.

  As an example, a residential user certificate might include a subject
  name of the form: C = "US" SP = "Massachusetts" L = "Boston" PA = "19
  North Square" CN = "Paul Revere."  The issuer of that certificate
  might have a DN of the form: C = "US"  SP = "Massachusetts" L =
  "Boston".  Note that the issuer DN is superior to the subject DN, as
  required by the IPRA policy described earlier.

  3.4.4.3  PERSONA CAs

  One or more CAs will be established to accommodate users who wish to
  conceal their identities while making use of PEM security features,
  e.g., to preserve the anonymity offered by "arbitrary" mailbox names
  in the current mail environment.  In this case the certifying
  authority is explicitly NOT vouching for the identity of the user.
  All such certificates are issued under a PERSONA CA, subordinate to a
  PCA with a PERSONA policy, to warn users explicitly that the subject
  DN is NOT a validated user identity.  To minimize the possibility of
  syntactic confusion with certificates which do purport to specify an
  authenticated user identity, a PERSONA certificate is issued as a
  form of organizational user certificate, not a residential user
  certificate.  There are no explicit, reserved words used to identify
  PERSONA user certificates.

  A CA issuing PERSONA certificates must institute procedures to ensure
  that it does not issue the same subject DN to multiple users (a
  constraint required for all certificates of any type issued by any
  CA).  There are no requirements on an issuer of PERSONA certificates
  to maintain any other records that might bind the true identity of
  the subject to his certificate.  However, a CA issuing such



Kent                                                           [Page 21]

RFC 1422           Certificate-Based Key Management        February 1993


  certificates must establish procedures (not specified in this
  document) in order to allow the holder of a PERSONA certificate to
  request that his certificate be revoked (i.e., listed on a CRL).

  As an example, a PERSONA user certificate might include a subject DN
  of the form:  C = "US" SP = "Massachusetts" L = "Boston" O =
  "Pseudonyms R US" CN = "Paul Revere."  The issuer of this certificate
  might have a DN of the form: C = "US"  SP = "Massachusetts" L =
  "Boston" O = "Pseudonyms R US".  Note the differences between this
  PERSONA user certificate for "Paul Revere" and the corresponding
  residential user certificate for the same common name.

  3.4.4.4  CA Responsibilities for CRL Management

  As X.500 directory servers become available, CRLs should be
  maintained and accessed via these servers.  However, prior to
  widespread deployment of X.500 directories, this document adopts some
  additional requirements for CRL management by CAs and PCAs.  As per
  X.509, each CA is required to maintain a CRL (in the format specified
  by this document in Appendix A) which contains entries for all
  certificates issued and later revoked by the CA.  Once a certificate
  is entered on a CRL it remains there until the validity interval
  expires.  Each PCA is required to maintain a CRL for revoked CA
  certificates within its domain.  The interval at which a CA issues a
  CRL is not fixed by this document, but the PCAs may establish minimum
  and maximum intervals for such issuance.

  As noted earlier, each PCA will provide access to a database
  containing CRLs issued by the IPRA, PCAs, and all CAs.  In support of
  this requirement, each CA must supply its current CRL to its PCA in a
  fashion consistent with CRL issuance rules imposed by the PCA and
  with the next scheduled issue date specified by the CA (see Section
  3.5.1).  CAs may distribute CRLs to subordinate UAs using the CRL
  processing type available in PEM messages (see RFC 1421).  CAs also
  may provide access to CRLs via the database mechanism described in
  RFC 1424 and alluded to immediately above.

  3.5  Certificate Revocation

  3.5.1  X.509 CRLs

  X.509 states that it is a CA's responsibility to maintain: "a time-
  stamped list of the certificates it issued which have been revoked."
  There are two primary reasons for a CA to revoke a certificate, i.e.,
  suspected compromise of a private component (invalidating the
  corresponding public component) or change of user affiliation
  (invalidating the DN).  The use of Certificate Revocation Lists
  (CRLs) as defined in X.509 is one means of propagating information



Kent                                                           [Page 22]

RFC 1422           Certificate-Based Key Management        February 1993


  relative to certificate revocation, though it is not a perfect
  mechanism.  In particular, an X.509 CRL indicates only the age of the
  information contained in it; it does not provide any basis for
  determining if the list is the most current CRL available from a
  given CA.

  The proposed architecture establishes a format for a CRL in which not
  only the date of issue, but also the next scheduled date of issue is
  specified.  Adopting this convention, when the next scheduled issue
  date arrives a CA (Throughout this section, when the term "CA" is
  employed, it should be interpreted broadly, to include the IPRA and
  PCAs as well as organizational, residential, and PERSONA CAs.) will
  issue a new CRL, even if there are no changes in the list of entries.
  In this fashion each CA can independently establish and advertise the
  frequency with which CRLs are issued by that CA.  Note that this does
  not preclude CRL issuance on a more frequent basis, e.g., in case of
  some emergency, but no system-wide mechanisms are architected for
  alerting users that such an unscheduled issuance has taken place.
  This scheduled CRL issuance convention allows users (UAs) to
  determine whether a given CRL is "out of date," a facility not
  available from the (1988) X.509 CRL format.

  The description of CRL management in the text and the format for CRLs
  specified in X.509 (1988) are inconsistent.  For example, the latter
  associates an issuer distinguished name with each revoked certificate
  even though the text states that a CRL contains entries for only a
  single issuer (which is separately specified in the CRL format).  The
  CRL format adopted for PEM is a (simplified) format consistent with
  the text of X.509, but not identical to the accompanying format. The
  ASN.1 format for CRLs used with PEM is provided in Appendix A.

  X.509 also defines a syntax for the "time-stamped list of revoked
  certificates representing other CAs."  This syntax, the
  "AuthorityRevocationList" (ARL) allows the list to include references
  to certificates issued by CAs other than the list maintainer.  There
  is no syntactic difference between these two lists except as they are
  stored in directories.  Since PEM is expected to be used prior to
  widespread directory deployment, this distinction between ARLs and
  CRLs is not syntactically significant.  As a simplification, this
  document specifies the use the CRL format defined below for
  revocation both of user and of CA certificates.

  3.5.2  PEM CRL Format

  Appendix A contains the ASN.1 description of CRLs specified by this
  document.  This section provides an informal description of CRL
  components analogous to that provided for certificates in Section
  3.3.



Kent                                                           [Page 23]

RFC 1422           Certificate-Based Key Management        February 1993


      1. signature (signature algorithm ID and parameters)

      2. issuer

      3. last update

      4. next update

      5. revoked certificates

  The "signature" is a data item completely analogous to the signature
  data item in a certificate. Similarly, the "issuer" is the DN of the
  CA which signed the CRL.  The "last update" and "next update" fields
  contain time and date values (UTCT format) which specify,
  respectively, when this CRL was issued and when the next CRL is
  scheduled to be issued.  Finally, "revoked certificates" is a
  sequence of ordered pairs, in which the first element is the serial
  number of the revoked certificate and the second element is the time
  and date of the revocation for that certificate.

  The semantics for this second element are not made clear in X.509.
  For example, the time and date specified might indicate when a
  private component was thought to have been compromised or it may
  reflect when the report of such compromise was reported to the CA.

  For uniformity, this document adopts the latter convention, i.e., the
  revocation date specifies the time and date at which a CA formally
  acknowledges a report of a compromise or a change or DN attributes.
  As with certificates, it is recommended that the UTCT values be of no
  finer granularity than minutes and that all values be stated in terms
  of Zulu.

  3.6  Certificate Validation

  3.6.1  Validation Basics

  Every UA must contain the public component of the IPRA as the root
  for its certificate validation database.  Public components
  associated with PCAs must be identified as such, so that the
  certificate validation process described below can operate correctly.
  Whenever a certificate for a PCA is entered into a UA cache, e.g., if
  encountered in a PEM message encapsulated header, the certificate
  must NOT be entered into the cache automatically.  Rather, the user
  must be notified and must explicitly direct the UA to enter any PCA
  certificate data into the cache.  This precaution is essential
  because introduction of a PCA certificate into the cache implies user
  recognition of the policy associated with the PCA.




Kent                                                           [Page 24]

RFC 1422           Certificate-Based Key Management        February 1993


  Validating a certificate begins with verifying that the signature
  affixed to the certificate is valid, i.e., that the hash value
  computed on the certificate contents matches the value that results
  from decrypting the signature field using the public component of the
  issuer.  In order to perform this operation the user must possess the
  public component of the issuer, either via some integrity-assured
  channel, or by extracting it from another (validated) certificate.
  In order to rapidly terminate this recursive validation process, we
  recommend each PCA sign certificates for all CAs within its domain,
  even CAs which are certified by other, superior CAs in the
  certification hierarchy.

  The public component needed to validate certificates signed by the
  IPRA is made available to each user as part of the registration or
  via the PEM installation process.  Thus a user will be able to
  validate any PCA certificate immediately.  CAs are certified by PCAs,
  so validation of a CA certificate requires processing a validation
  path of length two.  User certificates are issued by CAs (either
  immediately subordinate to PCAs or subordinate to other CAs), thus
  validation of a user certificate may require three or more steps.
  Local caching of validated certificates by a UA can be used to speed
  up this process significantly.

  Consider the situation in which a user receives a privacy enhanced
  message from an originator with whom the recipient has never
  previously corresponded, and assume that the message originator
  includes a full certification path in the PEM message header.  First
  the recipient can use the IPRA's public component to validate a PCA
  certificate contained in an Issuer-Certificate field.  Using the
  PCA's public component extracted from this certificate, the CA
  certificate in an Issuer-Certificate field also can be validated.
  This process cam be repeated until the certificate for the
  originator, from the Originator-Certificate field, is validated.

  Having performed this certificate validation process, the recipient
  can extract the originator's public component and use it to decrypt
  the content of the MIC-Info field.  By comparing the decrypted
  contents of this field against the MIC computed locally on the
  message the user verifies the data origin authenticity and integrity
  of the message.  It is recommended that implementations of privacy
  enhanced mail cache validated public components (acquired from
  incoming mail) to speed up this process.  If a message arrives from
  an originator whose public component is held in the recipient's cache
  (and if the cache is maintained in a fashion that ensures timely
  incorporation of received CRLs), the recipient can immediately employ
  that public component without the need for the certificate validation
  process described here. (For some digital signature algorithms, the
  processing required for certificate validation is considerably faster



Kent                                                           [Page 25]

RFC 1422           Certificate-Based Key Management        February 1993


  than that involved in signing a certificate.  Use of such algorithms
  serves to minimize the computational burden on UAs.)

  3.6.2  Display of Certificate Validation Data

  PEM provides authenticated identities for message recipients and
  originators expressed in the form of distinguished names.  Mail
  systems in which PEM is employed may employ identifiers other than
  DNs as the primary means of identifying recipients or originators.
  Thus, in order to benefit from these authentication facilities, each
  PEM implementation must employ some means of binding native mail
  system identifiers to distinguished names in a fashion which does not
  undermine this basic PEM functionality.

  For example, if a human user interacts directly with PEM, then the
  full DN of the originator of any message received using PEM should be
  displayed for the user.  Merely displaying the PEM-protected message
  content, containing an originator name from the native mail system,
  does not provide equivalent security functionality and could allow
  spoofing.  If the recipient of a message is a forwarding agent such
  as a list exploder or mail relay, display of the originator's DN is
  not a relevant requirement.  In all cases the essential requirement
  is that the ultimate recipient of a PEM message be able to ascertain
  the identity of the originator based on the PEM certification system,
  not on unauthenticated identification information, e.g., extracted
  from the native message system.

  Conversely, for the originator of an ENCRYPTED message, it is
  important that recipient identities be linked to the DNs as expressed
  in PEM certificates.  This can be effected in a variety of ways by
  the PEM implementation, e.g., by display of recipient DNs upon
  message submission or by a tightly controlled binding between local
  aliases and the DNs.  Here too, if the originator is a forwarding
  process this linkage might be effected via various mechanisms not
  applicable to direct human interaction.  Again, the essential
  requirement is to avoid procedures which might undermine the
  authentication services provided by PEM.

  As described above, it is a local matter how and what certification
  information is displayed for a human user in the course of submission
  or delivery of a PEM message.  Nonetheless all PEM implementations
  must provide a user with the ability to display a full certification
  path for any certificate employed in PEM upon demand.  Implementors
  are urged to not overwhelm the user with certification path
  information which might confuse him or distract him from the critical
  information cited above.





Kent                                                           [Page 26]

RFC 1422           Certificate-Based Key Management        February 1993


  3.6.3  Validation Procedure Details

  Every PEM implementation is required to perform the following
  validation steps for every public component employed in the
  submission of an ENCRYPTED PEM message or the delivery of an
  ENCRYPTED, MIC-ONLY, or MIC-CLEAR PEM message.  Each public component
  may be acquired from an internal source, e.g., from a (secure) cache
  at the originator/recipient or it may be obtained from an external
  source, e.g., the PEM header of an incoming message or a directory.
  The following procedures applies to the validation of certificates
  from either type of source.

  Validation of a public component involves constructing a
  certification path between the component and the public component of
  the IPRA.  The validity interval for every certificate in this path
  must be checked.  PEM software must, at a minimum, warn the user if
  any certificate in the path fails the validity interval check, though
  the form of this warning is a local matter.  For example, the warning
  might indicate which certificate in the path had expired.  Local
  security policy may prohibit use of expired certificates.

  Each certificate also must be checked against the current CRL from
  the certificate's issuer to ensure that revoked certificates are not
  employed.  If the UA does not have access to the current CRL for any
  certificate in the path, the user must be warned.  Again, the form of
  the warning is a local matter.  For example, the warning might
  indicate whether the CRL is unavailable or, if available but not
  current, the CRL issue date should be displayed. Local policy may
  prohibit use of a public component which cannot be checked against a
  current CRL, and in such cases the user should receive the same
  information provided by the warning indications described above.

  If any revoked certificates are encountered in the construction of a
  certification path, the user must be warned.  The form of the warning
  is a local matter, but it is recommended that this warning be more
  stringent than those previously alluded to above.  For example, this
  warning might display the issuer and subject DNs from the revoked
  certificate and the date of revocation, and then require the user to
  provide a positive response before the submission or delivery process
  may proceed.  In the case of message submission, the warning might
  display the identity of the recipient affected by this validation
  failure and the user might be provided with the option to specify
  that this recipient be dropped from recipient list processing without
  affecting PEM processing for the remaining recipients.  Local policy
  may prohibit PEM processing if a revoked certificate is encountered
  in the course of constructing a certification path.

  Note that in order to comply with these validation procedures, a



Kent                                                           [Page 27]

RFC 1422           Certificate-Based Key Management        February 1993


  certificate cache must maintain all of the information contained in a
  certificate, not just the DNs and the public component.  For example
  the serial number and validity interval must be associated with the
  cache entry to comply with the checks described above.  Also note
  that these procedures apply to human interaction in message
  submission and delivery and are not directly applicable to forwarding
  processes.  When non human interaction is involved, a compliant PEM
  implementation must provide parameters to enable a process to specify
  whether certificate validation will succeed or fail if any of the
  conditions arise which would result in warnings to a human user.

  Finally, in the course of validating certificates as described above,
  one additional check must be performed: the subject DN of every
  certificate must be subordinate to the certificate issuer DN, except
  if the issuer is the IPRA or a PCA (hence another reason to
  distinguish the IPRA and PCA entries in a certificate cache).  This
  requirement is levied upon all PEM implementations as part of
  maintaining the certification hierarchy constraints defined in this
  document.  Any certificate which does not comply with these
  requirements is considered invalid and must be rejected in PEM
  submission or delivery processing.  The user  must be notified of the
  nature of this fatal error.





























Kent                                                           [Page 28]

RFC 1422           Certificate-Based Key Management        February 1993


A.  Appendix A: ASN.1 Syntax for Certificates and CRLs

A.1  Certificate Syntax

  The X.509 certificate format is defined by the following ASN.1
  syntax:

  Certificate ::= SIGNED SEQUENCE{
          version [0]     Version DEFAULT v1988,
          serialNumber    CertificateSerialNumber,
          signature       AlgorithmIdentifier,
          issuer          Name,
          validity        Validity,
          subject         Name,
          subjectPublicKeyInfo    SubjectPublicKeyInfo}

  Version ::=     INTEGER {v1988(0)}

  CertificateSerialNumber ::=     INTEGER

  Validity ::=    SEQUENCE{
          notBefore       UTCTime,
          notAfter        UTCTime}

  SubjectPublicKeyInfo ::=        SEQUENCE{
          algorithm               AlgorithmIdentifier,
          subjectPublicKey        BIT STRING}


  AlgorithmIdentifier ::= SEQUENCE{
          algorithm       OBJECT IDENTIFIER,
          parameters      ANY DEFINED BY algorithm OPTIONAL}

  The components of this structure are defined by ASN.1 syntax defined
  in the X.500 Series Recommendations.  RFC 1423 provides references
  for and the values of AlgorithmIdentifiers used by PEM in the
  subjectPublicKeyInfo and the signature data items.  It also describes
  how a signature is generated and the results represented.  Because
  the certificate is a signed data object, the distinguished encoding
  rules (see X.509, section 8.7) must be applied prior to signing.











Kent                                                           [Page 29]

RFC 1422           Certificate-Based Key Management        February 1993


A.2  Certificate Revocation List Syntax

  The following ASN.1 syntax, derived from X.509 and aligned with the
  suggested format in recently submitted defect reports, defines the
  format of CRLs for use in the PEM environment.

  CertificateRevocationList ::= SIGNED SEQUENCE{
          signature       AlgorithmIdentifier,
          issuer          Name,
          lastUpdate      UTCTime,
          nextUpdate      UTCTime,
          revokedCertificates
                          SEQUENCE OF CRLEntry OPTIONAL}

  CRLEntry ::= SEQUENCE{
          userCertificate SerialNumber,
          revocationDate UTCTime}

References

  [1] CCITT Recommendation X.411 (1988), "Message Handling Systems:
      Message Transfer System: Abstract Service Definition and
      Procedures".

  [2] CCITT Recommendation X.509 (1988), "The Directory -
      Authentication Framework".

  [3] CCITT Recommendation X.520 (1988), "The Directory - Selected
      Attribute Types".

  [4] NIST Special Publication 500-183, "Stable Agreements for Open
      Systems Interconnection Protocols," Version 4, Edition 1,
      December 1990.

  [5] North American Directory Forum, "A Naming Scheme for c=US", RFC
      1255, NADF, September 1991.

  [6] Linn, J., "Privacy Enhancement for Internet Electronic Mail: Part
      I: Message Encryption and Authentication Procedures", RFC 1421,
      DEC, February 1993.

  [7] Balenson, D., "Privacy Enhancement for Internet Electronic Mail:
      Part III: Algorithms, Modes, and Identifiers", RFC 1423, TIS,
      February 1993.

  [8] Balaski, B., "Privacy Enhancement for Internet Electronic Mail:
      Part IV: Notary, Co-Issuer, CRL-Storing and CRL-Retrieving
      Services", RFC 1424, RSA Laboratories, February 1993.



Kent                                                           [Page 30]

RFC 1422           Certificate-Based Key Management        February 1993


  [9] North American Directory Forum, "NADF Standing Documents: A Brief
      Overview", RFC 1417, NADF, February 1993.

Patent Statement

  This version of Privacy Enhanced Mail (PEM) relies on the use of
  patented public key encryption technology for authentication and
  encryption.  The Internet Standards Process as defined in RFC 1310
  requires a written statement from the Patent holder that a license
  will be made available to applicants under reasonable terms and
  conditions prior to approving a specification as a Proposed, Draft or
  Internet Standard.

  The Massachusetts Institute of Technology and the Board of Trustees
  of the Leland Stanford Junior University have granted Public Key
  Partners (PKP) exclusive sub-licensing rights to the following
  patents issued in the United States, and all of their corresponding
  foreign patents:

     Cryptographic Apparatus and Method
     ("Diffie-Hellman")............................... No. 4,200,770

     Public Key Cryptographic Apparatus
     and Method ("Hellman-Merkle").................... No. 4,218,582

     Cryptographic Communications System and
     Method ("RSA")................................... No. 4,405,829

     Exponential Cryptographic Apparatus
     and Method ("Hellman-Pohlig").................... No. 4,424,414

  These patents are stated by PKP to cover all known methods of
  practicing the art of Public Key encryption, including the variations
  collectively known as El Gamal.

  Public Key Partners has provided written assurance to the Internet
  Society that parties will be able to obtain, under reasonable,
  nondiscriminatory terms, the right to use the technology covered by
  these patents.  This assurance is documented in RFC 1170 titled
  "Public Key Standards and Licenses".  A copy of the written assurance
  dated April 20, 1990, may be obtained from the Internet Assigned
  Number Authority (IANA).

  The Internet Society, Internet Architecture Board, Internet
  Engineering Steering Group and the Corporation for National Research
  Initiatives take no position on the validity or scope of the patents
  and patent applications, nor on the appropriateness of the terms of
  the assurance.  The Internet Society and other groups mentioned above



Kent                                                           [Page 31]

RFC 1422           Certificate-Based Key Management        February 1993


  have not made any determination as to any other intellectual property
  rights which may apply to the practice of this standard. Any further
  consideration of these matters is the user's own responsibility.

Security Considerations

  This entire document is about security.

Author's Address

  Steve Kent
  BBN Communications
  50 Moulton Street
  Cambridge, MA 02138

  Phone: (617) 873-3988
  EMail: [email protected]


































Kent                                                           [Page 32]


You are viewing proxied material from gopher.fnord.one. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.