rfc1510

落日映苍穹つ 2021-11-22 02:58 267阅读 0赞

Network Working Group                                            J. Kohl  
Request for Comments: 1510                 Digital Equipment Corporation  
                                                               C. Neuman  
                                                                     ISI  
                                                          September 1993

The Kerberos Network Authentication Service (V5)

Status of this Memo

This RFC 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" for the standardization state and status  
   of this protocol.  Distribution of this memo is unlimited.

Abstract

This document gives an overview and specification of Version 5 of the  
   protocol for the Kerberos network authentication system. Version 4,  
   described elsewhere \[1,2\], is presently in production use at MIT's  
   Project Athena, and at other Internet sites.

Overview

Project Athena, Athena, Athena MUSE, Discuss, Hesiod, Kerberos,  
   Moira, and Zephyr are trademarks of the Massachusetts Institute of  
   Technology (MIT).  No commercial use of these trademarks may be made  
   without prior written permission of MIT.

This RFC describes the concepts and model upon which the Kerberos  
   network authentication system is based. It also specifies Version 5  
   of the Kerberos protocol.

The motivations, goals, assumptions, and rationale behind most design  
   decisions are treated cursorily; for Version 4 they are fully  
   described in the Kerberos portion of the Athena Technical Plan \[1\].  
   The protocols are under review, and are not being submitted for  
   consideration as an Internet standard at this time.  Comments are  
   encouraged.  Requests for addition to an electronic mailing list for  
   discussion of Kerberos, [kerberos@MIT.EDU][kerberos_MIT.EDU], may be addressed to  
   [kerberos-request@MIT.EDU][kerberos-request_MIT.EDU].  This mailing list is gatewayed onto the  
   Usenet as the group comp.protocols.kerberos.  Requests for further  
   information, including documents and code availability, may be sent  
   to [info-kerberos@MIT.EDU][info-kerberos_MIT.EDU].

Kohl & Neuman                                                   \[Page 1\]  
  
RFC 1510                        Kerberos                  September 1993

Background

The Kerberos model is based in part on Needham and Schroeder's  
   trusted third-party authentication protocol \[3\] and on modifications  
   suggested by Denning and Sacco \[4\].  The original design and  
   implementation of Kerberos Versions 1 through 4 was the work of two  
   former Project Athena staff members, Steve Miller of Digital  
   Equipment Corporation and Clifford Neuman (now at the Information  
   Sciences Institute of the University of Southern California), along  
   with Jerome Saltzer, Technical Director of Project Athena, and  
   Jeffrey Schiller, MIT Campus Network Manager.  Many other members of  
   Project Athena have also contributed to the work on Kerberos.  
   Version 4 is publicly available, and has seen wide use across the  
   Internet.

Version 5 (described in this document) has evolved from Version 4  
   based on new requirements and desires for features not available in  
   Version 4.  Details on the differences between Kerberos Versions 4  
   and 5 can be found in \[5\].

Table of Contents

1. Introduction .......................................    5  
   1.1. Cross-Realm Operation ............................    7  
   1.2. Environmental assumptions ........................    8  
   1.3. Glossary of terms ................................    9  
   2. Ticket flag uses and requests ......................   12  
   2.1. Initial and pre-authenticated tickets ............   12  
   2.2. Invalid tickets ..................................   12  
   2.3. Renewable tickets ................................   12  
   2.4. Postdated tickets ................................   13  
   2.5. Proxiable and proxy tickets ......................   14  
   2.6. Forwardable tickets ..............................   15  
   2.7. Other KDC options ................................   15  
   3. Message Exchanges ..................................   16  
   3.1. The Authentication Service Exchange ..............   16  
   3.1.1. Generation of KRB\_AS\_REQ message ...............   17  
   3.1.2. Receipt of KRB\_AS\_REQ message ..................   17  
   3.1.3. Generation of KRB\_AS\_REP message ...............   17  
   3.1.4. Generation of KRB\_ERROR message ................   19  
   3.1.5. Receipt of KRB\_AS\_REP message ..................   19  
   3.1.6. Receipt of KRB\_ERROR message ...................   20  
   3.2. The Client/Server Authentication Exchange ........   20  
   3.2.1. The KRB\_AP\_REQ message .........................   20  
   3.2.2. Generation of a KRB\_AP\_REQ message .............   20  
   3.2.3. Receipt of KRB\_AP\_REQ message ..................   21  
   3.2.4. Generation of a KRB\_AP\_REP message .............   23  
   3.2.5. Receipt of KRB\_AP\_REP message ..................   23

Kohl & Neuman                                                   \[Page 2\]  
  
RFC 1510                        Kerberos                  September 1993

3.2.6. Using the encryption key .......................   24  
   3.3. The Ticket-Granting Service (TGS) Exchange .......   24  
   3.3.1. Generation of KRB\_TGS\_REQ message ..............   25  
   3.3.2. Receipt of KRB\_TGS\_REQ message .................   26  
   3.3.3. Generation of KRB\_TGS\_REP message ..............   27  
   3.3.3.1. Encoding the transited field .................   29  
   3.3.4. Receipt of KRB\_TGS\_REP message .................   31  
   3.4. The KRB\_SAFE Exchange ............................   31  
   3.4.1. Generation of a KRB\_SAFE message ...............   31  
   3.4.2. Receipt of KRB\_SAFE message ....................   32  
   3.5. The KRB\_PRIV Exchange ............................   33  
   3.5.1. Generation of a KRB\_PRIV message ...............   33  
   3.5.2. Receipt of KRB\_PRIV message ....................   33  
   3.6. The KRB\_CRED Exchange ............................   34  
   3.6.1. Generation of a KRB\_CRED message ...............   34  
   3.6.2. Receipt of KRB\_CRED message ....................   34  
   4. The Kerberos Database ..............................   35  
   4.1. Database contents ................................   35  
   4.2. Additional fields ................................   36  
   4.3. Frequently Changing Fields .......................   37  
   4.4. Site Constants ...................................   37  
   5. Message Specifications .............................   38  
   5.1. ASN.1 Distinguished Encoding Representation ......   38  
   5.2. ASN.1 Base Definitions ...........................   38  
   5.3. Tickets and Authenticators .......................   42  
   5.3.1. Tickets ........................................   42  
   5.3.2. Authenticators .................................   47  
   5.4. Specifications for the AS and TGS exchanges ......   49  
   5.4.1. KRB\_KDC\_REQ definition .........................   49  
   5.4.2. KRB\_KDC\_REP definition .........................   56  
   5.5. Client/Server (CS) message specifications ........   58  
   5.5.1. KRB\_AP\_REQ definition ..........................   58  
   5.5.2. KRB\_AP\_REP definition ..........................   60  
   5.5.3. Error message reply ............................   61  
   5.6. KRB\_SAFE message specification ...................   61  
   5.6.1. KRB\_SAFE definition ............................   61  
   5.7. KRB\_PRIV message specification ...................   62  
   5.7.1. KRB\_PRIV definition ............................   62  
   5.8. KRB\_CRED message specification ...................   63  
   5.8.1. KRB\_CRED definition ............................   63  
   5.9. Error message specification ......................   65  
   5.9.1. KRB\_ERROR definition ...........................   66  
   6. Encryption and Checksum Specifications .............   67  
   6.1. Encryption Specifications ........................   68  
   6.2. Encryption Keys ..................................   71  
   6.3. Encryption Systems ...............................   71  
   6.3.1. The NULL Encryption System (null) ..............   71  
   6.3.2. DES in CBC mode with a CRC-32 checksum (descbc-crc)71

Kohl & Neuman                                                   \[Page 3\]  
  
RFC 1510                        Kerberos                  September 1993

6.3.3. DES in CBC mode with an MD4 checksum (descbc-md4)  72  
   6.3.4. DES in CBC mode with an MD5 checksum (descbc-md5)  72  
   6.4. Checksums ........................................   74  
   6.4.1. The CRC-32 Checksum (crc32) ....................   74  
   6.4.2. The RSA MD4 Checksum (rsa-md4) .................   75  
   6.4.3. RSA MD4 Cryptographic Checksum Using DES  
   (rsa-md4-des) .........................................   75  
   6.4.4. The RSA MD5 Checksum (rsa-md5) .................   76  
   6.4.5. RSA MD5 Cryptographic Checksum Using DES  
   (rsa-md5-des) .........................................   76  
   6.4.6. DES cipher-block chained checksum (des-mac)  
   6.4.7. RSA MD4 Cryptographic Checksum Using DES  
   alternative (rsa-md4-des-k) ...........................   77  
   6.4.8. DES cipher-block chained checksum alternative  
   (des-mac-k) ...........................................   77  
   7. Naming Constraints .................................   78  
   7.1. Realm Names ......................................   77  
   7.2. Principal Names ..................................   79  
   7.2.1. Name of server principals ......................   80  
   8. Constants and other defined values .................   80  
   8.1. Host address types ...............................   80  
   8.2. KDC messages .....................................   81  
   8.2.1. IP transport ...................................   81  
   8.2.2. OSI transport ..................................   82  
   8.2.3. Name of the TGS ................................   82  
   8.3. Protocol constants and associated values .........   82  
   9. Interoperability requirements ......................   86  
   9.1. Specification 1 ..................................   86  
   9.2. Recommended KDC values ...........................   88  
   10. Acknowledgments ...................................   88  
   11. References ........................................   89  
   12. Security Considerations ...........................   90  
   13. Authors' Addresses ................................   90  
   A. Pseudo-code for protocol processing ................   91  
   A.1. KRB\_AS\_REQ generation ............................   91  
   A.2. KRB\_AS\_REQ verification and KRB\_AS\_REP generation    92  
   A.3. KRB\_AS\_REP verification ..........................   95  
   A.4. KRB\_AS\_REP and KRB\_TGS\_REP common checks .........   96  
   A.5. KRB\_TGS\_REQ generation ...........................   97  
   A.6. KRB\_TGS\_REQ verification and KRB\_TGS\_REP generation  98  
   A.7. KRB\_TGS\_REP verification .........................  104  
   A.8. Authenticator generation .........................  104  
   A.9. KRB\_AP\_REQ generation ............................  105  
   A.10. KRB\_AP\_REQ verification .........................  105  
   A.11. KRB\_AP\_REP generation ...........................  106  
   A.12. KRB\_AP\_REP verification .........................  107  
   A.13. KRB\_SAFE generation .............................  107  
   A.14. KRB\_SAFE verification ...........................  108

Kohl & Neuman                                                   \[Page 4\]  
  
RFC 1510                        Kerberos                  September 1993

A.15. KRB\_SAFE and KRB\_PRIV common checks .............  108  
   A.16. KRB\_PRIV generation .............................  109  
   A.17. KRB\_PRIV verification ...........................  110  
   A.18. KRB\_CRED generation .............................  110  
   A.19. KRB\_CRED verification ...........................  111  
   A.20. KRB\_ERROR generation ............................  112

1.  Introduction

Kerberos provides a means of verifying the identities of principals,  
   (e.g., a workstation user or a network server) on an open  
   (unprotected) network.  This is accomplished without relying on  
   authentication by the host operating system, without basing trust on  
   host addresses, without requiring physical security of all the hosts  
   on the network, and under the assumption that packets traveling along  
   the network can be read, modified, and inserted at will. (Note,  
   however, that many applications use Kerberos' functions only upon the  
   initiation of a stream-based network connection, and assume the  
   absence of any "hijackers" who might subvert such a connection.  Such  
   use implicitly trusts the host addresses involved.)  Kerberos  
   performs authentication under these conditions as a trusted third-  
   party authentication service by using conventional cryptography,  
   i.e., shared secret key.  (shared secret key - Secret and private are  
   often used interchangeably in the literature.  In our usage, it takes  
   two (or more) to share a secret, thus a shared DES key is a secret  
   key.  Something is only private when no one but its owner knows it.  
   Thus, in public key cryptosystems, one has a public and a private  
   key.)

The authentication process proceeds as follows: A client sends a  
   request to the authentication server (AS) requesting "credentials"  
   for a given server.  The AS responds with these credentials,  
   encrypted in the client's key.  The credentials consist of 1) a  
   "ticket" for the server and 2) a temporary encryption key (often  
   called a "session key").  The client transmits the ticket (which  
   contains the client's identity and a copy of the session key, all  
   encrypted in the server's key) to the server.  The session key (now  
   shared by the client and server) is used to authenticate the client,  
   and may optionally be used to authenticate the server.  It may also  
   be used to encrypt further communication between the two parties or  
   to exchange a separate sub-session key to be used to encrypt further  
   communication.

The implementation consists of one or more authentication servers  
   running on physically secure hosts.  The authentication servers  
   maintain a database of principals (i.e., users and servers) and their  
   secret keys. Code libraries provide encryption and implement the  
   Kerberos protocol.  In order to add authentication to its

Kohl & Neuman                                                   \[Page 5\]  
  
RFC 1510                        Kerberos                  September 1993

transactions, a typical network application adds one or two calls to  
   the Kerberos library, which results in the transmission of the  
   necessary messages to achieve authentication.

The Kerberos protocol consists of several sub-protocols (or  
   exchanges).  There are two methods by which a client can ask a  
   Kerberos server for credentials.  In the first approach, the client  
   sends a cleartext request for a ticket for the desired server to the  
   AS. The reply is sent encrypted in the client's secret key. Usually  
   this request is for a ticket-granting ticket (TGT) which can later be  
   used with the ticket-granting server (TGS).  In the second method,  
   the client sends a request to the TGS.  The client sends the TGT to  
   the TGS in the same manner as if it were contacting any other  
   application server which requires Kerberos credentials.  The reply is  
   encrypted in the session key from the TGT.

Once obtained, credentials may be used to verify the identity of the  
   principals in a transaction, to ensure the integrity of messages  
   exchanged between them, or to preserve privacy of the messages.  The  
   application is free to choose whatever protection may be necessary.

To verify the identities of the principals in a transaction, the  
   client transmits the ticket to the server.  Since the ticket is sent  
   "in the clear" (parts of it are encrypted, but this encryption  
   doesn't thwart replay) and might be intercepted and reused by an  
   attacker, additional information is sent to prove that the message  
   was originated by the principal to whom the ticket was issued.  This  
   information (called the authenticator) is encrypted in the session  
   key, and includes a timestamp.  The timestamp proves that the message  
   was recently generated and is not a replay.  Encrypting the  
   authenticator in the session key proves that it was generated by a  
   party possessing the session key.  Since no one except the requesting  
   principal and the server know the session key (it is never sent over  
   the network in the clear) this guarantees the identity of the client.

The integrity of the messages exchanged between principals can also  
   be guaranteed using the session key (passed in the ticket and  
   contained in the credentials).  This approach provides detection of  
   both replay attacks and message stream modification attacks.  It is  
   accomplished by generating and transmitting a collision-proof  
   checksum (elsewhere called a hash or digest function) of the client's  
   message, keyed with the session key.  Privacy and integrity of the  
   messages exchanged between principals can be secured by encrypting  
   the data to be passed using the session key passed in the ticket, and  
   contained in the credentials.

The authentication exchanges mentioned above require read-only access  
   to the Kerberos database.  Sometimes, however, the entries in the

Kohl & Neuman                                                   \[Page 6\]  
  
RFC 1510                        Kerberos                  September 1993

database must be modified, such as when adding new principals or  
   changing a principal's key.  This is done using a protocol between a  
   client and a third Kerberos server, the Kerberos Administration  
   Server (KADM).  The administration protocol is not described in this  
   document. There is also a protocol for maintaining multiple copies of  
   the Kerberos database, but this can be considered an implementation  
   detail and may vary to support different database technologies.

1.1.  Cross-Realm Operation

The Kerberos protocol is designed to operate across organizational  
   boundaries.  A client in one organization can be authenticated to a  
   server in another.  Each organization wishing to run a Kerberos  
   server establishes its own "realm".  The name of the realm in which a  
   client is registered is part of the client's name, and can be used by  
   the end-service to decide whether to honor a request.

By establishing "inter-realm" keys, the administrators of two realms  
   can allow a client authenticated in the local realm to use its  
   authentication remotely (Of course, with appropriate permission the  
   client could arrange registration of a separately-named principal in  
   a remote realm, and engage in normal exchanges with that realm's  
   services. However, for even small numbers of clients this becomes  
   cumbersome, and more automatic methods as described here are  
   necessary).  The exchange of inter-realm keys (a separate key may be  
   used for each direction) registers the ticket-granting service of  
   each realm as a principal in the other realm.  A client is then able  
   to obtain a ticket-granting ticket for the remote realm's ticket-  
   granting service from its local realm. When that ticket-granting  
   ticket is used, the remote ticket-granting service uses the inter-  
   realm key (which usually differs from its own normal TGS key) to  
   decrypt the ticket-granting ticket, and is thus certain that it was  
   issued by the client's own TGS. Tickets issued by the remote ticket-  
   granting service will indicate to the end-service that the client was  
   authenticated from another realm.

A realm is said to communicate with another realm if the two realms  
   share an inter-realm key, or if the local realm shares an inter-realm  
   key with an intermediate realm that communicates with the remote  
   realm.  An authentication path is the sequence of intermediate realms  
   that are transited in communicating from one realm to another.

Realms are typically organized hierarchically. Each realm shares a  
   key with its parent and a different key with each child.  If an  
   inter-realm key is not directly shared by two realms, the  
   hierarchical organization allows an authentication path to be easily  
   constructed.  If a hierarchical organization is not used, it may be  
   necessary to consult some database in order to construct an

Kohl & Neuman                                                   \[Page 7\]  
  
RFC 1510                        Kerberos                  September 1993

authentication path between realms.

Although realms are typically hierarchical, intermediate realms may  
   be bypassed to achieve cross-realm authentication through alternate  
   authentication paths (these might be established to make  
   communication between two realms more efficient).  It is important  
   for the end-service to know which realms were transited when deciding  
   how much faith to place in the authentication process. To facilitate  
   this decision, a field in each ticket contains the names of the  
   realms that were involved in authenticating the client.

1.2.  Environmental assumptions

Kerberos imposes a few assumptions on the environment in which it can  
   properly function:

+    "Denial of service" attacks are not solved with Kerberos.  There  
        are places in these protocols where an intruder intruder can  
        prevent an application from participating in the proper  
        authentication steps.  Detection and solution of such attacks  
        (some of which can appear to be not-uncommon "normal" failure  
        modes for the system) is usually best left to the human  
        administrators and users.

+    Principals must keep their secret keys secret.  If an intruder  
        somehow steals a principal's key, it will be able to masquerade  
        as that principal or impersonate any server to the legitimate  
        principal.

+    "Password guessing" attacks are not solved by Kerberos.  If a  
        user chooses a poor password, it is possible for an attacker to  
        successfully mount an offline dictionary attack by repeatedly  
        attempting to decrypt, with successive entries from a  
        dictionary, messages obtained which are encrypted under a key  
        derived from the user's password.

+    Each host on the network must have a clock which is "loosely  
        synchronized" to the time of the other hosts; this  
        synchronization is used to reduce the bookkeeping needs of  
        application servers when they do replay detection.  The degree  
        of "looseness" can be configured on a per-server basis.  If the  
        clocks are synchronized over the network, the clock  
        synchronization protocol must itself be secured from network  
        attackers.

+    Principal identifiers are not recycled on a short-term basis.  A  
        typical mode of access control will use access control lists  
        (ACLs) to grant permissions to particular principals.  If a

Kohl & Neuman                                                   \[Page 8\]  
  
RFC 1510                        Kerberos                  September 1993

stale ACL entry remains for a deleted principal and the  
        principal identifier is reused, the new principal will inherit  
        rights specified in the stale ACL entry. By not re-using  
        principal identifiers, the danger of inadvertent access is  
        removed.

1.3.  Glossary of terms

Below is a list of terms used throughout this document.

Authentication      Verifying the claimed identity of a  
                       principal.

Authentication header A record containing a Ticket and an  
                         Authenticator to be presented to a  
                         server as part of the authentication  
                         process.

Authentication path  A sequence of intermediate realms transited  
                        in the authentication process when  
                        communicating from one realm to another.

Authenticator       A record containing information that can  
                       be shown to have been recently generated  
                       using the session key known only by  the  
                       client and server.

Authorization       The process of determining whether a  
                       client may use a service, which objects  
                       the client is allowed to access, and the  
                       type of access allowed for each.

Capability          A token that grants the bearer permission  
                       to access an object or service.  In  
                       Kerberos, this might be a ticket whose  
                       use is restricted by the contents of the  
                       authorization data field, but which  
                       lists no network addresses, together  
                       with the session key necessary to use  
                       the ticket.

Kohl & Neuman                                                   \[Page 9\]  
  
RFC 1510                        Kerberos                  September 1993

Ciphertext          The output of an encryption function.  
                       Encryption transforms plaintext into  
                       ciphertext.

Client              A process that makes use of a network  
                       service on behalf of a user.  Note that  
                       in some cases a Server may itself be a  
                       client of some other server (e.g., a  
                       print server may be a client of a file  
                       server).

Credentials         A ticket plus the secret session key  
                       necessary to successfully use that  
                       ticket in an authentication exchange.

KDC                 Key Distribution Center, a network service  
                       that supplies tickets and temporary  
                       session keys; or an instance of that  
                       service or the host on which it runs.  
                       The KDC services both initial ticket and  
                       ticket-granting ticket requests.  The  
                       initial ticket portion is sometimes  
                       referred to as the Authentication Server  
                       (or service).  The ticket-granting  
                       ticket portion is sometimes referred to  
                       as the ticket-granting server (or service).

Kerberos            Aside from the 3-headed dog guarding  
                       Hades, the name given to Project  
                       Athena's authentication service, the  
                       protocol used by that service, or the  
                       code used to implement the authentication  
                       service.

Plaintext           The input to an encryption function  or  
                       the output of a decryption function.  
                       Decryption transforms ciphertext into  
                       plaintext.

Principal           A uniquely named client or server  
                       instance that participates in a network  
                       communication.

Kohl & Neuman                                                  \[Page 10\]  
  
RFC 1510                        Kerberos                  September 1993

Principal identifier The name used to uniquely identify each  
                        different principal.

Seal                To encipher a record containing several  
                       fields in such a way that the fields  
                       cannot be individually replaced without  
                       either knowledge of the encryption key  
                       or leaving evidence of tampering.

Secret key          An encryption key shared by a principal  
                       and the KDC, distributed outside the  
                       bounds of the system, with a long lifetime.  
                       In the case of a human user's  
                       principal, the secret key is derived  
                       from a password.

Server              A particular Principal which provides a  
                       resource to network clients.

Service             A resource provided to network clients;  
                       often provided by more than one server  
                       (for example, remote file service).

Session key         A temporary encryption key used between  
                       two principals, with a lifetime limited  
                       to the duration of a single login "session".

Sub-session key     A temporary encryption key used between  
                       two principals, selected and exchanged  
                       by the principals using the session key,  
                       and with a lifetime limited to the duration  
                       of a single association.

Ticket              A record that helps a client authenticate  
                       itself to a server; it contains the  
                       client's identity, a session key, a  
                       timestamp, and other information, all  
                       sealed using the server's secret key.  
                       It only serves to authenticate a client  
                       when presented along with a fresh  
                       Authenticator.

Kohl & Neuman                                                  \[Page 11\]  
  
RFC 1510                        Kerberos                  September 1993

2.  Ticket flag uses and requests

Each Kerberos ticket contains a set of flags which are used to  
   indicate various attributes of that ticket.  Most flags may be  
   requested by a client when the ticket is obtained; some are  
   automatically turned on and off by a Kerberos server as required.  
   The following sections explain what the various flags mean, and gives  
   examples of reasons to use such a flag.

2.1.  Initial and pre-authenticated tickets

The INITIAL flag indicates that a ticket was issued using the AS  
   protocol and not issued based on a ticket-granting ticket.  
   Application servers that want to require the knowledge of a client's  
   secret key (e.g., a passwordchanging program) can insist that this  
   flag be set in any tickets they accept, and thus be assured that the  
   client's key was recently presented to the application client.

The PRE-AUTHENT and HW-AUTHENT flags provide addition information  
   about the initial authentication, regardless of whether the current  
   ticket was issued directly (in which case INITIAL will also be set)  
   or issued on the basis of a ticket-granting ticket (in which case the  
   INITIAL flag is clear, but the PRE-AUTHENT and HW-AUTHENT flags are  
   carried forward from the ticket-granting ticket).

2.2.  Invalid tickets

The INVALID flag indicates that a ticket is invalid.  Application  
   servers must reject tickets which have this flag set.  A postdated  
   ticket will usually be issued in this form. Invalid tickets must be  
   validated by the KDC before use, by presenting them to the KDC in a  
   TGS request with the VALIDATE option specified.  The KDC will only  
   validate tickets after their starttime has passed.  The validation is  
   required so that postdated tickets which have been stolen before  
   their starttime can be rendered permanently invalid (through a hot-  
   list mechanism).

2.3.  Renewable tickets

Applications may desire to hold tickets which can be valid for long  
   periods of time.  However, this can expose their credentials to  
   potential theft for equally long periods, and those stolen  
   credentials would be valid until the expiration time of the  
   ticket(s).  Simply using shortlived tickets and obtaining new ones  
   periodically would require the client to have long-term access to its  
   secret key, an even greater risk.  Renewable tickets can be used to  
   mitigate the consequences of theft.  Renewable tickets have two  
   "expiration times": the first is when the current instance of the

Kohl & Neuman                                                  \[Page 12\]  
  
RFC 1510                        Kerberos                  September 1993

ticket expires, and the second is the latest permissible value for an  
   individual expiration time.  An application client must periodically  
   (i.e., before it expires) present a renewable ticket to the KDC, with  
   the RENEW option set in the KDC request.  The KDC will issue a new  
   ticket with a new session key and a later expiration time.  All other  
   fields of the ticket are left unmodified by the renewal process.  
   When the latest permissible expiration time arrives, the ticket  
   expires permanently.  At each renewal, the KDC may consult a hot-list  
   to determine if the ticket had been reported stolen since its last  
   renewal; it will refuse to renew such stolen tickets, and thus the  
   usable lifetime of stolen tickets is reduced.

The RENEWABLE flag in a ticket is normally only interpreted by the  
   ticket-granting service (discussed below in section 3.3).  It can  
   usually be ignored by application servers.  However, some  
   particularly careful application servers may wish to disallow  
   renewable tickets.

If a renewable ticket is not renewed by its  expiration time, the KDC  
   will not renew the ticket.  The RENEWABLE flag is reset by default,  
   but a client may request it be  set  by setting  the RENEWABLE option  
   in the KRB\_AS\_REQ message.  If it is set, then the renew-till field  
   in the ticket  contains the time after which the ticket may not be  
   renewed.

2.4.  Postdated tickets

Applications may occasionally need to obtain tickets for use much  
   later, e.g., a batch submission system would need tickets to be valid  
   at the time the batch job is serviced.  However, it is dangerous to  
   hold valid tickets in a batch queue, since they will be on-line  
   longer and more prone to theft.  Postdated tickets provide a way to  
   obtain these tickets from the KDC at job submission time, but to  
   leave them "dormant" until they are activated and validated by a  
   further request of the KDC.  If a ticket theft were reported in the  
   interim, the KDC would refuse to validate the ticket, and the thief  
   would be foiled.

The MAY-POSTDATE flag in a ticket is normally only interpreted by the  
   ticket-granting service.  It can be ignored by application servers.  
   This flag must be set in a ticket-granting ticket in order to issue a  
   postdated ticket based on the presented ticket. It is reset by  
   default; it may be requested by a client by setting the ALLOW-  
   POSTDATE option in the KRB\_AS\_REQ message.  This flag does not allow  
   a client to obtain a postdated ticket-granting ticket; postdated  
   ticket-granting tickets can only by obtained by requesting the  
   postdating in the KRB\_AS\_REQ message.  The life (endtime-starttime)  
   of a postdated ticket will be the remaining life of the ticket-

Kohl & Neuman                                                  \[Page 13\]  
  
RFC 1510                        Kerberos                  September 1993

granting ticket at the time of the request, unless the RENEWABLE  
   option is also set, in which case it can be the full life (endtime-  
   starttime) of the ticket-granting ticket.  The KDC may limit how far  
   in the future a ticket may be postdated.

The POSTDATED flag indicates that a ticket has been postdated.  The  
   application server can check the authtime field in the ticket to see  
   when the original authentication occurred.  Some services may choose  
   to reject postdated tickets, or they may only accept them within a  
   certain period after the original authentication. When the KDC issues  
   a POSTDATED ticket, it will also be marked as INVALID, so that the  
   application client must present the ticket to the KDC to be validated  
   before use.

2.5.  Proxiable and proxy tickets

At times it may be necessary for a principal to allow a service  to  
   perform an operation on its behalf.  The service must be able to take  
   on the identity of the client, but only for  a particular purpose.  A  
   principal can allow a service to take on the principal's identity for  
   a particular purpose by granting it a proxy.

The PROXIABLE flag in a ticket is normally only interpreted by the  
   ticket-granting service. It can be ignored by application servers.  
   When set, this flag tells the ticket-granting server that it is OK to  
   issue a new ticket (but not a ticket-granting ticket) with a  
   different network address based on this ticket.  This flag is set by  
   default.

This flag allows a client to pass a proxy to a server to perform a  
   remote request on its behalf, e.g., a print service client can give  
   the print server a proxy to access the client's files on a particular  
   file server in order to satisfy a print request.

In order to complicate the use of stolen credentials, Kerberos  
   tickets are usually valid from only those network addresses  
   specifically included in the ticket (It is permissible to request or  
   issue tickets with no network addresses specified, but we do not  
   recommend it).  For this reason, a client wishing to grant a proxy  
   must request a new ticket valid for the network address of the  
   service to be granted the proxy.

The PROXY flag is set in a ticket by the  TGS  when  it issues a  
   proxy ticket.  Application servers may check this flag and require  
   additional authentication  from  the  agent presenting the proxy in  
   order to provide an audit trail.

Kohl & Neuman                                                  \[Page 14\]  
  
RFC 1510                        Kerberos                  September 1993

2.6.  Forwardable tickets

Authentication forwarding is an instance of the proxy case where the  
   service is granted complete use of the client's identity.  An example  
   where it might be used is when a user logs in to a remote system and  
   wants authentication to work from that system as if the login were  
   local.

The FORWARDABLE flag in a ticket is normally only interpreted by the  
   ticket-granting service.  It can be ignored by application servers.  
   The FORWARDABLE flag has an interpretation similar to that of the  
   PROXIABLE flag, except ticket-granting tickets may also be issued  
   with different network addresses.  This flag is reset by default, but  
   users may request that it be set by setting the FORWARDABLE option in  
   the AS request when they request their initial ticket-granting  
   ticket.

This flag allows for authentication forwarding without requiring the  
   user to enter a password again.  If the flag is not set, then  
   authentication forwarding is not permitted, but the same end result  
   can still be achieved if the user engages in the AS exchange with the  
   requested network addresses and supplies a password.

The FORWARDED flag is set by the TGS when a client presents a ticket  
   with the FORWARDABLE flag set and requests it be set by specifying  
   the FORWARDED KDC option and supplying a set of addresses for the new  
   ticket.  It is also set in all tickets issued based on tickets with  
   the FORWARDED flag set.  Application servers may wish to process  
   FORWARDED tickets differently than non-FORWARDED tickets.

2.7.  Other KDC options

There are two additional options which may be set in a client's  
   request of the KDC.  The RENEWABLE-OK option indicates that the  
   client will accept a renewable ticket if a ticket with the requested  
   life cannot otherwise be provided.  If a ticket with the requested  
   life cannot be provided, then the KDC may issue a renewable ticket  
   with a renew-till equal to the the requested endtime.  The value of  
   the renew-till field may still be adjusted by site-determined limits  
   or limits imposed by the individual principal or server.

The ENC-TKT-IN-SKEY option is honored only by the ticket-granting  
   service.  It indicates that the to-be-issued ticket for the end  
   server is to be encrypted in the session key from the additional  
   ticket-granting ticket provided with the request.  See section 3.3.3  
   for specific details.

Kohl & Neuman                                                  \[Page 15\]  
  
RFC 1510                        Kerberos                  September 1993

3.  Message Exchanges

The following sections describe the interactions between network  
   clients and servers and the messages involved in those exchanges.

3.1.  The Authentication Service Exchange

Summary

Message direction       Message type    Section  
         1. Client to Kerberos   KRB\_AS\_REQ      5.4.1  
         2. Kerberos to client   KRB\_AS\_REP or   5.4.2  
                                 KRB\_ERROR       5.9.1

The Authentication Service (AS) Exchange between the client and the  
   Kerberos Authentication Server is usually initiated by a client when  
   it wishes to obtain authentication credentials for a given server but  
   currently holds no credentials.  The client's secret key is used for  
   encryption and decryption.  This exchange is typically used at the  
   initiation of a login session, to obtain credentials for a Ticket-  
   Granting Server, which will subsequently be used to obtain  
   credentials for other servers (see section 3.3) without requiring  
   further use of the client's secret key.  This exchange is also used  
   to request credentials for services which must not be mediated  
   through the Ticket-Granting Service, but rather require a principal's  
   secret key, such as the password-changing service.  (The password-  
   changing request must not be honored unless the requester can provide  
   the old password (the user's current secret key).  Otherwise, it  
   would be possible for someone to walk up to an unattended session and  
   change another user's password.)  This exchange does not by itself  
   provide any assurance of the the identity of the user.  (To  
   authenticate a user logging on to a local system, the credentials  
   obtained in the AS exchange may first be used in a TGS exchange to  
   obtain credentials for a local server.  Those credentials must then  
   be verified by the local server through successful completion of the  
   Client/Server exchange.)

The exchange consists of two messages: KRB\_AS\_REQ from the client to  
   Kerberos, and KRB\_AS\_REP or KRB\_ERROR in reply. The formats for these  
   messages are described in sections 5.4.1, 5.4.2, and 5.9.1.

In the request, the client sends (in cleartext) its own identity and  
   the identity of the server for which it is requesting credentials.  
   The response, KRB\_AS\_REP, contains a ticket for the client to present  
   to the server, and a session key that will be shared by the client  
   and the server.  The session key and additional information are  
   encrypted in the client's secret key.  The KRB\_AS\_REP message  
   contains information which can be used to detect replays, and to

Kohl & Neuman                                                  \[Page 16\]  
  
RFC 1510                        Kerberos                  September 1993

associate it with the message to which it replies.  Various errors  
   can occur; these are indicated by an error response (KRB\_ERROR)  
   instead of the KRB\_AS\_REP response.  The error message is not  
   encrypted.  The KRB\_ERROR message also contains information which can  
   be used to associate it with the message to which it replies.  The  
   lack of encryption in the KRB\_ERROR message precludes the ability to  
   detect replays or fabrications of such messages.

In the normal case the authentication server does not know whether  
   the client is actually the principal named in the request.  It simply  
   sends a reply without knowing or caring whether they are the same.  
   This is acceptable because nobody but the principal whose identity  
   was given in the request will be able to use the reply. Its critical  
   information is encrypted in that principal's key.  The initial  
   request supports an optional field that can be used to pass  
   additional information that might be needed for the initial exchange.  
   This field may be used for preauthentication if desired, but the  
   mechanism is not currently specified.

3.1.1. Generation of KRB\_AS\_REQ message

The client may specify a number of options in the initial request.  
   Among these options are whether preauthentication is to be performed;  
   whether the requested ticket is to be renewable, proxiable, or  
   forwardable; whether it should be postdated or allow postdating of  
   derivative tickets; and whether a renewable ticket will be accepted  
   in lieu of a non-renewable ticket if the requested ticket expiration  
   date cannot be satisfied by a nonrenewable ticket (due to  
   configuration constraints; see section 4).  See section A.1 for  
   pseudocode.

The client prepares the KRB\_AS\_REQ message and sends it to the KDC.

3.1.2. Receipt of KRB\_AS\_REQ message

If all goes well, processing the KRB\_AS\_REQ message will result in  
   the creation of a ticket for the client to present to the server.  
   The format for the ticket is described in section 5.3.1.  The  
   contents of the ticket are determined as follows.

3.1.3. Generation of KRB\_AS\_REP message

The authentication server looks up the client and server principals  
   named in the KRB\_AS\_REQ in its database, extracting their respective  
   keys.  If required, the server pre-authenticates the request, and if  
   the pre-authentication check fails, an error message with the code  
   KDC\_ERR\_PREAUTH\_FAILED is returned. If the server cannot accommodate  
   the requested encryption type, an error message with code

Kohl & Neuman                                                  \[Page 17\]  
  
RFC 1510                        Kerberos                  September 1993

KDC\_ERR\_ETYPE\_NOSUPP is returned. Otherwise it generates a "random"  
   session key ("Random" means that, among other things, it should be  
   impossible to guess the next session key based on knowledge of past  
   session keys.  This can only be achieved in a pseudo-random number  
   generator if it is based on cryptographic principles.  It would be  
   more desirable to use a truly random number generator, such as one  
   based on measurements of random physical phenomena.).

If the requested start time is absent or indicates a time in the  
   past, then the start time of the ticket is set to the authentication  
   server's current time. If it indicates a time in the future, but the  
   POSTDATED option has not been specified, then the error  
   KDC\_ERR\_CANNOT\_POSTDATE is returned.  Otherwise the requested start  
   time is checked against the policy of the local realm (the  
   administrator might decide to prohibit certain types or ranges of  
   postdated tickets), and if acceptable, the ticket's start time is set  
   as requested and the INVALID flag is set in the new ticket. The  
   postdated ticket must be validated before use by presenting it to the  
   KDC after the start time has been reached.

The expiration time of the ticket will be set to the minimum of the  
   following:

+The expiration time (endtime) requested in the KRB\_AS\_REQ  
    message.

+The ticket's start time plus the maximum allowable lifetime  
    associated with the client principal (the authentication  
    server's database includes a maximum ticket lifetime field  
    in each principal's record; see section 4).

+The ticket's start time plus the maximum allowable lifetime  
    associated with the server principal.

+The ticket's start time plus the maximum lifetime set by  
    the policy of the local realm.

If the requested expiration time minus the start time (as determined  
   above) is less than a site-determined minimum lifetime, an error  
   message with code KDC\_ERR\_NEVER\_VALID is returned.  If the requested  
   expiration time for the ticket exceeds what was determined as above,  
   and if the "RENEWABLE-OK" option was requested, then the "RENEWABLE"  
   flag is set in the new ticket, and the renew-till value is set as if  
   the "RENEWABLE" option were requested (the field and option names are  
   described fully in section 5.4.1).  If the RENEWABLE option has been  
   requested or if the RENEWABLE-OK option has been set and a renewable  
   ticket is to be issued, then the renew-till field is set to the  
   minimum of:

Kohl & Neuman                                                  \[Page 18\]  
  
RFC 1510                        Kerberos                  September 1993

+Its requested value.

+The start time of the ticket plus the minimum of the two  
    maximum renewable lifetimes associated with the principals'  
    database entries.

+The start time of the ticket plus the maximum renewable  
    lifetime set by the policy of the local realm.

The flags field of the new ticket will have the following options set  
   if they have been requested and if the policy of the local realm  
   allows: FORWARDABLE, MAY-POSTDATE, POSTDATED, PROXIABLE, RENEWABLE.  
   If the new ticket is postdated (the start time is in the future), its  
   INVALID flag will also be set.

If all of the above succeed, the server formats a KRB\_AS\_REP message  
   (see section 5.4.2), copying the addresses in the request into the  
   caddr of the response, placing any required pre-authentication data  
   into the padata of the response, and encrypts the ciphertext part in  
   the client's key using the requested encryption method, and sends it  
   to the client.  See section A.2 for pseudocode.

3.1.4. Generation of KRB\_ERROR message

Several errors can occur, and the Authentication Server responds by  
   returning an error message, KRB\_ERROR, to the client, with the  
   error-code and e-text fields set to appropriate values.  The error  
   message contents and details are described in Section 5.9.1.

3.1.5. Receipt of KRB\_AS\_REP message

If the reply message type is KRB\_AS\_REP, then the client verifies  
   that the cname and crealm fields in the cleartext portion of the  
   reply match what it requested.  If any padata fields are present,  
   they may be used to derive the proper secret key to decrypt the  
   message.  The client decrypts the encrypted part of the response  
   using its secret key, verifies that the nonce in the encrypted part  
   matches the nonce it supplied in its request (to detect replays).  It  
   also verifies that the sname and srealm in the response match those  
   in the request, and that the host address field is also correct.  It  
   then stores the ticket, session key, start and expiration times, and  
   other information for later use.  The key-expiration field from the  
   encrypted part of the response may be checked to notify the user of  
   impending key expiration (the client program could then suggest  
   remedial action, such as a password change).  See section A.3 for  
   pseudocode.

Proper decryption of the KRB\_AS\_REP message is not sufficient to

Kohl & Neuman                                                  \[Page 19\]  
  
RFC 1510                        Kerberos                  September 1993

verify the identity of the user; the user and an attacker could  
   cooperate to generate a KRB\_AS\_REP format message which decrypts  
   properly but is not from the proper KDC.  If the host wishes to  
   verify the identity of the user, it must require the user to present  
   application credentials which can be verified using a securely-stored  
   secret key.  If those credentials can be verified, then the identity  
   of the user can be assured.

3.1.6. Receipt of KRB\_ERROR message

If the reply message type is KRB\_ERROR, then the client interprets it  
   as an error and performs whatever application-specific tasks are  
   necessary to recover.

3.2.  The Client/Server Authentication Exchange

Summary

Message direction                         Message type    Section  
   Client to Application server              KRB\_AP\_REQ      5.5.1  
   \[optional\] Application server to client   KRB\_AP\_REP or   5.5.2  
                                             KRB\_ERROR       5.9.1

The client/server authentication (CS) exchange is used by network  
   applications to authenticate the client to the server and vice versa.  
   The client must have already acquired credentials for the server  
   using the AS or TGS exchange.

3.2.1. The KRB\_AP\_REQ message

The KRB\_AP\_REQ contains authentication information which should be  
   part of the first message in an authenticated transaction.  It  
   contains a ticket, an authenticator, and some additional bookkeeping  
   information (see section 5.5.1 for the exact format).  The ticket by  
   itself is insufficient to authenticate a client, since tickets are  
   passed across the network in cleartext(Tickets contain both an  
   encrypted and unencrypted portion, so cleartext here refers to the  
   entire unit, which can be copied from one message and replayed in  
   another without any cryptographic skill.), so the authenticator is  
   used to prevent invalid replay of tickets by proving to the server  
   that the client knows the session key of the ticket and thus is  
   entitled to use it.  The KRB\_AP\_REQ message is referred to elsewhere  
   as the "authentication header."

3.2.2. Generation of a KRB\_AP\_REQ message

When a client wishes to initiate authentication to a server, it  
   obtains (either through a credentials cache, the AS exchange, or the

Kohl & Neuman                                                  \[Page 20\]  
  
RFC 1510                        Kerberos                  September 1993

TGS exchange) a ticket and session key for the desired service.  The  
   client may re-use any tickets it holds until they expire.  The client  
   then constructs a new Authenticator from the the system time, its  
   name, and optionally an application specific checksum, an initial  
   sequence number to be used in KRB\_SAFE or KRB\_PRIV messages, and/or a  
   session subkey to be used in negotiations for a session key unique to  
   this particular session.  Authenticators may not be re-used and will  
   be rejected if replayed to a server (Note that this can make  
   applications based on unreliable transports difficult to code  
   correctly, if the transport might deliver duplicated messages.  In  
   such cases, a new authenticator must be generated for each retry.).  
   If a sequence number is to be included, it should be randomly chosen  
   so that even after many messages have been exchanged it is not likely  
   to collide with other sequence numbers in use.

The client may indicate a requirement of mutual authentication or the  
   use of a session-key based ticket by setting the appropriate flag(s)  
   in the ap-options field of the message.

The Authenticator is encrypted in the session key and combined with  
   the ticket to form the KRB\_AP\_REQ message which is then sent to the  
   end server along with any additional application-specific  
   information.  See section A.9 for pseudocode.

3.2.3. Receipt of KRB\_AP\_REQ message

Authentication is based on the server's current time of day (clocks  
   must be loosely synchronized), the authenticator, and the ticket.  
   Several errors are possible.  If an error occurs, the server is  
   expected to reply to the client with a KRB\_ERROR message.  This  
   message may be encapsulated in the application protocol if its "raw"  
   form is not acceptable to the protocol. The format of error messages  
   is described in section 5.9.1.

The algorithm for verifying authentication information is as follows.  
   If the message type is not KRB\_AP\_REQ, the server returns the  
   KRB\_AP\_ERR\_MSG\_TYPE error. If the key version indicated by the Ticket  
   in the KRB\_AP\_REQ is not one the server can use (e.g., it indicates  
   an old key, and the server no longer possesses a copy of the old  
   key), the KRB\_AP\_ERR\_BADKEYVER error is returned.  If the USE-  
   SESSION-KEY flag is set in the ap-options field, it indicates to the  
   server that the ticket is encrypted in the session key from the  
   server's ticket-granting ticket rather than its secret key (This is  
   used for user-to-user authentication as described in \[6\]).  Since it  
   is possible for the server to be registered in multiple realms, with  
   different keys in each, the srealm field in the unencrypted portion  
   of the ticket in the KRB\_AP\_REQ is used to specify which secret key  
   the server should use to decrypt that ticket.  The KRB\_AP\_ERR\_NOKEY

Kohl & Neuman                                                  \[Page 21\]  
  
RFC 1510                        Kerberos                  September 1993

error code is returned if the server doesn't have the proper key to  
   decipher the ticket.

The ticket is decrypted using the version of the server's key  
   specified by the ticket.  If the decryption routines detect a  
   modification of the ticket (each encryption system must provide  
   safeguards to detect modified ciphertext; see section 6), the  
   KRB\_AP\_ERR\_BAD\_INTEGRITY error is returned (chances are good that  
   different keys were used to encrypt and decrypt).

The authenticator is decrypted using the session key extracted from  
   the decrypted ticket.  If decryption shows it to have been modified,  
   the KRB\_AP\_ERR\_BAD\_INTEGRITY error is returned.  The name and realm  
   of the client from the ticket are compared against the same fields in  
   the authenticator.  If they don't match, the KRB\_AP\_ERR\_BADMATCH  
   error is returned (they might not match, for example, if the wrong  
   session key was used to encrypt the authenticator).  The addresses in  
   the ticket (if any) are then searched for an address matching the  
   operating-system reported address of the client.  If no match is  
   found or the server insists on ticket addresses but none are present  
   in the ticket, the KRB\_AP\_ERR\_BADADDR error is returned.

If the local (server) time and the client time in the authenticator  
   differ by more than the allowable clock skew (e.g., 5 minutes), the  
   KRB\_AP\_ERR\_SKEW error is returned.  If the server name, along with  
   the client name, time and microsecond fields from the Authenticator  
   match any recently-seen such tuples, the KRB\_AP\_ERR\_REPEAT error is  
   returned (Note that the rejection here is restricted to  
   authenticators from the same principal to the same server.  Other  
   client principals communicating with the same server principal should  
   not be have their authenticators rejected if the time and microsecond  
   fields happen to match some other client's authenticator.).  The  
   server must remember any authenticator presented within the allowable  
   clock skew, so that a replay attempt is guaranteed to fail. If a  
   server loses track of any authenticator presented within the  
   allowable clock skew, it must reject all requests until the clock  
   skew interval has passed.  This assures that any lost or re-played  
   authenticators will fall outside the allowable clock skew and can no  
   longer be successfully replayed (If this is not done, an attacker  
   could conceivably record the ticket and authenticator sent over the  
   network to a server, then disable the client's host, pose as the  
   disabled host, and replay the ticket and authenticator to subvert the  
   authentication.).  If a sequence number is provided in the  
   authenticator, the server saves it for later use in processing  
   KRB\_SAFE and/or KRB\_PRIV messages.  If a subkey is present, the  
   server either saves it for later use or uses it to help generate its  
   own choice for a subkey to be returned in a KRB\_AP\_REP message.

Kohl & Neuman                                                  \[Page 22\]  
  
RFC 1510                        Kerberos                  September 1993

The server computes the age of the ticket: local (server) time minus  
   the start time inside the Ticket.  If the start time is later than  
   the current time by more than the allowable clock skew or if the  
   INVALID flag is set in the ticket, the KRB\_AP\_ERR\_TKT\_NYV error is  
   returned.  Otherwise, if the current time is later than end time by  
   more than the allowable clock skew, the KRB\_AP\_ERR\_TKT\_EXPIRED error  
   is returned.

If all these checks succeed without an error, the server is assured  
   that the client possesses the credentials of the principal named in  
   the ticket and thus, the client has been authenticated to the server.  
   See section A.10 for pseudocode.

3.2.4. Generation of a KRB\_AP\_REP message

Typically, a client's request will include both the authentication  
   information and its initial request in the same message, and the  
   server need not explicitly reply to the KRB\_AP\_REQ.  However, if  
   mutual authentication (not only authenticating the client to the  
   server, but also the server to the client) is being performed, the  
   KRB\_AP\_REQ message will have MUTUAL-REQUIRED set in its ap-options  
   field, and a KRB\_AP\_REP message is required in response.  As with the  
   error message, this message may be encapsulated in the application  
   protocol if its "raw" form is not acceptable to the application's  
   protocol.  The timestamp and microsecond field used in the reply must  
   be the client's timestamp and microsecond field (as provided in the  
   authenticator). \[Note: In the Kerberos version 4 protocol, the  
   timestamp in the reply was the client's timestamp plus one.  This is  
   not necessary in version 5 because version 5 messages are formatted  
   in such a way that it is not possible to create the reply by  
   judicious message surgery (even in encrypted form) without knowledge  
   of the appropriate encryption keys.\]  If a sequence number is to be  
   included, it should be randomly chosen as described above for the  
   authenticator.  A subkey may be included if the server desires to  
   negotiate a different subkey.  The KRB\_AP\_REP message is encrypted in  
   the session key extracted from the ticket.  See section A.11 for  
   pseudocode.

3.2.5. Receipt of KRB\_AP\_REP message

If a KRB\_AP\_REP message is returned, the client uses the session key  
   from the credentials obtained for the server (Note that for  
   encrypting the KRB\_AP\_REP message, the sub-session key is not used,  
   even if present in the Authenticator.) to decrypt the message, and  
   verifies that the timestamp and microsecond fields match those in the  
   Authenticator it sent to the server.  If they match, then the client  
   is assured that the server is genuine. The sequence number and subkey  
   (if present) are retained for later use.  See section A.12 for

Kohl & Neuman                                                  \[Page 23\]  
  
RFC 1510                        Kerberos                  September 1993

pseudocode.

3.2.6. Using the encryption key

After the KRB\_AP\_REQ/KRB\_AP\_REP exchange has occurred, the client and  
   server share an encryption key which can be used by the application.  
   The "true session key" to be used for KRB\_PRIV, KRB\_SAFE, or other  
   application-specific uses may be chosen by the application based on  
   the subkeys in the KRB\_AP\_REP message and the authenticator  
   (Implementations of the protocol may wish to provide routines to  
   choose subkeys based on session keys and random numbers and to  
   orchestrate a negotiated key to be returned in the KRB\_AP\_REP  
   message.).  In some cases, the use of this session key will be  
   implicit in the protocol; in others the method of use must be chosen  
   from a several alternatives.  We leave the protocol negotiations of  
   how to use the key (e.g., selecting an encryption or checksum type)  
   to the application programmer; the Kerberos protocol does not  
   constrain the implementation options.

With both the one-way and mutual authentication exchanges, the peers  
   should take care not to send sensitive information to each other  
   without proper assurances.  In particular, applications that require  
   privacy or integrity should use the KRB\_AP\_REP or KRB\_ERROR responses  
   from the server to client to assure both client and server of their  
   peer's identity.  If an application protocol requires privacy of its  
   messages, it can use the KRB\_PRIV message (section 3.5). The KRB\_SAFE  
   message (section 3.4) can be used to assure integrity.

3.3.  The Ticket-Granting Service (TGS) Exchange

Summary

Message direction       Message type     Section  
         1. Client to Kerberos   KRB\_TGS\_REQ      5.4.1  
         2. Kerberos to client   KRB\_TGS\_REP or   5.4.2  
                                 KRB\_ERROR        5.9.1

The TGS exchange between a client and the Kerberos Ticket-Granting  
   Server is initiated by a client when it wishes to obtain  
   authentication credentials for a given server (which might be  
   registered in a remote realm), when it wishes to renew or validate an  
   existing ticket, or when it wishes to obtain a proxy ticket.  In the  
   first case, the client must already have acquired a ticket for the  
   Ticket-Granting Service using the AS exchange (the ticket-granting  
   ticket is usually obtained when a client initially authenticates to  
   the system, such as when a user logs in).  The message format for the  
   TGS exchange is almost identical to that for the AS exchange.  The  
   primary difference is that encryption and decryption in the TGS

Kohl & Neuman                                                  \[Page 24\]  
  
RFC 1510                        Kerberos                  September 1993

exchange does not take place under the client's key.  Instead, the  
   session key from the ticket-granting ticket or renewable ticket, or  
   sub-session key from an Authenticator is used.  As is the case for  
   all application servers, expired tickets are not accepted by the TGS,  
   so once a renewable or ticket-granting ticket expires, the client  
   must use a separate exchange to obtain valid tickets.

The TGS exchange consists of two messages: A request (KRB\_TGS\_REQ)  
   from the client to the Kerberos Ticket-Granting Server, and a reply  
   (KRB\_TGS\_REP or KRB\_ERROR).  The KRB\_TGS\_REQ message includes  
   information authenticating the client plus a request for credentials.  
   The authentication information consists of the authentication header  
   (KRB\_AP\_REQ) which includes the client's previously obtained ticket-  
   granting, renewable, or invalid ticket.  In the ticket-granting  
   ticket and proxy cases, the request may include one or more of: a  
   list of network addresses, a collection of typed authorization data  
   to be sealed in the ticket for authorization use by the application  
   server, or additional tickets (the use of which are described later).  
   The TGS reply (KRB\_TGS\_REP) contains the requested credentials,  
   encrypted in the session key from the ticket-granting ticket or  
   renewable ticket, or if present, in the subsession key from the  
   Authenticator (part of the authentication header). The KRB\_ERROR  
   message contains an error code and text explaining what went wrong.  
   The KRB\_ERROR message is not encrypted.  The KRB\_TGS\_REP message  
   contains information which can be used to detect replays, and to  
   associate it with the message to which it replies.  The KRB\_ERROR  
   message also contains information which can be used to associate it  
   with the message to which it replies, but the lack of encryption in  
   the KRB\_ERROR message precludes the ability to detect replays or  
   fabrications of such messages.

3.3.1. Generation of KRB\_TGS\_REQ message

Before sending a request to the ticket-granting service, the client  
   must determine in which realm the application server is registered  
   \[Note: This can be accomplished in several ways.  It might be known  
   beforehand (since the realm is part of the principal identifier), or  
   it might be stored in a nameserver.  Presently, however, this  
   information is obtained from a configuration file.  If the realm to  
   be used is obtained from a nameserver, there is a danger of being  
   spoofed if the nameservice providing the realm name is not  
   authenticated.  This might result in the use of a realm which has  
   been compromised, and would result in an attacker's ability to  
   compromise the authentication of the application server to the  
   client.\].  If the client does not already possess a ticket-granting  
   ticket for the appropriate realm, then one must be obtained.  This is  
   first attempted by requesting a ticket-granting ticket for the  
   destination realm from the local Kerberos server (using the

Kohl & Neuman                                                  \[Page 25\]  
  
RFC 1510                        Kerberos                  September 1993

KRB\_TGS\_REQ message recursively).  The Kerberos server may return a  
   TGT for the desired realm in which case one can proceed.  
   Alternatively, the Kerberos server may return a TGT for a realm which  
   is "closer" to the desired realm (further along the standard  
   hierarchical path), in which case this step must be repeated with a  
   Kerberos server in the realm specified in the returned TGT.  If  
   neither are returned, then the request must be retried with a  
   Kerberos server for a realm higher in the hierarchy.  This request  
   will itself require a ticket-granting ticket for the higher realm  
   which must be obtained by recursively applying these directions.

Once the client obtains a ticket-granting ticket for the appropriate  
   realm, it determines which Kerberos servers serve that realm, and  
   contacts one. The list might be obtained through a configuration file  
   or network service; as long as the secret keys exchanged by realms  
   are kept secret, only denial of service results from a false Kerberos  
   server.

As in the AS exchange, the client may specify a number of options in  
   the KRB\_TGS\_REQ message.  The client prepares the KRB\_TGS\_REQ  
   message, providing an authentication header as an element of the  
   padata field, and including the same fields as used in the KRB\_AS\_REQ  
   message along with several optional fields: the enc-authorization-  
   data field for application server use and additional tickets required  
   by some options.

In preparing the authentication header, the client can select a sub-  
   session key under which the response from the Kerberos server will be  
   encrypted (If the client selects a sub-session key, care must be  
   taken to ensure the randomness of the selected subsession key.  One  
   approach would be to generate a random number and XOR it with the  
   session key from the ticket-granting ticket.). If the sub-session key  
   is not specified, the session key from the ticket-granting ticket  
   will be used.  If the enc-authorization-data is present, it must be  
   encrypted in the sub-session key, if present, from the authenticator  
   portion of the authentication header, or if not present in the  
   session key from the ticket-granting ticket.

Once prepared, the message is sent to a Kerberos server for the  
   destination realm.  See section A.5 for pseudocode.

3.3.2. Receipt of KRB\_TGS\_REQ message

The KRB\_TGS\_REQ message is processed in a manner similar to the  
   KRB\_AS\_REQ message, but there are many additional checks to be  
   performed.  First, the Kerberos server must determine which server  
   the accompanying ticket is for and it must select the appropriate key  
   to decrypt it. For a normal KRB\_TGS\_REQ message, it will be for the

Kohl & Neuman                                                  \[Page 26\]  
  
RFC 1510                        Kerberos                  September 1993

ticket granting service, and the TGS's key will be used.  If the TGT  
   was issued by another realm, then the appropriate inter-realm key  
   must be used.  If the accompanying ticket is not a ticket granting  
   ticket for the current realm, but is for an application server in the  
   current realm, the RENEW, VALIDATE, or PROXY options are specified in  
   the request, and the server for which a ticket is requested is the  
   server named in the accompanying ticket, then the KDC will decrypt  
   the ticket in the authentication header using the key of the server  
   for which it was issued.  If no ticket can be found in the padata  
   field, the KDC\_ERR\_PADATA\_TYPE\_NOSUPP error is returned.

Once the accompanying ticket has been decrypted, the user-supplied  
   checksum in the Authenticator must be verified against the contents  
   of the request, and the message rejected if the checksums do not  
   match (with an error code of KRB\_AP\_ERR\_MODIFIED) or if the checksum  
   is not keyed or not collision-proof (with an error code of  
   KRB\_AP\_ERR\_INAPP\_CKSUM).  If the checksum type is not supported, the  
   KDC\_ERR\_SUMTYPE\_NOSUPP error is returned.  If the authorization-data  
   are present, they are decrypted using the sub-session key from the  
   Authenticator.

If any of the decryptions indicate failed integrity checks, the  
   KRB\_AP\_ERR\_BAD\_INTEGRITY error is returned.

3.3.3. Generation of KRB\_TGS\_REP message

The KRB\_TGS\_REP message shares its format with the KRB\_AS\_REP  
   (KRB\_KDC\_REP), but with its type field set to KRB\_TGS\_REP.  The  
   detailed specification is in section 5.4.2.

The response will include a ticket for the requested server.  The  
   Kerberos database is queried to retrieve the record for the requested  
   server (including the key with which the ticket will be encrypted).  
   If the request is for a ticket granting ticket for a remote realm,  
   and if no key is shared with the requested realm, then the Kerberos  
   server will select the realm "closest" to the requested realm with  
   which it does share a key, and use that realm instead. This is the  
   only case where the response from the KDC will be for a different  
   server than that requested by the client.

By default, the address field, the client's name and realm, the list  
   of transited realms, the time of initial authentication, the  
   expiration time, and the authorization data of the newly-issued  
   ticket will be copied from the ticket-granting ticket (TGT) or  
   renewable ticket.  If the transited field needs to be updated, but  
   the transited type is not supported, the KDC\_ERR\_TRTYPE\_NOSUPP error  
   is returned.

Kohl & Neuman                                                  \[Page 27\]  
  
RFC 1510                        Kerberos                  September 1993

If the request specifies an endtime, then the endtime of the new  
   ticket is set to the minimum of (a) that request, (b) the endtime  
   from the TGT, and (c) the starttime of the TGT plus the minimum of  
   the maximum life for the application server and the maximum life for  
   the local realm (the maximum life for the requesting principal was  
   already applied when the TGT was issued).  If the new ticket is to be  
   a renewal, then the endtime above is replaced by the minimum of (a)  
   the value of the renew\_till field of the ticket and (b) the starttime  
   for the new ticket plus the life (endtimestarttime) of the old  
   ticket.

If the FORWARDED option has been requested, then the resulting ticket  
   will contain the addresses specified by the client.  This option will  
   only be honored if the FORWARDABLE flag is set in the TGT.  The PROXY  
   option is similar; the resulting ticket will contain the addresses  
   specified by the client.  It will be honored only if the PROXIABLE  
   flag in the TGT is set.  The PROXY option will not be honored on  
   requests for additional ticket-granting tickets.

If the requested start time is absent or indicates a time in the  
   past, then the start time of the ticket is set to the authentication  
   server's current time.  If it indicates a time in the future, but the  
   POSTDATED option has not been specified or the MAY-POSTDATE flag is  
   not set in the TGT, then the error KDC\_ERR\_CANNOT\_POSTDATE is  
   returned.  Otherwise, if the ticket-granting ticket has the  
   MAYPOSTDATE flag set, then the resulting ticket will be postdated and  
   the requested starttime is checked against the policy of the local  
   realm. If acceptable, the ticket's start time is set as requested,  
   and the INVALID flag is set.  The postdated ticket must be validated  
   before use by presenting it to the KDC after the starttime has been  
   reached. However, in no case may the starttime, endtime, or renew-  
   till time of a newly-issued postdated ticket extend beyond the  
   renew-till time of the ticket-granting ticket.

If the ENC-TKT-IN-SKEY option has been specified and an additional  
   ticket has been included in the request, the KDC will decrypt the  
   additional ticket using the key for the server to which the  
   additional ticket was issued and verify that it is a ticket-granting  
   ticket.  If the name of the requested server is missing from the  
   request, the name of the client in the additional ticket will be  
   used.  Otherwise the name of the requested server will be compared to  
   the name of the client in the additional ticket and if different, the  
   request will be rejected.  If the request succeeds, the session key  
   from the additional ticket will be used to encrypt the new ticket  
   that is issued instead of using the key of the server for which the  
   new ticket will be used (This allows easy implementation of user-to-  
   user authentication \[6\], which uses ticket-granting ticket session  
   keys in lieu of secret server keys in situations where such secret

Kohl & Neuman                                                  \[Page 28\]  
  
RFC 1510                        Kerberos                  September 1993

keys could be easily compromised.).

If the name of the server in the ticket that is presented to the KDC  
   as part of the authentication header is not that of the ticket-  
   granting server itself, and the server is registered in the realm of  
   the KDC, If the RENEW option is requested, then the KDC will verify  
   that the RENEWABLE flag is set in the ticket and that the renew\_till  
   time is still in the future.  If the VALIDATE option is rqeuested,  
   the KDC will check that the starttime has passed and the INVALID flag  
   is set.  If the PROXY option is requested, then the KDC will check  
   that the PROXIABLE flag is set in the ticket.  If the tests succeed,  
   the KDC will issue the appropriate new ticket.

Whenever a request is made to the ticket-granting server, the  
   presented ticket(s) is(are) checked against a hot-list of tickets  
   which have been canceled.  This hot-list might be implemented by  
   storing a range of issue dates for "suspect tickets"; if a presented  
   ticket had an authtime in that range, it would be rejected.  In this  
   way, a stolen ticket-granting ticket or renewable ticket cannot be  
   used to gain additional tickets (renewals or otherwise) once the  
   theft has been reported.  Any normal ticket obtained before it was  
   reported stolen will still be valid (because they require no  
   interaction with the KDC), but only until their normal expiration  
   time.

The ciphertext part of the response in the KRB\_TGS\_REP message is  
   encrypted in the sub-session key from the Authenticator, if present,  
   or the session key key from the ticket-granting ticket.  It is not  
   encrypted using the client's secret key.  Furthermore, the client's  
   key's expiration date and the key version number fields are left out  
   since these values are stored along with the client's database  
   record, and that record is not needed to satisfy a request based on a  
   ticket-granting ticket.  See section A.6 for pseudocode.

3.3.3.1.  Encoding the transited field

If the identity of the server in the TGT that is presented to the KDC  
   as part of the authentication header is that of the ticket-granting  
   service, but the TGT was issued from another realm, the KDC will look  
   up the inter-realm key shared with that realm and use that key to  
   decrypt the ticket.  If the ticket is valid, then the KDC will honor  
   the request, subject to the constraints outlined above in the section  
   describing the AS exchange.  The realm part of the client's identity  
   will be taken from the ticket-granting ticket.  The name of the realm  
   that issued the ticket-granting ticket will be added to the transited  
   field of the ticket to be issued.  This is accomplished by reading  
   the transited field from the ticket-granting ticket (which is treated  
   as an unordered set of realm names), adding the new realm to the set,

Kohl & Neuman                                                  \[Page 29\]  
  
RFC 1510                        Kerberos                  September 1993

then constructing and writing out its encoded (shorthand) form (this  
   may involve a rearrangement of the existing encoding).

Note that the ticket-granting service does not add the name of its  
   own realm.  Instead, its responsibility is to add the name of the  
   previous realm.  This prevents a malicious Kerberos server from  
   intentionally leaving out its own name (it could, however, omit other  
   realms' names).

The names of neither the local realm nor the principal's realm are to  
   be included in the transited field.  They appear elsewhere in the  
   ticket and both are known to have taken part in authenticating the  
   principal.  Since the endpoints are not included, both local and  
   single-hop inter-realm authentication result in a transited field  
   that is empty.

Because the name of each realm transited  is  added  to this field,  
   it might potentially be very long.  To decrease the length of this  
   field, its contents are encoded.  The initially supported encoding is  
   optimized for the normal case of inter-realm communication: a  
   hierarchical arrangement of realms using either domain or X.500 style  
   realm names. This encoding (called DOMAIN-X500-COMPRESS) is now  
   described.

Realm names in the transited field are separated by a ",".  The ",",  
   "\\", trailing "."s, and leading spaces (" ") are special characters,  
   and if they are part of a realm name, they must be quoted in the  
   transited field by preceding them with a "\\".

A realm name ending with a "." is interpreted as  being prepended to  
   the previous realm.  For example, we can encode traversal of EDU,  
   MIT.EDU,  ATHENA.MIT.EDU,  WASHINGTON.EDU, and CS.WASHINGTON.EDU as:

"EDU,MIT.,ATHENA.,WASHINGTON.EDU,CS.".

Note that if ATHENA.MIT.EDU, or CS.WASHINGTON.EDU were endpoints,  
   that they would not be included in this field, and we would have:

"EDU,MIT.,WASHINGTON.EDU"

A realm name beginning with a "/" is interpreted as being appended to  
   the previous realm (For the purpose of appending, the realm preceding  
   the first listed realm is considered to be the null realm ("")).  If  
   it is to stand by itself, then it should be preceded by a space ("  
   ").  For example, we can encode traversal of /COM/HP/APOLLO, /COM/HP,  
   /COM, and /COM/DEC as:

"/COM,/HP,/APOLLO, /COM/DEC".

Kohl & Neuman                                                  \[Page 30\]  
  
RFC 1510                        Kerberos                  September 1993

Like the example above, if /COM/HP/APOLLO and /COM/DEC are endpoints,  
   they they would not be included in this field, and we would have:

"/COM,/HP"

A null subfield preceding or following a "," indicates that all  
   realms between the previous realm and the next realm have been  
   traversed (For the purpose of interpreting null subfields, the  
   client's realm is considered to precede those in the transited field,  
   and the server's realm is considered to follow them.). Thus, ","  
   means that all realms along the path between the client and the  
   server have been traversed.  ",EDU, /COM," means that that all realms  
   from the client's realm up to EDU (in a domain style hierarchy) have  
   been traversed, and that everything from /COM down to the server's  
   realm in an X.500 style has also been traversed.  This could occur if  
   the EDU realm in one hierarchy shares an inter-realm key directly  
   with the /COM realm in another hierarchy.

3.3.4. Receipt of KRB\_TGS\_REP message

When the KRB\_TGS\_REP is received by the client, it is processed in  
   the same manner as the KRB\_AS\_REP processing described above.  The  
   primary difference is that the ciphertext part of the response must  
   be decrypted using the session key from the ticket-granting ticket  
   rather than the client's secret key.  See section A.7 for pseudocode.

3.4.  The KRB\_SAFE Exchange

The KRB\_SAFE message may be used by clients requiring the ability to  
   detect modifications of messages they exchange.  It achieves this by  
   including a keyed collisionproof checksum of the user data and some  
   control information.  The checksum is keyed with an encryption key  
   (usually the last key negotiated via subkeys, or the session key if  
   no negotiation has occured).

3.4.1. Generation of a KRB\_SAFE message

When an application wishes to send a KRB\_SAFE message, it collects  
   its data and the appropriate control information and computes a  
   checksum over them.  The checksum algorithm should be some sort of  
   keyed one-way hash function (such as the RSA-MD5-DES checksum  
   algorithm specified in section 6.4.5, or the DES MAC), generated  
   using the sub-session key if present, or the session key.  Different  
   algorithms may be selected by changing the checksum type in the  
   message.  Unkeyed or non-collision-proof checksums are not suitable  
   for this use.

The control information for the KRB\_SAFE message includes both a

Kohl & Neuman                                                  \[Page 31\]  
  
RFC 1510                        Kerberos                  September 1993

timestamp and a sequence number.  The designer of an application  
   using the KRB\_SAFE message must choose at least one of the two  
   mechanisms.  This choice should be based on the needs of the  
   application protocol.

Sequence numbers are useful when all messages sent will be received  
   by one's peer.  Connection state is presently required to maintain  
   the session key, so maintaining the next sequence number should not  
   present an additional problem.

If the application protocol is expected to tolerate lost messages  
   without them being resent, the use of the timestamp is the  
   appropriate replay detection mechanism.  Using timestamps is also the  
   appropriate mechanism for multi-cast protocols where all of one's  
   peers share a common sub-session key, but some messages will be sent  
   to a subset of one's peers.

After computing the checksum, the client then transmits the  
   information and checksum to the recipient in the message format  
   specified in section 5.6.1.

3.4.2. Receipt of KRB\_SAFE message

When an application receives a KRB\_SAFE message, it verifies it as  
   follows.  If any error occurs, an error code is reported for use by  
   the application.

The message is first checked by verifying that the protocol version  
   and type fields match the current version and KRB\_SAFE, respectively.  
   A mismatch generates a KRB\_AP\_ERR\_BADVERSION or KRB\_AP\_ERR\_MSG\_TYPE  
   error.  The application verifies that the checksum used is a  
   collisionproof keyed checksum, and if it is not, a  
   KRB\_AP\_ERR\_INAPP\_CKSUM error is generated.  The recipient verifies  
   that the operating system's report of the sender's address matches  
   the sender's address in the message, and (if a recipient address is  
   specified or the recipient requires an address) that one of the  
   recipient's addresses appears as the recipient's address in the  
   message.  A failed match for either case generates a  
   KRB\_AP\_ERR\_BADADDR error.  Then the timestamp and usec and/or the  
   sequence number fields are checked.  If timestamp and usec are  
   expected and not present, or they are present but not current, the  
   KRB\_AP\_ERR\_SKEW error is generated.  If the server name, along with  
   the client name, time and microsecond fields from the Authenticator  
   match any recently-seen such tuples, the KRB\_AP\_ERR\_REPEAT error is  
   generated.  If an incorrect sequence number is included, or a  
   sequence number is expected but not present, the KRB\_AP\_ERR\_BADORDER  
   error is generated.  If neither a timestamp and usec or a sequence  
   number is present, a KRB\_AP\_ERR\_MODIFIED error is generated.

Kohl & Neuman                                                  \[Page 32\]  
  
RFC 1510                        Kerberos                  September 1993

Finally, the checksum is computed over the data and control  
   information, and if it doesn't match the received checksum, a  
   KRB\_AP\_ERR\_MODIFIED error is generated.

If all the checks succeed, the application is assured that the  
   message was generated by its peer and was not modified in transit.

3.5.  The KRB\_PRIV Exchange

The KRB\_PRIV message may be used by clients requiring confidentiality  
   and the ability to detect modifications of exchanged messages.  It  
   achieves this by encrypting the messages and adding control  
   information.

3.5.1. Generation of a KRB\_PRIV message

When an application wishes to send a KRB\_PRIV message, it collects  
   its data and the appropriate control information (specified in  
   section 5.7.1) and encrypts them under an encryption key (usually the  
   last key negotiated via subkeys, or the session key if no negotiation  
   has occured).  As part of the control information, the client must  
   choose to use either a timestamp or a sequence number (or both); see  
   the discussion in section 3.4.1 for guidelines on which to use.  
   After the user data and control information are encrypted, the client  
   transmits the ciphertext and some "envelope" information to the  
   recipient.

3.5.2. Receipt of KRB\_PRIV message

When an application receives a KRB\_PRIV message, it verifies it as  
   follows.  If any error occurs, an error code is reported for use by  
   the application.

The message is first checked by verifying that the protocol version  
   and type fields match the current version and KRB\_PRIV, respectively.  
   A mismatch generates a KRB\_AP\_ERR\_BADVERSION or KRB\_AP\_ERR\_MSG\_TYPE  
   error.  The application then decrypts the ciphertext and processes  
   the resultant plaintext. If decryption shows the data to have been  
   modified, a KRB\_AP\_ERR\_BAD\_INTEGRITY error is generated.  The  
   recipient verifies that the operating system's report of the sender's  
   address matches the sender's address in the message, and (if a  
   recipient address is specified or the recipient requires an address)  
   that one of the recipient's addresses appears as the recipient's  
   address in the message.  A failed match for either case generates a  
   KRB\_AP\_ERR\_BADADDR error.  Then the timestamp and usec and/or the  
   sequence number fields are checked. If timestamp and usec are  
   expected and not present, or they are present but not current, the  
   KRB\_AP\_ERR\_SKEW error is generated.  If the server name, along with

Kohl & Neuman                                                  \[Page 33\]  
  
RFC 1510                        Kerberos                  September 1993

the client name, time and microsecond fields from the Authenticator  
   match any recently-seen such tuples, the KRB\_AP\_ERR\_REPEAT error is  
   generated.  If an incorrect sequence number is included, or a  
   sequence number is expected but not present, the KRB\_AP\_ERR\_BADORDER  
   error is generated.  If neither a timestamp and usec or a sequence  
   number is present, a KRB\_AP\_ERR\_MODIFIED error is generated.

If all the checks succeed, the application can assume the message was  
   generated by its peer, and was securely transmitted (without  
   intruders able to see the unencrypted contents).

3.6.  The KRB\_CRED Exchange

The KRB\_CRED message may be used by clients requiring the ability to  
   send Kerberos credentials from one host to another.  It achieves this  
   by sending the tickets together with encrypted data containing the  
   session keys and other information associated with the tickets.

3.6.1. Generation of a KRB\_CRED message

When an application wishes to send a KRB\_CRED message it first (using  
   the KRB\_TGS exchange) obtains credentials to be sent to the remote  
   host.  It then constructs a KRB\_CRED message using the ticket or  
   tickets so obtained, placing the session key needed to use each  
   ticket in the key field of the corresponding KrbCredInfo sequence of  
   the encrypted part of the the KRB\_CRED message.

Other information associated with each ticket and obtained during the  
   KRB\_TGS exchange is also placed in the corresponding KrbCredInfo  
   sequence in the encrypted part of the KRB\_CRED message.  The current  
   time and, if specifically required by the application the nonce, s-  
   address, and raddress fields, are placed in the encrypted part of the  
   KRB\_CRED message which is then encrypted under an encryption key  
   previosuly exchanged in the KRB\_AP exchange (usually the last key  
   negotiated via subkeys, or the session key if no negotiation has  
   occured).

3.6.2. Receipt of KRB\_CRED message

When an application receives a KRB\_CRED message, it verifies it.  If  
   any error occurs, an error code is reported for use by the  
   application.  The message is verified by checking that the protocol  
   version and type fields match the current version and KRB\_CRED,  
   respectively.  A mismatch generates a KRB\_AP\_ERR\_BADVERSION or  
   KRB\_AP\_ERR\_MSG\_TYPE error.  The application then decrypts the  
   ciphertext and processes the resultant plaintext. If decryption shows  
   the data to have been modified, a KRB\_AP\_ERR\_BAD\_INTEGRITY error is  
   generated.

Kohl & Neuman                                                  \[Page 34\]  
  
RFC 1510                        Kerberos                  September 1993

If present or required, the recipient verifies that the operating  
   system's report of the sender's address matches the sender's address  
   in the message, and that one of the recipient's addresses appears as  
   the recipient's address in the message.  A failed match for either  
   case generates a KRB\_AP\_ERR\_BADADDR error.  The timestamp and usec  
   fields (and the nonce field if required) are checked next.  If the  
   timestamp and usec are not present, or they are present but not  
   current, the KRB\_AP\_ERR\_SKEW error is generated.

If all the checks succeed, the application stores each of the new  
   tickets in its ticket cache together with the session key and other  
   information in the corresponding KrbCredInfo sequence from the  
   encrypted part of the KRB\_CRED message.

4.  The Kerberos Database

The Kerberos server must have access to a database containing the  
   principal identifiers and secret keys of principals to be  
   authenticated (The implementation of the Kerberos server need not  
   combine the database and the server on the same machine; it is  
   feasible to store the principal database in, say, a network name  
   service, as long as the entries stored therein are protected from  
   disclosure to and modification by unauthorized parties.  However, we  
   recommend against such strategies, as they can make system management  
   and threat analysis quite complex.).

4.1.  Database contents

A database entry should contain at least the following fields:

Field                Value

name                 Principal's identifier  
   key                  Principal's secret key  
   p\_kvno               Principal's key version  
   max\_life             Maximum lifetime for Tickets  
   max\_renewable\_life   Maximum total lifetime for renewable  
                        Tickets

The name field is an encoding of the principal's identifier.  The key  
   field contains an encryption key.  This key is the principal's secret  
   key.  (The key can be encrypted before storage under a Kerberos  
   "master key" to protect it in case the database is compromised but  
   the master key is not.  In that case, an extra field must be added to  
   indicate the master key version used, see below.) The p\_kvno field is  
   the key version number of the principal's secret key.  The max\_life  
   field contains the maximum allowable lifetime (endtime - starttime)  
   for any Ticket issued for this principal.  The max\_renewable\_life

Kohl & Neuman                                                  \[Page 35\]  
  
RFC 1510                        Kerberos                  September 1993

field contains the maximum allowable total lifetime for any renewable  
   Ticket issued for this principal.  (See section 3.1 for a description  
   of how these lifetimes are used in determining the lifetime of a  
   given Ticket.)

A server may provide KDC service to several realms, as long as the  
   database representation provides a mechanism to distinguish between  
   principal records with identifiers which differ only in the realm  
   name.

When an application server's key changes, if the change is routine  
   (i.e.,  not the result of disclosure of the old key), the old key  
   should be retained by the server until all tickets that had been  
   issued using that key have expired.  Because of this, it is possible  
   for several keys to be active for a single principal.  Ciphertext  
   encrypted in a principal's key is always tagged with the version of  
   the key that was used for encryption, to help the recipient find the  
   proper key for decryption.

When more than one key is active for a particular principal, the  
   principal will have more than one record in the Kerberos database.  
   The keys and key version numbers will differ between the records (the  
   rest of the fields may or may not be the same). Whenever Kerberos  
   issues a ticket, or responds to a request for initial authentication,  
   the most recent key (known by the Kerberos server) will be used for  
   encryption.  This is the key with the highest key version number.

4.2.  Additional fields

Project Athena's KDC implementation uses additional fields in its  
   database:

Field        Value

K\_kvno       Kerberos' key version  
   expiration   Expiration date for entry  
   attributes   Bit field of attributes  
   mod\_date     Timestamp of last modification  
   mod\_name     Modifying principal's identifier

The K\_kvno field indicates the key version of the Kerberos master key  
   under which the principal's secret key is encrypted.

After an entry's expiration date has passed, the KDC will return an  
   error to any client attempting to gain tickets as or for the  
   principal.  (A database may want to maintain two expiration dates:  
   one for the principal, and one for the principal's current key.  This  
   allows password aging to work independently of the principal's

Kohl & Neuman                                                  \[Page 36\]  
  
RFC 1510                        Kerberos                  September 1993

expiration date.  However, due to the limited space in the responses,  
   the KDC must combine the key expiration and principal expiration date  
   into a single value called "key\_exp", which is used as a hint to the  
   user to take administrative action.)

The attributes field is a bitfield used to govern the operations  
   involving the principal.  This field might be useful in conjunction  
   with user registration procedures, for site-specific policy  
   implementations (Project Athena currently uses it for their user  
   registration process controlled by the system-wide database service,  
   Moira \[7\]), or to identify the "string to key" conversion algorithm  
   used for a principal's key.  (See the discussion of the padata field  
   in section 5.4.2 for details on why this can be useful.)  Other bits  
   are used to indicate that certain ticket options should not be  
   allowed in tickets encrypted under a principal's key (one bit each):  
   Disallow issuing postdated tickets, disallow issuing forwardable  
   tickets, disallow issuing tickets based on TGT authentication,  
   disallow issuing renewable tickets, disallow issuing proxiable  
   tickets, and disallow issuing tickets for which the principal is the  
   server.

The mod\_date field contains the time of last modification of the  
   entry, and the mod\_name field contains the name of the principal  
   which last modified the entry.

4.3.  Frequently Changing Fields

Some KDC implementations may wish to maintain the last time that a  
   request was made by a particular principal.  Information that might  
   be maintained includes the time of the last request, the time of the  
   last request for a ticket-granting ticket, the time of the last use  
   of a ticket-granting ticket, or other times.  This information can  
   then be returned to the user in the last-req field (see section 5.2).

Other frequently changing information that can be maintained is the  
   latest expiration time for any tickets that have been issued using  
   each key.  This field would be used to indicate how long old keys  
   must remain valid to allow the continued use of outstanding tickets.

4.4.  Site Constants

The KDC implementation should have the following configurable  
   constants or options, to allow an administrator to make and enforce  
   policy decisions:

+ The minimum supported lifetime (used to determine whether the  
      KDC\_ERR\_NEVER\_VALID error should be returned). This constant  
      should reflect reasonable expectations of round-trip time to the

Kohl & Neuman                                                  \[Page 37\]  
  
RFC 1510                        Kerberos                  September 1993

KDC, encryption/decryption time, and processing time by the client  
      and target server, and it should allow for a minimum "useful"  
      lifetime.

+ The maximum allowable total (renewable) lifetime of a ticket  
      (renew\_till - starttime).

+ The maximum allowable lifetime of a ticket (endtime - starttime).

+ Whether to allow the issue of tickets with empty address fields  
      (including the ability to specify that such tickets may only be  
      issued if the request specifies some authorization\_data).

+ Whether proxiable, forwardable, renewable or post-datable tickets  
      are to be issued.

5.  Message Specifications

The following sections describe the exact contents and encoding of  
   protocol messages and objects.  The ASN.1 base definitions are  
   presented in the first subsection.  The remaining subsections specify  
   the protocol objects (tickets and authenticators) and messages.  
   Specification of encryption and checksum techniques, and the fields  
   related to them, appear in section 6.

5.1.  ASN.1 Distinguished Encoding Representation

All uses of ASN.1 in Kerberos shall use the Distinguished Encoding  
   Representation of the data elements as described in the X.509  
   specification, section 8.7 \[8\].

5.2.  ASN.1 Base Definitions

The following ASN.1 base definitions are used in the rest of this  
   section. Note that since the underscore character (\_) is not  
   permitted in ASN.1 names, the hyphen (-) is used in its place for the  
   purposes of ASN.1 names.

Realm ::=           GeneralString  
   PrincipalName ::=   SEQUENCE \{  
                       name-type\[0\]     INTEGER,  
                       name-string\[1\]   SEQUENCE OF GeneralString  
   \}

Kerberos realms are encoded as GeneralStrings. Realms shall not  
   contain a character with the code 0 (the ASCII NUL).  Most realms  
   will usually consist of several components separated by periods (.),  
   in the style of Internet Domain Names, or separated by slashes (/) in

Kohl & Neuman                                                  \[Page 38\]  
  
RFC 1510                        Kerberos                  September 1993

the style of X.500 names.  Acceptable forms for realm names are  
   specified in section 7.  A PrincipalName is a typed sequence of  
   components consisting of the following sub-fields:

name-type This field specifies the type of name that follows.  
             Pre-defined values for this field are  
             specified in section 7.2.  The name-type should be  
             treated as a hint.  Ignoring the name type, no two  
             names can be the same (i.e., at least one of the  
             components, or the realm, must be different).  
             This constraint may be eliminated in the future.

name-string This field encodes a sequence of components that  
               form a name, each component encoded as a General  
               String.  Taken together, a PrincipalName and a Realm  
               form a principal identifier.  Most PrincipalNames  
               will have only a few components (typically one or two).

KerberosTime ::=   GeneralizedTime  
                              -- Specifying UTC time zone (Z)

The timestamps used in Kerberos are encoded as GeneralizedTimes.  An  
   encoding shall specify the UTC time zone (Z) and shall not include  
   any fractional portions of the seconds.  It further shall not include  
   any separators.  Example: The only valid format for UTC time 6  
   minutes, 27 seconds after 9 pm on 6 November 1985 is 19851106210627Z.

HostAddress ::=     SEQUENCE  \{  
                        addr-type\[0\]             INTEGER,  
                        address\[1\]               OCTET STRING  
    \}

HostAddresses ::=   SEQUENCE OF SEQUENCE \{  
                        addr-type\[0\]             INTEGER,  
                        address\[1\]               OCTET STRING  
    \}

The host adddress encodings consists of two fields:

addr-type  This field specifies the type of  address that  
              follows. Pre-defined values for this field are  
              specified in section 8.1.

address   This field encodes a single address of type addr-type.

The two forms differ slightly. HostAddress contains exactly one

Kohl & Neuman                                                  \[Page 39\]  
  
RFC 1510                        Kerberos                  September 1993

address; HostAddresses contains a sequence of possibly many  
   addresses.

AuthorizationData ::=   SEQUENCE OF SEQUENCE \{  
                           ad-type\[0\]               INTEGER,  
                           ad-data\[1\]               OCTET STRING  
   \}

ad-data   This field contains authorization data to be  
             interpreted according to the value of the  
             corresponding ad-type field.

ad-type   This field specifies the format for the ad-data  
             subfield.  All negative values are reserved for  
             local use.  Non-negative values are reserved for  
             registered use.

APOptions ::=   BIT STRING \{  
                                   reserved(0),  
                                   use-session-key(1),  
                                   mutual-required(2)  
                   \}

TicketFlags ::=   BIT STRING \{  
                                     reserved(0),  
                                     forwardable(1),  
                                     forwarded(2),  
                                     proxiable(3),  
                                     proxy(4),  
                                     may-postdate(5),  
                                     postdated(6),  
                                     invalid(7),  
                                     renewable(8),  
                                     initial(9),  
                                     pre-authent(10),  
                                     hw-authent(11)  
                   \}

KDCOptions ::=   BIT STRING \{  
                                   reserved(0),  
                                   forwardable(1),  
                                   forwarded(2),  
                                   proxiable(3),  
                                   proxy(4),  
                                   allow-postdate(5),  
                                   postdated(6),

Kohl & Neuman                                                  \[Page 40\]  
  
RFC 1510                        Kerberos                  September 1993

unused7(7),  
                                   renewable(8),  
                                   unused9(9),  
                                   unused10(10),  
                                   unused11(11),  
                                   renewable-ok(27),  
                                   enc-tkt-in-skey(28),  
                                   renew(30),  
                                   validate(31)  
                  \}

LastReq ::=   SEQUENCE OF SEQUENCE \{  
                          lr-type\[0\]               INTEGER,  
                          lr-value\[1\]              KerberosTime  
            \}

lr-type   This field indicates how the following lr-value  
             field is to be interpreted.  Negative values indicate  
             that the information pertains only to the  
             responding server.  Non-negative values pertain to  
             all servers for the realm.

If the lr-type field is zero (0), then no information  
             is conveyed by the lr-value subfield.  If the  
             absolute value of the lr-type field is one (1),  
             then the lr-value subfield is the time of last  
             initial request for a TGT.  If it is two (2), then  
             the lr-value subfield is the time of last initial  
             request.  If it is three (3), then the lr-value  
             subfield is the time of issue for the newest  
             ticket-granting ticket used. If it is four (4),  
             then the lr-value subfield is the time of the last  
             renewal.  If it is five (5), then the lr-value  
             subfield is the time of last request (of any  
             type).

lr-value  This field contains the time of the last request.  
             The time must be interpreted according to the contents  
             of the accompanying lr-type subfield.

See section 6 for the definitions of Checksum, ChecksumType,  
   EncryptedData, EncryptionKey, EncryptionType, and KeyType.

Kohl & Neuman                                                  \[Page 41\]  
  
RFC 1510                        Kerberos                  September 1993

5.3.  Tickets and Authenticators

This section describes the format and encryption parameters for  
   tickets and authenticators.  When a ticket or authenticator is  
   included in a protocol message it is treated as an opaque object.

5.3.1. Tickets

A ticket is a record that helps a client authenticate to a service.  
   A Ticket contains the following information:

Ticket ::=                    \[APPLICATION 1\] SEQUENCE \{  
                              tkt-vno\[0\]                   INTEGER,  
                              realm\[1\]                     Realm,  
                              sname\[2\]                     PrincipalName,  
                              enc-part\[3\]                  EncryptedData  
\}  
\-- Encrypted part of ticket  
EncTicketPart ::=     \[APPLICATION 3\] SEQUENCE \{  
                      flags\[0\]             TicketFlags,  
                      key\[1\]               EncryptionKey,  
                      crealm\[2\]            Realm,  
                      cname\[3\]             PrincipalName,  
                      transited\[4\]         TransitedEncoding,  
                      authtime\[5\]          KerberosTime,  
                      starttime\[6\]         KerberosTime OPTIONAL,  
                      endtime\[7\]           KerberosTime,  
                      renew-till\[8\]        KerberosTime OPTIONAL,  
                      caddr\[9\]             HostAddresses OPTIONAL,  
                      authorization-data\[10\]   AuthorizationData OPTIONAL  
\}  
\-- encoded Transited field  
TransitedEncoding ::=         SEQUENCE \{  
                              tr-type\[0\]  INTEGER, -- must be registered  
                              contents\[1\]          OCTET STRING  
\}

The encoding of EncTicketPart is encrypted in the key shared by  
   Kerberos and the end server (the server's secret key).  See section 6  
   for the format of the ciphertext.

tkt-vno   This field specifies the version number for the ticket  
             format.  This document describes version number 5.

realm     This field specifies the realm that issued a ticket.  It  
             also serves to identify the realm part of the server's  
             principal identifier.  Since a Kerberos server can only  
             issue tickets for servers within its realm, the two will

Kohl & Neuman                                                  \[Page 42\]  
  
RFC 1510                        Kerberos                  September 1993

always be identical.

sname     This field specifies the name part of the server's  
             identity.

enc-part  This field holds the encrypted encoding of the  
             EncTicketPart sequence.

flags     This field indicates which of various options were used or  
             requested when the ticket was issued.  It is a bit-field,  
             where the selected options are indicated by the bit being  
             set (1), and the unselected options and reserved fields  
             being reset (0).  Bit 0 is the most significant bit.  The  
             encoding of the bits is specified in section 5.2.  The  
             flags are described in more detail above in section 2.  The  
             meanings of the flags are:

Bit(s)    Name        Description

0         RESERVED    Reserved for future expansion of this  
                                   field.

1         FORWARDABLE The FORWARDABLE flag is normally only  
                                   interpreted by the TGS, and can be  
                                   ignored by end servers.  When set,  
                                   this flag tells the ticket-granting  
                                   server that it is OK to issue a new  
                                   ticket- granting ticket with a  
                                   different network address based on  
                                   the presented ticket.

2         FORWARDED   When set, this flag indicates that  
                                   the ticket has either been forwarded  
                                   or was issued based on authentication  
                                   involving a forwarded ticket-granting  
                                   ticket.

3         PROXIABLE   The PROXIABLE flag is normally only  
                                   interpreted by the TGS, and can be  
                                   ignored by end servers. The PROXIABLE  
                                   flag has an interpretation identical  
                                   to that of the FORWARDABLE flag,  
                                   except that the PROXIABLE flag tells  
                                   the ticket-granting server that only  
                                   non- ticket-granting tickets may be  
                                   issued with different network  
                                   addresses.

Kohl & Neuman                                                  \[Page 43\]  
  
RFC 1510                        Kerberos                  September 1993

4         PROXY      When set, this flag indicates that a  
                                   ticket is a proxy.

5         MAY-POSTDATE The MAY-POSTDATE flag is normally  
                                   only interpreted by the TGS, and can  
                                   be ignored by end servers.  This flag  
                                   tells the ticket-granting server that  
                                   a post- dated ticket may be issued  
                                   based on this ticket-granting ticket.

6         POSTDATED   This flag indicates that this ticket  
                                   has been postdated.  The end-service  
                                   can check the authtime field to see  
                                   when the original authentication  
                                   occurred.

7         INVALID     This flag indicates that a ticket is  
                                   invalid, and it must be validated by  
                                   the KDC before use.  Application  
                                   servers must reject tickets which  
                                   have this flag set.

8         RENEWABLE   The RENEWABLE flag is normally only  
                                   interpreted by the TGS, and can  
                                   usually be ignored by end servers  
                                   (some particularly careful servers  
                                   may wish to disallow renewable  
                                   tickets).  A renewable ticket can be  
                                   used to obtain a replacement ticket  
                                   that expires at a later date.

9         INITIAL     This flag indicates that this ticket  
                                   was issued using the AS protocol, and  
                                   not issued based on a ticket-granting  
                                   ticket.

10        PRE-AUTHENT This flag indicates that during  
                                   initial authentication, the client  
                                   was authenticated by the KDC before a  
                                   ticket was issued.  The strength of  
                                   the preauthentication method is not  
                                   indicated, but is acceptable to the  
                                   KDC.

11        HW-AUTHENT  This flag indicates that the protocol  
                                   employed for initial authentication  
                                   required the use of hardware expected  
                                   to be possessed solely by the named

Kohl & Neuman                                                  \[Page 44\]  
  
RFC 1510                        Kerberos                  September 1993

client.  The hardware authentication  
                                   method is selected by the KDC and the  
                                   strength of the method is not  
                                   indicated.

12-31     RESERVED    Reserved for future use.

key       This field exists in the ticket and the KDC response and is  
             used to pass the session key from Kerberos to the  
             application server and the client.  The field's encoding is  
             described in section 6.2.

crealm    This field contains the name of the realm in which the  
             client is registered and in which initial authentication  
             took place.

cname     This field contains the name part of the client's principal  
             identifier.

transited This field lists the names of the Kerberos realms that took  
             part in authenticating the user to whom this ticket was  
             issued.  It does not specify the order in which the realms  
             were transited.  See section 3.3.3.1 for details on how  
             this field encodes the traversed realms.

authtime  This field indicates the time of initial authentication for  
             the named principal.  It is the time of issue for the  
             original ticket on which this ticket is based.  It is  
             included in the ticket to provide additional information to  
             the end service, and  to provide  the necessary information  
             for implementation of a \`hot list' service at the KDC.   An  
             end service that is particularly paranoid could refuse to  
             accept tickets for which the initial authentication  
             occurred "too far" in the past.

This field is also returned as part of the response from  
             the KDC.  When returned as part of the response to initial  
             authentication (KRB\_AS\_REP), this is the current time on  
             the Kerberos server (It is NOT recommended that this time  
             value be used to adjust the workstation's clock since the  
             workstation cannot reliably determine that such a  
             KRB\_AS\_REP actually came from the proper KDC in a timely  
             manner.).

starttime This field in the ticket specifies the time after which the  
             ticket is valid.  Together with endtime, this field  
             specifies the life of the ticket.   If it is absent from  
             the ticket, its value should be treated as that of the

Kohl & Neuman                                                  \[Page 45\]  
  
RFC 1510                        Kerberos                  September 1993

authtime field.

endtime   This field contains the time after which the ticket will  
             not be honored (its expiration time).  Note that individual  
             services may place their own limits on the life of a ticket  
             and may reject tickets which have not yet expired.  As  
             such, this is really an upper bound on the expiration time  
             for the ticket.

renew-till This field is only present in tickets that have the  
             RENEWABLE flag set in the flags field.  It indicates the  
             maximum endtime that may be included in a renewal.  It can  
             be thought of as the absolute expiration time for the  
             ticket, including all renewals.

caddr     This field in a ticket contains zero (if omitted) or more  
             (if present) host addresses.  These are the addresses from  
             which the ticket can be used.  If there are no addresses,  
             the ticket can be used from any location.  The decision  
             by the KDC to issue or by the end server to accept zero-  
             address tickets is a policy decision and is left to the  
             Kerberos and end-service administrators; they may refuse to  
             issue or accept such tickets.  The suggested and default  
             policy, however, is that such tickets will only be issued  
             or accepted when additional information that can be used to  
             restrict the use of the ticket is included in the  
             authorization\_data field.  Such a ticket is a capability.

Network addresses are included in the ticket to make it  
             harder for an attacker to use stolen credentials. Because  
             the session key is not sent over the network in cleartext,  
             credentials can't be stolen simply by listening to the  
             network; an attacker has to gain access to the session key  
             (perhaps through operating system security breaches or a  
             careless user's unattended session) to make use of stolen  
             tickets.

It is important to note that the network address from which  
             a connection is received cannot be reliably determined.  
             Even if it could be, an attacker who has compromised the  
             client's workstation could use the credentials from there.  
             Including the network addresses only makes it more  
             difficult, not impossible, for an attacker to walk off with  
             stolen credentials and then use them from a "safe"  
             location.

Kohl & Neuman                                                  \[Page 46\]  
  
RFC 1510                        Kerberos                  September 1993

authorization-data The authorization-data field is used to pass  
             authorization data from the principal on whose behalf a  
             ticket was issued to the application service.  If no  
             authorization data is included, this field will be left  
             out.  The data in this field are specific to the end  
             service.  It is expected that the field will contain the  
             names of service specific objects, and the rights to those  
             objects.  The format for this field is described in section  
             5.2.  Although Kerberos is not concerned with the format of  
             the contents of the subfields, it does carry type  
             information (ad-type).

By using the authorization\_data field, a principal is able  
             to issue a proxy that is valid for a specific purpose.  For  
             example, a client wishing to print a file can obtain a file  
             server proxy to be passed to the print server.  By  
             specifying the name of the file in the authorization\_data  
             field, the file server knows that the print server can only  
             use the client's rights when accessing the particular file  
             to be printed.

It is interesting to note that if one specifies the  
             authorization-data field of a proxy and leaves the host  
             addresses blank, the resulting ticket and session key can  
             be treated as a capability.  See \[9\] for some suggested  
             uses of this field.

The authorization-data field is optional and does not have  
             to be included in a ticket.

5.3.2. Authenticators

An authenticator is a record sent with a ticket to a server to  
   certify the client's knowledge of the encryption key in the ticket,  
   to help the server detect replays, and to help choose a "true session  
   key" to use with the particular session.  The encoding is encrypted  
   in the ticket's session key shared by the client and the server:

\-- Unencrypted authenticator  
Authenticator ::=    \[APPLICATION 2\] SEQUENCE    \{  
               authenticator-vno\[0\]          INTEGER,  
               crealm\[1\]                     Realm,  
               cname\[2\]                      PrincipalName,  
               cksum\[3\]                      Checksum OPTIONAL,  
               cusec\[4\]                      INTEGER,  
               ctime\[5\]                      KerberosTime,  
               subkey\[6\]                     EncryptionKey OPTIONAL,  
               seq-number\[7\]                 INTEGER OPTIONAL,

Kohl & Neuman                                                  \[Page 47\]  
  
RFC 1510                        Kerberos                  September 1993

authorization-data\[8\]         AuthorizationData OPTIONAL  
                     \}

authenticator-vno This field specifies the version number for the  
             format of the authenticator. This document specifies  
             version 5.

crealm and cname These fields are the same as those described for the  
             ticket in section 5.3.1.

cksum     This field contains a checksum of the the application data  
             that accompanies the KRB\_AP\_REQ.

cusec     This field contains the microsecond part of the client's  
             timestamp.  Its value (before encryption) ranges from 0 to  
             999999.  It often appears along with ctime.  The two fields  
             are used together to specify a reasonably accurate  
             timestamp.

ctime     This field contains the current time on the client's host.

subkey    This field contains the client's choice for an encryption  
             key which is to be used to protect this specific  
             application session. Unless an application specifies  
             otherwise, if this field is left out the session key from  
             the ticket will be used.

seq-number This optional field includes the initial sequence number  
             to be used by the KRB\_PRIV or KRB\_SAFE messages when  
             sequence numbers are used to detect replays (It may also be  
             used by application specific messages).  When included in  
             the authenticator this field specifies the initial sequence  
             number for messages from the client to the server.  When  
             included in the AP-REP message, the initial sequence number  
             is that for messages from the server to the client.  When  
             used in KRB\_PRIV or KRB\_SAFE messages, it is incremented by  
             one after each message is sent.

For sequence numbers to adequately support the detection of  
             replays they should be non-repeating, even across  
             connection boundaries. The initial sequence number should  
             be random and uniformly distributed across the full space  
             of possible sequence numbers, so that it cannot be guessed  
             by an attacker and so that it and the successive sequence  
             numbers do not repeat other sequences.

Kohl & Neuman                                                  \[Page 48\]  
  
RFC 1510                        Kerberos                  September 1993

authorization-data This field is the same as described for the ticket  
             in section 5.3.1.  It is optional and will only appear when  
             additional restrictions are to be placed on the use of a  
             ticket, beyond those carried in the ticket itself.

5.4.  Specifications for the AS and TGS exchanges

This section specifies the format of the messages used in exchange  
   between the client and the Kerberos server.  The format of possible  
   error messages appears in section 5.9.1.

5.4.1. KRB\_KDC\_REQ definition

The KRB\_KDC\_REQ message has no type of its own.  Instead, its type is  
   one of KRB\_AS\_REQ or KRB\_TGS\_REQ depending on whether the request is  
   for an initial ticket or an additional ticket.  In either case, the  
   message is sent from the client to the Authentication Server to  
   request credentials for a service.

The message fields are:

AS-REQ ::=         \[APPLICATION 10\] KDC-REQ  
TGS-REQ ::=        \[APPLICATION 12\] KDC-REQ

KDC-REQ ::=        SEQUENCE \{  
           pvno\[1\]               INTEGER,  
           msg-type\[2\]           INTEGER,  
           padata\[3\]             SEQUENCE OF PA-DATA OPTIONAL,  
           req-body\[4\]           KDC-REQ-BODY  
\}

PA-DATA ::=        SEQUENCE \{  
           padata-type\[1\]        INTEGER,  
           padata-value\[2\]       OCTET STRING,  
                         -- might be encoded AP-REQ  
\}

KDC-REQ-BODY ::=   SEQUENCE \{  
            kdc-options\[0\]       KDCOptions,  
            cname\[1\]             PrincipalName OPTIONAL,  
                         -- Used only in AS-REQ  
            realm\[2\]             Realm, -- Server's realm  
                         -- Also client's in AS-REQ  
            sname\[3\]             PrincipalName OPTIONAL,  
            from\[4\]              KerberosTime OPTIONAL,  
            till\[5\]              KerberosTime,  
            rtime\[6\]             KerberosTime OPTIONAL,  
            nonce\[7\]             INTEGER,

Kohl & Neuman                                                  \[Page 49\]  
  
RFC 1510                        Kerberos                  September 1993

etype\[8\]             SEQUENCE OF INTEGER, -- EncryptionType,  
                         -- in preference order  
            addresses\[9\]         HostAddresses OPTIONAL,  
            enc-authorization-data\[10\]   EncryptedData OPTIONAL,  
                         -- Encrypted AuthorizationData encoding  
            additional-tickets\[11\]       SEQUENCE OF Ticket OPTIONAL  
\}

The fields in this message are:

pvno      This field is included in each message, and specifies the  
             protocol version number.  This document specifies protocol  
             version 5.

msg-type  This field indicates the type of a protocol message.  It  
             will almost always be the same as the application  
             identifier associated with a message.  It is included to  
             make the identifier more readily accessible to the  
             application.  For the KDC-REQ message, this type will be  
             KRB\_AS\_REQ or KRB\_TGS\_REQ.

padata    The padata (pre-authentication data) field contains a of  
             authentication information which may be needed before  
             credentials can be issued or decrypted.  In the case of  
             requests for additional tickets (KRB\_TGS\_REQ), this field  
             will include an element with padata-type of PA-TGS-REQ and  
             data of an authentication header (ticket-granting ticket  
             and authenticator). The checksum in the authenticator  
             (which must be collisionproof) is to be computed over the  
             KDC-REQ-BODY encoding.  In most requests for initial  
             authentication (KRB\_AS\_REQ) and most replies (KDC-REP), the  
             padata field will be left out.

This field may also contain information needed by certain  
             extensions to the Kerberos protocol.  For example, it might  
             be used to initially verify the identity of a client before  
             any response is returned.  This is accomplished with a  
             padata field with padata-type equal to PA-ENC-TIMESTAMP and  
             padata-value defined as follows:

padata-type     ::= PA-ENC-TIMESTAMP  
   padata-value    ::= EncryptedData -- PA-ENC-TS-ENC

PA-ENC-TS-ENC   ::= SEQUENCE \{  
           patimestamp\[0\]               KerberosTime, -- client's time  
           pausec\[1\]                    INTEGER OPTIONAL  
   \}

Kohl & Neuman                                                  \[Page 50\]  
  
RFC 1510                        Kerberos                  September 1993

with patimestamp containing the client's time and pausec  
             containing the microseconds which may be omitted if a  
             client will not generate more than one request per second.  
             The ciphertext (padata-value) consists of the PA-ENC-TS-ENC  
             sequence, encrypted using the client's secret key.

The padata field can also contain information needed to  
             help the KDC or the client select the key needed for  
             generating or decrypting the response.  This form of the  
             padata is useful for supporting the use of certain  
             "smartcards" with Kerberos.  The details of such extensions  
             are beyond the scope of this specification.  See \[10\] for  
             additional uses of this field.

padata-type The padata-type element of the padata field indicates the  
             way that the padata-value element is to be interpreted.  
             Negative values of padata-type are reserved for  
             unregistered use; non-negative values are used for a  
             registered interpretation of the element type.

req-body  This field is a placeholder delimiting the extent of the  
             remaining fields.  If a checksum is to be calculated over  
             the request, it is calculated over an encoding of the KDC-  
             REQ-BODY sequence which is enclosed within the req-body  
             field.

kdc-options This field appears in the KRB\_AS\_REQ and KRB\_TGS\_REQ  
             requests to the KDC and indicates the flags that the client  
             wants set on the tickets as well as other information that  
             is to modify the behavior of the KDC. Where appropriate,  
             the name of an option may be the same as the flag that is  
             set by that option.  Although in most case, the bit in the  
             options field will be the same as that in the flags field,  
             this is not guaranteed, so it is not acceptable to simply  
             copy the options field to the flags field.  There are  
             various checks that must be made before honoring an option  
             anyway.

The kdc\_options field is a bit-field, where the selected  
             options are indicated by the bit being set (1), and the  
             unselected options and reserved fields being reset (0).  
             The encoding of the bits is specified in section 5.2.  The  
             options are described in more detail above in section 2.  
             The meanings of the options are:

Kohl & Neuman                                                  \[Page 51\]  
  
RFC 1510                        Kerberos                  September 1993

Bit(s)  Name         Description

0       RESERVED     Reserved for future expansion of this  
                                  field.

1       FORWARDABLE  The FORWARDABLE option indicates that  
                                  the ticket to be issued is to have its  
                                  forwardable flag set.  It may only be  
                                  set on the initial request, or in a  
                                  subsequent request if the ticket-  
                                  granting ticket on which it is based  
                                  is also forwardable.

2       FORWARDED    The FORWARDED option is only specified  
                                  in a request to the ticket-granting  
                                  server and will only be honored if the  
                                  ticket-granting ticket in the request  
                                  has its FORWARDABLE bit set.  This  
                                  option indicates that this is a  
                                  request for forwarding. The  
                                  address(es) of the host from which the  
                                  resulting ticket is to be valid are  
                                  included in the addresses field of the  
                                  request.

3       PROXIABLE    The PROXIABLE option indicates that  
                                  the ticket to be issued is to have its  
                                  proxiable flag set. It may only be set  
                                  on the initial request, or in a  
                                  subsequent request if the ticket-  
                                  granting ticket on which it is based  
                                  is also proxiable.

4       PROXY        The PROXY option indicates that this  
                                  is a request for a proxy.  This option  
                                  will only be honored if the ticket-  
                                  granting ticket in the request has its  
                                  PROXIABLE bit set.  The address(es) of  
                                  the host from which the resulting  
                                  ticket is to be valid are included in  
                                  the addresses field of the request.

5       ALLOW-POSTDATE The ALLOW-POSTDATE option indicates  
                                  that the ticket to be issued is to  
                                  have its MAY-POSTDATE flag set.  It  
                                  may only be set on the initial  
                                  request, or in a subsequent request if

Kohl & Neuman                                                  \[Page 52\]  
  
RFC 1510                        Kerberos                  September 1993

the ticket-granting ticket on which it  
                                  is based also has its MAY-POSTDATE  
                                  flag set.

6       POSTDATED    The POSTDATED option indicates that  
                                  this is a request for a postdated  
                                  ticket.  This option will only be  
                                  honored if the ticket-granting ticket  
                                  on which it is based has its MAY-  
                                  POSTDATE flag set.  The resulting  
                                  ticket will also have its INVALID flag  
                                  set, and that flag may be reset by a  
                                  subsequent request to the KDC after  
                                  the starttime in the ticket has been  
                                  reached.

7       UNUSED       This option is presently unused.

8       RENEWABLE    The RENEWABLE option indicates that  
                                  the ticket to be issued is to have its  
                                  RENEWABLE flag set.  It may only be  
                                  set on the initial request, or when  
                                  the ticket-granting ticket on which  
                                  the request is based is also  
                                  renewable.  If this option is  
                                  requested, then the rtime field in the  
                                  request contains the desired absolute  
                                  expiration time for the ticket.

9-26    RESERVED     Reserved for future use.

27      RENEWABLE-OK The RENEWABLE-OK option indicates that  
                                  a renewable ticket will be acceptable  
                                  if a ticket with the requested life  
                                  cannot otherwise be provided.  If a  
                                  ticket with the requested life cannot  
                                  be provided, then a renewable ticket  
                                  may be issued with a renew-till equal  
                                  to the the requested endtime.  The  
                                  value of the renew-till field may  
                                  still be limited by local limits, or  
                                  limits selected by the individual  
                                  principal or server.

28      ENC-TKT-IN-SKEY This option is used only by the  
                                  ticket-granting service.  The ENC-  
                                  TKT-IN-SKEY option indicates that the  
                                  ticket for the end server is to be

Kohl & Neuman                                                  \[Page 53\]  
  
RFC 1510                        Kerberos                  September 1993

encrypted in the session key from the  
                                  additional ticket-granting ticket  
                                  provided.

29      RESERVED     Reserved for future use.

30      RENEW        This option is used only by the  
                                  ticket-granting service.  The RENEW  
                                  option indicates that the present  
                                  request is for a renewal.  The ticket  
                                  provided is encrypted in the secret  
                                  key for the server on which it is  
                                  valid.  This option will only be  
                                  honored if the ticket to be renewed  
                                  has its RENEWABLE flag set and if the  
                                  time in its renew till field has not  
                                  passed.  The ticket to be renewed is  
                                  passed in the padata field as part of  
                                  the authentication header.

31      VALIDATE     This option is used only by the  
                                  ticket-granting service.  The VALIDATE  
                                  option indicates that the request is  
                                  to validate a postdated ticket.  It  
                                  will only be honored if the ticket  
                                  presented is postdated, presently has  
                                  its INVALID flag set, and would be  
                                  otherwise usable at this time.  A  
                                  ticket cannot be validated before its  
                                  starttime.  The ticket presented for  
                                  validation is encrypted in the key of  
                                  the server for which it is valid and  
                                  is passed in the padata field as part  
                                  of the authentication header.

cname and sname These fields are the same as those described for the  
             ticket in section 5.3.1.  sname may only be absent when the  
             ENC-TKT-IN-SKEY option is specified.  If absent, the name  
             of the server is taken from the name of the client in the  
             ticket passed as additional-tickets.

enc-authorization-data The enc-authorization-data, if present (and it  
             can only be present in the TGS\_REQ form), is an encoding of  
             the desired authorization-data encrypted under the sub-  
             session key if present in the Authenticator, or  
             alternatively from the session key in the ticket-granting  
             ticket, both from the padata field in the KRB\_AP\_REQ.

Kohl & Neuman                                                  \[Page 54\]  
  
RFC 1510                        Kerberos                  September 1993

realm     This field specifies the realm part of the server's  
             principal identifier. In the AS exchange, this is also the  
             realm part of the client's principal identifier.

from      This field is included in the KRB\_AS\_REQ and KRB\_TGS\_REQ  
             ticket requests when the requested ticket is to be  
             postdated.  It specifies the desired start time for the  
             requested ticket.

till      This field contains the expiration date requested by the  
             client in a ticket request.

rtime     This field is the requested renew-till time sent from a  
             client to the KDC in a ticket request.  It is optional.

nonce     This field is part of the KDC request and response.  It it  
             intended to hold a random number generated by the client.  
             If the same number is included in the encrypted response  
             from the KDC, it provides evidence that the response is  
             fresh and has not been replayed by an attacker.  Nonces  
             must never be re-used.  Ideally, it should be gen erated  
             randomly, but if the correct time is known, it may suffice  
             (Note, however, that if the time is used as the nonce, one  
             must make sure that the workstation time is monotonically  
             increasing.  If the time is ever reset backwards, there is  
             a small, but finite, probability that a nonce will be  
             reused.).

etype     This field specifies the desired encryption algorithm to be  
             used in the response.

addresses This field is included in the initial request for tickets,  
             and optionally included in requests for additional tickets  
             from the ticket-granting server.  It specifies the  
             addresses from which the requested ticket is to be valid.  
             Normally it includes the addresses for the client's host.  
             If a proxy is requested, this field will contain other  
             addresses.  The contents of this field are usually copied  
             by the KDC into the caddr field of the resulting ticket.

additional-tickets Additional tickets may be optionally included in a  
             request to the ticket-granting server.  If the ENC-TKT-IN-  
             SKEY option has been specified, then the session key from  
             the additional ticket will be used in place of the server's  
             key to encrypt the new ticket.  If more than one option  
             which requires additional tickets has been specified, then  
             the additional tickets are used in the order specified by  
             the ordering of the options bits (see kdc-options, above).

Kohl & Neuman                                                  \[Page 55\]  
  
RFC 1510                        Kerberos                  September 1993

The application code will be either ten (10) or twelve (12) depending  
   on whether the request is for an initial ticket (AS-REQ) or for an  
   additional ticket (TGS-REQ).

The optional fields (addresses, authorization-data and additional-  
   tickets) are only included if necessary to perform the operation  
   specified in the kdc-options field.

It should be noted that in KRB\_TGS\_REQ, the protocol version number  
   appears twice and two different message types appear: the KRB\_TGS\_REQ  
   message contains these fields as does the authentication header  
   (KRB\_AP\_REQ) that is passed in the padata field.

5.4.2. KRB\_KDC\_REP definition

The KRB\_KDC\_REP message format is used for the reply from the KDC for  
   either an initial (AS) request or a subsequent (TGS) request.  There  
   is no message type for KRB\_KDC\_REP.  Instead, the type will be either  
   KRB\_AS\_REP or KRB\_TGS\_REP.  The key used to encrypt the ciphertext  
   part of the reply depends on the message type.  For KRB\_AS\_REP, the  
   ciphertext is encrypted in the client's secret key, and the client's  
   key version number is included in the key version number for the  
   encrypted data.  For KRB\_TGS\_REP, the ciphertext is encrypted in the  
   sub-session key from the Authenticator, or if absent, the session key  
   from the ticket-granting ticket used in the request.  In that case,  
   no version number will be present in the EncryptedData sequence.

The KRB\_KDC\_REP message contains the following fields:

AS-REP ::=    \[APPLICATION 11\] KDC-REP  
   TGS-REP ::=   \[APPLICATION 13\] KDC-REP

KDC-REP ::=   SEQUENCE \{  
                 pvno\[0\]                    INTEGER,  
                 msg-type\[1\]                INTEGER,  
                 padata\[2\]                  SEQUENCE OF PA-DATA OPTIONAL,  
                 crealm\[3\]                  Realm,  
                 cname\[4\]                   PrincipalName,  
                 ticket\[5\]                  Ticket,  
                 enc-part\[6\]                EncryptedData  
   \}

EncASRepPart ::=    \[APPLICATION 25\[25\]\] EncKDCRepPart  
   EncTGSRepPart ::=   \[APPLICATION 26\] EncKDCRepPart

EncKDCRepPart ::=   SEQUENCE \{  
               key\[0\]                       EncryptionKey,  
               last-req\[1\]                  LastReq,

Kohl & Neuman                                                  \[Page 56\]  
  
RFC 1510                        Kerberos                  September 1993

nonce\[2\]                     INTEGER,  
               key-expiration\[3\]            KerberosTime OPTIONAL,  
               flags\[4\]                     TicketFlags,  
               authtime\[5\]                  KerberosTime,  
               starttime\[6\]                 KerberosTime OPTIONAL,  
               endtime\[7\]                   KerberosTime,  
               renew-till\[8\]                KerberosTime OPTIONAL,  
               srealm\[9\]                    Realm,  
               sname\[10\]                    PrincipalName,  
               caddr\[11\]                    HostAddresses OPTIONAL  
   \}

NOTE: In EncASRepPart, the application code in the encrypted  
         part of a message provides an additional check that  
         the message was decrypted properly.

pvno and msg-type These fields are described above in section 5.4.1.  
             msg-type is either KRB\_AS\_REP or KRB\_TGS\_REP.

padata    This field is described in detail in section 5.4.1.  One  
             possible use for this field is to encode an alternate  
             "mix-in" string to be used with a string-to-key algorithm  
             (such as is described in section 6.3.2). This ability is  
             useful to ease transitions if a realm name needs to change  
             (e.g., when a company is acquired); in such a case all  
             existing password-derived entries in the KDC database would  
             be flagged as needing a special mix-in string until the  
             next password change.

crealm, cname, srealm and sname These fields are the same as those  
             described for the ticket in section 5.3.1.

ticket    The newly-issued ticket, from section 5.3.1.

enc-part  This field is a place holder for the ciphertext and related  
             information that forms the encrypted part of a message.  
             The description of the encrypted part of the message  
             follows each appearance of this field.  The encrypted part  
             is encoded as described in section 6.1.

key       This field is the same as described for the ticket in  
             section 5.3.1.

last-req  This field is returned by the KDC and specifies the time(s)  
             of the last request by a principal.  Depending on what  
             information is available, this might be the last time that  
             a request for a ticket-granting ticket was made, or the  
             last time that a request based on a ticket-granting ticket

Kohl & Neuman                                                  \[Page 57\]  
  
RFC 1510                        Kerberos                  September 1993

was successful.  It also might cover all servers for a  
             realm, or just the particular server. Some implementations  
             may display this information to the user to aid in  
             discovering unauthorized use of one's identity.  It is  
             similar in spirit to the last login time displayed when  
             logging into timesharing systems.

nonce     This field is described above in section 5.4.1.

key-expiration The key-expiration field is part of the response from  
             the KDC and specifies the time that the client's secret key  
             is due to expire.  The expiration might be the result of  
             password aging or an account expiration.  This field will  
             usually be left out of the TGS reply since the response to  
             the TGS request is encrypted in a session key and no client  
             information need be retrieved from the KDC database.  It is  
             up to the application client (usually the login program) to  
             take appropriate action (such as notifying the user) if the  
             expira    tion time is imminent.

flags, authtime, starttime, endtime, renew-till and caddr These  
             fields are duplicates of those found in the encrypted  
             portion of the attached ticket (see section 5.3.1),  
             provided so the client may verify they match the intended  
             request and to assist in proper ticket caching.  If the  
             message is of type KRB\_TGS\_REP, the caddr field will only  
             be filled in if the request was for a proxy or forwarded  
             ticket, or if the user is substituting a subset of the  
             addresses from the ticket granting ticket.  If the client-  
             requested addresses are not present or not used, then the  
             addresses contained in the ticket will be the same as those  
             included in the ticket-granting ticket.

5.5.  Client/Server (CS) message specifications

This section specifies the format of the messages used for the  
   authentication of the client to the application server.

5.5.1. KRB\_AP\_REQ definition

The KRB\_AP\_REQ message contains the Kerberos protocol version number,  
   the message type KRB\_AP\_REQ, an options field to indicate any options  
   in use, and the ticket and authenticator themselves.  The KRB\_AP\_REQ  
   message is often referred to as the "authentication header".

AP-REQ ::=      \[APPLICATION 14\] SEQUENCE \{  
                   pvno\[0\]                       INTEGER,  
                   msg-type\[1\]                   INTEGER,

Kohl & Neuman                                                  \[Page 58\]  
  
RFC 1510                        Kerberos                  September 1993

ap-options\[2\]                 APOptions,  
                   ticket\[3\]                     Ticket,  
                   authenticator\[4\]              EncryptedData  
   \}

pvno and msg-type These fields are described above in section 5.4.1.  
             msg-type is KRB\_AP\_REQ.

ap-options This field appears in the application request (KRB\_AP\_REQ)  
             and affects the way the request is processed.  It is a  
             bit-field, where the selected options are indicated by the  
             bit being set (1), and the unselected options and reserved  
             fields being reset (0).  The encoding of the bits is  
             specified in section 5.2.  The meanings of the options are:

Bit(s)  Name           Description

0       RESERVED       Reserved for future expansion of  
                                  this field.

1       USE-SESSION-KEYThe USE-SESSION-KEY option indicates  
                                  that the ticket the client is  
                                  presenting to a server is encrypted in  
                                  the session key from the server's  
                                  ticket-granting ticket. When this  
                                  option is not specified, the ticket is  
                                  encrypted in the server's secret key.

2       MUTUAL-REQUIREDThe MUTUAL-REQUIRED option tells the  
                                  server that the client requires mutual  
                                  authentication, and that it must  
                                  respond with a KRB\_AP\_REP message.

3-31    RESERVED       Reserved for future use.

ticket    This field is a ticket authenticating the client to the  
             server.

authenticator This contains the authenticator, which includes the  
             client's choice of a subkey.  Its encoding is described in  
             section 5.3.2.

Kohl & Neuman                                                  \[Page 59\]  
  
RFC 1510                        Kerberos                  September 1993

5.5.2.  KRB\_AP\_REP definition

The KRB\_AP\_REP message contains the Kerberos protocol version number,  
   the message type, and an encrypted timestamp. The message is sent in  
   in response to an application request (KRB\_AP\_REQ) where the mutual  
   authentication option has been selected in the ap-options field.

AP-REP ::=         \[APPLICATION 15\] SEQUENCE \{  
              pvno\[0\]                   INTEGER,  
              msg-type\[1\]               INTEGER,  
              enc-part\[2\]               EncryptedData  
   \}

EncAPRepPart ::=   \[APPLICATION 27\]     SEQUENCE \{  
              ctime\[0\]                  KerberosTime,  
              cusec\[1\]                  INTEGER,  
              subkey\[2\]                 EncryptionKey OPTIONAL,  
              seq-number\[3\]             INTEGER OPTIONAL  
   \}

NOTE: in EncAPRepPart, the application code in the encrypted part of  
   a message provides an additional check that the message was decrypted  
   properly.

The encoded EncAPRepPart is encrypted in the shared session key of  
   the ticket.  The optional subkey field can be used in an  
   application-arranged negotiation to choose a per association session  
   key.

pvno and msg-type These fields are described above in section 5.4.1.  
             msg-type is KRB\_AP\_REP.

enc-part  This field is described above in section 5.4.2.

ctime     This field contains the current time on the client's host.

cusec     This field contains the microsecond part of the client's  
             timestamp.

subkey    This field contains an encryption key which is to be used  
             to protect this specific application session.  See section  
             3.2.6 for specifics on how this field is used to negotiate  
             a key.  Unless an application specifies otherwise, if this  
             field is left out, the sub-session key from the  
             authenticator, or if also left out, the session key from  
             the ticket will be used.

Kohl & Neuman                                                  \[Page 60\]  
  
RFC 1510                        Kerberos                  September 1993

5.5.3. Error message reply

If an error occurs while processing the application request, the  
   KRB\_ERROR message will be sent in response.  See section 5.9.1 for  
   the format of the error message.  The cname and crealm fields may be  
   left out if the server cannot determine their appropriate values from  
   the corresponding KRB\_AP\_REQ message.  If the authenticator was  
   decipherable, the ctime and cusec fields will contain the values from  
   it.

5.6.  KRB\_SAFE message specification

This section specifies the format of a message that can be used by  
   either side (client or server) of an application to send a tamper-  
   proof message to its peer. It presumes that a session key has  
   previously been exchanged (for example, by using the  
   KRB\_AP\_REQ/KRB\_AP\_REP messages).

5.6.1. KRB\_SAFE definition

The KRB\_SAFE message contains user data along with a collision-proof  
   checksum keyed with the session key.  The message fields are:

KRB-SAFE ::=        \[APPLICATION 20\] SEQUENCE \{  
               pvno\[0\]               INTEGER,  
               msg-type\[1\]           INTEGER,  
               safe-body\[2\]          KRB-SAFE-BODY,  
               cksum\[3\]              Checksum  
   \}

KRB-SAFE-BODY ::=   SEQUENCE \{  
               user-data\[0\]          OCTET STRING,  
               timestamp\[1\]          KerberosTime OPTIONAL,  
               usec\[2\]               INTEGER OPTIONAL,  
               seq-number\[3\]         INTEGER OPTIONAL,  
               s-address\[4\]          HostAddress,  
               r-address\[5\]          HostAddress OPTIONAL  
   \}

pvno and msg-type These fields are described above in section 5.4.1.  
             msg-type is KRB\_SAFE.

safe-body This field is a placeholder for the body of the KRB-SAFE  
             message.  It is to be encoded separately and then have the  
             checksum computed over it, for use in the cksum field.

cksum     This field contains the checksum of the application data.  
             Checksum details are described in section 6.4.  The

Kohl & Neuman                                                  \[Page 61\]  
  
RFC 1510                        Kerberos                  September 1993

checksum is computed over the encoding of the KRB-SAFE-BODY  
             sequence.

user-data This field is part of the KRB\_SAFE and KRB\_PRIV messages  
             and contain the application specific data that is being  
             passed from the sender to the recipient.

timestamp This field is part of the KRB\_SAFE and KRB\_PRIV messages.  
             Its contents are the current time as known by the sender of  
             the message. By checking the timestamp, the recipient of  
             the message is able to make sure that it was recently  
             generated, and is not a replay.

usec      This field is part of the KRB\_SAFE and KRB\_PRIV headers.  
             It contains the microsecond part of the timestamp.

seq-number This field is described above in section 5.3.2.

s-address This field specifies the address in use by the sender of  
             the message.

r-address This field specifies the address in use by the recipient of  
             the message.  It may be omitted for some uses (such as  
             broadcast protocols), but the recipient may arbitrarily  
             reject such messages.  This field along with s-address can  
             be used to help detect messages which have been incorrectly  
             or maliciously delivered to the wrong recipient.

5.7.  KRB\_PRIV message specification

This section specifies the format of a message that can be used by  
   either side (client or server) of an application to securely and  
   privately send a message to its peer.  It presumes that a session key  
   has previously been exchanged (for example, by using the  
   KRB\_AP\_REQ/KRB\_AP\_REP messages).

5.7.1. KRB\_PRIV definition

The KRB\_PRIV message contains user data encrypted in the Session Key.  
   The message fields are:

KRB-PRIV ::=         \[APPLICATION 21\] SEQUENCE \{  
                pvno\[0\]                   INTEGER,  
                msg-type\[1\]               INTEGER,  
                enc-part\[3\]               EncryptedData  
   \}

Kohl & Neuman                                                  \[Page 62\]  
  
RFC 1510                        Kerberos                  September 1993

EncKrbPrivPart ::=   \[APPLICATION 28\] SEQUENCE \{  
                user-data\[0\]              OCTET STRING,  
                timestamp\[1\]              KerberosTime OPTIONAL,  
                usec\[2\]                   INTEGER OPTIONAL,  
                seq-number\[3\]             INTEGER OPTIONAL,  
                s-address\[4\]              HostAddress, -- sender's addr  
                r-address\[5\]              HostAddress OPTIONAL  
                                                      -- recip's addr  
   \}

NOTE: In EncKrbPrivPart, the application code in the encrypted part  
   of a message provides an additional check that the message was  
   decrypted properly.

pvno and msg-type These fields are described above in section 5.4.1.  
             msg-type is KRB\_PRIV.

enc-part  This field holds an encoding of the EncKrbPrivPart sequence  
             encrypted under the session key (If supported by the  
             encryption method in use, an initialization vector may be  
             passed to the encryption procedure, in order to achieve  
             proper cipher chaining.  The initialization vector might  
             come from the last block of the ciphertext from the  
             previous KRB\_PRIV message, but it is the application's  
             choice whether or not to use such an initialization vector.  
             If left out, the default initialization vector for the  
             encryption algorithm will be used.).  This encrypted  
             encoding is used for the enc-part field of the KRB-PRIV  
             message.  See section 6 for the format of the ciphertext.

user-data, timestamp, usec, s-address and r-address These fields are  
             described above in section 5.6.1.

seq-number This field is described above in section 5.3.2.

5.8.  KRB\_CRED message specification

This section specifies the format of a message that can be used to  
   send Kerberos credentials from one principal to another.  It is  
   presented here to encourage a common mechanism to be used by  
   applications when forwarding tickets or providing proxies to  
   subordinate servers.  It presumes that a session key has already been  
   exchanged perhaps by using the KRB\_AP\_REQ/KRB\_AP\_REP messages.

5.8.1. KRB\_CRED definition

The KRB\_CRED message contains a sequence of tickets to be sent and  
   information needed to use the tickets, including the session key from

Kohl & Neuman                                                  \[Page 63\]  
  
RFC 1510                        Kerberos                  September 1993

each.  The information needed to use the tickets is encryped under an  
   encryption key previously exchanged.  The message fields are:

KRB-CRED         ::= \[APPLICATION 22\]   SEQUENCE \{  
                    pvno\[0\]                INTEGER,  
                    msg-type\[1\]            INTEGER, -- KRB\_CRED  
                    tickets\[2\]             SEQUENCE OF Ticket,  
                    enc-part\[3\]            EncryptedData  
   \}

EncKrbCredPart   ::= \[APPLICATION 29\]   SEQUENCE \{  
                    ticket-info\[0\]         SEQUENCE OF KrbCredInfo,  
                    nonce\[1\]               INTEGER OPTIONAL,  
                    timestamp\[2\]           KerberosTime OPTIONAL,  
                    usec\[3\]                INTEGER OPTIONAL,  
                    s-address\[4\]           HostAddress OPTIONAL,  
                    r-address\[5\]           HostAddress OPTIONAL  
   \}

KrbCredInfo      ::=                    SEQUENCE \{  
                    key\[0\]                 EncryptionKey,  
                    prealm\[1\]              Realm OPTIONAL,  
                    pname\[2\]               PrincipalName OPTIONAL,  
                    flags\[3\]               TicketFlags OPTIONAL,  
                    authtime\[4\]            KerberosTime OPTIONAL,  
                    starttime\[5\]           KerberosTime OPTIONAL,  
                    endtime\[6\]             KerberosTime OPTIONAL  
                    renew-till\[7\]          KerberosTime OPTIONAL,  
                    srealm\[8\]              Realm OPTIONAL,  
                    sname\[9\]               PrincipalName OPTIONAL,  
                    caddr\[10\]              HostAddresses OPTIONAL  
   \}

pvno and msg-type These fields are described above in section 5.4.1.  
             msg-type is KRB\_CRED.

tickets  
               These are the tickets obtained from the KDC specifically  
             for use by the intended recipient.  Successive tickets are  
             paired with the corresponding KrbCredInfo sequence from the  
             enc-part of the KRB-CRED message.

enc-part  This field holds an encoding of the EncKrbCredPart sequence  
             encrypted under the session key shared between the sender  
             and the intended recipient.  This encrypted encoding is  
             used for the enc-part field of the KRB-CRED message.  See  
             section 6 for the format of the ciphertext.

Kohl & Neuman                                                  \[Page 64\]  
  
RFC 1510                        Kerberos                  September 1993

nonce     If practical, an application may require the inclusion of a  
             nonce generated by the recipient of the message. If the  
             same value is included as the nonce in the message, it  
             provides evidence that the message is fresh and has not  
             been replayed by an attacker.  A nonce must never be re-  
             used; it should be generated randomly by the recipient of  
             the message and provided to the sender of the mes  sage in  
             an application specific manner.

timestamp and usec These fields specify the time that the KRB-CRED  
             message was generated.  The time is used to provide  
             assurance that the message is fresh.

s-address and r-address These fields are described above in section  
             5.6.1.  They are used optionally to provide additional  
             assurance of the integrity of the KRB-CRED message.

key       This field exists in the corresponding ticket passed by the  
             KRB-CRED message and is used to pass the session key from  
             the sender to the intended recipient.  The field's encoding  
             is described in section 6.2.

The following fields are optional.   If present, they can be  
   associated with the credentials in the remote ticket file.  If left  
   out, then it is assumed that the recipient of the credentials already  
   knows their value.

prealm and pname The name and realm of the delegated principal  
             identity.

flags, authtime,  starttime,  endtime, renew-till,  srealm, sname,  
             and caddr These fields contain the values of the  
             corresponding fields from the ticket found in the ticket  
             field.  Descriptions of the fields are identical to the  
             descriptions in the KDC-REP message.

5.9.  Error message specification

This section specifies the format for the KRB\_ERROR message.  The  
   fields included in the message are intended to return as much  
   information as possible about an error.  It is not expected that all  
   the information required by the fields will be available for all  
   types of errors.  If the appropriate information is not available  
   when the message is composed, the corresponding field will be left  
   out of the message.

Note that since the KRB\_ERROR message is not protected by any  
   encryption, it is quite possible for an intruder to synthesize or

Kohl & Neuman                                                  \[Page 65\]  
  
RFC 1510                        Kerberos                  September 1993

modify such a message.  In particular, this means that the client  
   should not use any fields in this message for security-critical  
   purposes, such as setting a system clock or generating a fresh  
   authenticator.  The message can be useful, however, for advising a  
   user on the reason for some failure.

5.9.1. KRB\_ERROR definition

The KRB\_ERROR message consists of the following fields:

KRB-ERROR ::=   \[APPLICATION 30\] SEQUENCE \{  
                   pvno\[0\]               INTEGER,  
                   msg-type\[1\]           INTEGER,  
                   ctime\[2\]              KerberosTime OPTIONAL,  
                   cusec\[3\]              INTEGER OPTIONAL,  
                   stime\[4\]              KerberosTime,  
                   susec\[5\]              INTEGER,  
                   error-code\[6\]         INTEGER,  
                   crealm\[7\]             Realm OPTIONAL,  
                   cname\[8\]              PrincipalName OPTIONAL,  
                   realm\[9\]              Realm, -- Correct realm  
                   sname\[10\]             PrincipalName, -- Correct name  
                   e-text\[11\]            GeneralString OPTIONAL,  
                   e-data\[12\]            OCTET STRING OPTIONAL  
   \}

pvno and msg-type These fields are described above in section 5.4.1.  
             msg-type is KRB\_ERROR.

ctime     This field is described above in section 5.4.1.

cusec     This field is described above in section 5.5.2.

stime     This field contains the current time on the server.  It is  
             of type KerberosTime.

susec     This field contains the microsecond part of the server's  
             timestamp.  Its value ranges from 0 to 999. It appears  
             along with stime. The two fields are used in conjunction to  
             specify a reasonably accurate timestamp.

error-code This field contains the error code returned by Kerberos or  
             the server when a request fails.  To interpret the value of  
             this field see the list of error codes in section 8.  
             Implementations are encouraged to provide for national  
             language support in the display of error messages.

crealm, cname, srealm and sname These fields are described above in

Kohl & Neuman                                                  \[Page 66\]  
  
RFC 1510                        Kerberos                  September 1993

section 5.3.1.

e-text    This field contains additional text to help explain the  
             error code associated with the failed request (for example,  
             it might include a principal name which was unknown).

e-data    This field contains additional data about the error for use  
             by the application to help it recover from or handle the  
             error.  If the errorcode is KDC\_ERR\_PREAUTH\_REQUIRED, then  
             the e-data field will contain an encoding of a sequence of  
             padata fields, each corresponding to an acceptable pre-  
             authentication method and optionally containing data for  
             the method:

METHOD-DATA ::=    SEQUENCE of PA-DATA

If the error-code is KRB\_AP\_ERR\_METHOD, then the e-data field will  
   contain an encoding of the following sequence:

METHOD-DATA ::=    SEQUENCE \{  
                         method-type\[0\]   INTEGER,  
                         method-data\[1\]   OCTET STRING OPTIONAL  
       \}

method-type will indicate the required alternate method; method-data  
   will contain any required additional information.

6.  Encryption and Checksum Specifications

The Kerberos protocols described in this document are designed to use  
   stream encryption ciphers, which can be simulated using commonly  
   available block encryption ciphers, such as the Data Encryption  
   Standard \[11\], in conjunction with block chaining and checksum  
   methods \[12\].  Encryption is used to prove the identities of the  
   network entities participating in message exchanges.  The Key  
   Distribution Center for each realm is trusted by all principals  
   registered in that realm to store a secret key in confidence.  Proof  
   of knowledge of this secret key is used to verify the authenticity of  
   a principal.

The KDC uses the principal's secret key (in the AS exchange) or a  
   shared session key (in the TGS exchange) to encrypt responses to  
   ticket requests; the ability to obtain the secret key or session key  
   implies the knowledge of the appropriate keys and the identity of the  
   KDC. The ability of a principal to decrypt the KDC response and  
   present a Ticket and a properly formed Authenticator (generated with  
   the session key from the KDC response) to a service verifies the  
   identity of the principal; likewise the ability of the service to

Kohl & Neuman                                                  \[Page 67\]  
  
RFC 1510                        Kerberos                  September 1993

extract the session key from the Ticket and prove its knowledge  
   thereof in a response verifies the identity of the service.

The Kerberos protocols generally assume that the encryption used is  
   secure from cryptanalysis; however, in some cases, the order of  
   fields in the encrypted portions of messages are arranged to minimize  
   the effects of poorly chosen keys.  It is still important to choose  
   good keys.  If keys are derived from user-typed passwords, those  
   passwords need to be well chosen to make brute force attacks more  
   difficult.  Poorly chosen keys still make easy targets for intruders.

The following sections specify the encryption and checksum mechanisms  
   currently defined for Kerberos.  The encodings, chaining, and padding  
   requirements for each are described.  For encryption methods, it is  
   often desirable to place random information (often referred to as a  
   confounder) at the start of the message.  The requirements for a  
   confounder are specified with each encryption mechanism.

Some encryption systems use a block-chaining method to improve the  
   the security characteristics of the ciphertext.  However, these  
   chaining methods often don't provide an integrity check upon  
   decryption.  Such systems (such as DES in CBC mode) must be augmented  
   with a checksum of the plaintext which can be verified at decryption  
   and used to detect any tampering or damage.  Such checksums should be  
   good at detecting burst errors in the input.  If any damage is  
   detected, the decryption routine is expected to return an error  
   indicating the failure of an integrity check. Each encryption type is  
   expected to provide and verify an appropriate checksum. The  
   specification of each encryption method sets out its checksum  
   requirements.

Finally, where a key is to be derived from a user's password, an  
   algorithm for converting the password to a key of the appropriate  
   type is included.  It is desirable for the string to key function to  
   be one-way, and for the mapping to be different in different realms.  
   This is important because users who are registered in more than one  
   realm will often use the same password in each, and it is desirable  
   that an attacker compromising the Kerberos server in one realm not  
   obtain or derive the user's key in another.

For a discussion of the integrity characteristics of the candidate  
   encryption and checksum methods considered for Kerberos, the the  
   reader is referred to \[13\].

6.1.  Encryption Specifications

The following ASN.1 definition describes all encrypted messages.  The  
   enc-part field which appears in the unencrypted part of messages in

Kohl & Neuman                                                  \[Page 68\]  
  
RFC 1510                        Kerberos                  September 1993

section 5 is a sequence consisting of an encryption type, an optional  
   key version number, and the ciphertext.

EncryptedData ::=   SEQUENCE \{  
                       etype\[0\]     INTEGER, -- EncryptionType  
                       kvno\[1\]      INTEGER OPTIONAL,  
                       cipher\[2\]    OCTET STRING -- ciphertext  
   \}

etype     This field identifies which encryption algorithm was used  
             to encipher the cipher.  Detailed specifications for  
             selected encryption types appear later in this section.

kvno      This field contains the version number of the key under  
             which data is encrypted.  It is only present in messages  
             encrypted under long lasting keys, such as principals'  
             secret keys.

cipher    This field contains the enciphered text, encoded as an  
             OCTET STRING.

The cipher field is generated by applying the specified encryption  
   algorithm to data composed of the message and algorithm-specific  
   inputs.  Encryption mechanisms defined for use with Kerberos must  
   take sufficient measures to guarantee the integrity of the plaintext,  
   and we recommend they also take measures to protect against  
   precomputed dictionary attacks.  If the encryption algorithm is not  
   itself capable of doing so, the protections can often be enhanced by  
   adding a checksum and a confounder.

The suggested format for the data to be encrypted includes a  
   confounder, a checksum, the encoded plaintext, and any necessary  
   padding.  The msg-seq field contains the part of the protocol message  
   described in section 5 which is to be encrypted.  The confounder,  
   checksum, and padding are all untagged and untyped, and their length  
   is exactly sufficient to hold the appropriate item.  The type and  
   length is implicit and specified by the particular encryption type  
   being used (etype).  The format for the data to be encrypted is  
   described in the following diagram:

+-----------+----------+-------------+-----+  
         |confounder |   check  |   msg-seq   | pad |  
         +-----------+----------+-------------+-----+

The format cannot be described in ASN.1, but for those who prefer an  
   ASN.1-like notation:

Kohl & Neuman                                                  \[Page 69\]  
  
RFC 1510                        Kerberos                  September 1993

CipherText ::=   ENCRYPTED       SEQUENCE \{  
         confounder\[0\]   UNTAGGED OCTET STRING(conf\_length)     OPTIONAL,  
         check\[1\]        UNTAGGED OCTET STRING(checksum\_length) OPTIONAL,  
         msg-seq\[2\]      MsgSequence,  
         pad             UNTAGGED OCTET STRING(pad\_length) OPTIONAL  
\}

In the above specification, UNTAGGED OCTET STRING(length) is the  
   notation for an octet string with its tag and length removed.  It is  
   not a valid ASN.1 type.  The tag bits and length must be removed from  
   the confounder since the purpose of the confounder is so that the  
   message starts with random data, but the tag and its length are  
   fixed.  For other fields, the length and tag would be redundant if  
   they were included because they are specified by the encryption type.

One generates a random confounder of the appropriate length, placing  
   it in confounder; zeroes out check; calculates the appropriate  
   checksum over confounder, check, and msg-seq, placing the result in  
   check; adds the necessary padding; then encrypts using the specified  
   encryption type and the appropriate key.

Unless otherwise specified, a definition of an encryption algorithm  
   that specifies a checksum, a length for the confounder field, or an  
   octet boundary for padding uses this ciphertext format (The ordering  
   of the fields in the CipherText is important.  Additionally, messages  
   encoded in this format must include a length as part of the msg-seq  
   field.  This allows the recipient to verify that the message has not  
   been truncated.  Without a length, an attacker could use a chosen  
   plaintext attack to generate a message which could be truncated,  
   while leaving the checksum intact.  Note that if the msg-seq is an  
   encoding of an ASN.1 SEQUENCE or OCTET STRING, then the length is  
   part of that encoding.). Those fields which are not specified will be  
   omitted.

In the interest of allowing all implementations using a particular  
   encryption type to communicate with all others using that type, the  
   specification of an encryption type defines any checksum that is  
   needed as part of the encryption process.  If an alternative checksum  
   is to be used, a new encryption type must be defined.

Some cryptosystems require additional information beyond the key and  
   the data to be encrypted. For example, DES, when used in cipher-  
   block-chaining mode, requires an initialization vector.  If required,  
   the description for each encryption type must specify the source of  
   such additional information.

Kohl & Neuman                                                  \[Page 70\]  
  
RFC 1510                        Kerberos                  September 1993

6.2.  Encryption Keys

The sequence below shows the encoding of an encryption key:

EncryptionKey ::=   SEQUENCE \{  
                              keytype\[0\]    INTEGER,  
                              keyvalue\[1\]   OCTET STRING  
          \}

keytype   This field specifies the type of encryption key that  
             follows in the keyvalue field.  It will almost always  
             correspond to the encryption algorithm used to generate the  
             EncryptedData, though more than one algorithm may use the  
             same type of key (the mapping is many to one).  This might  
             happen, for example, if the encryption algorithm uses an  
             alternate checksum algorithm for an integrity check, or a  
             different chaining mechanism.

keyvalue  This field contains the key itself, encoded as an octet  
             string.

All negative values for the  encryption key type are reserved for  
   local use.  All non-negative values are reserved for officially  
   assigned type fields and interpretations.

6.3.  Encryption Systems

6.3.1. The NULL Encryption System (null)

If no encryption is in use, the encryption system is said to be the  
   NULL encryption system.  In the NULL encryption system there is no  
   checksum, confounder or padding.  The ciphertext is simply the  
   plaintext.  The NULL Key is used by the null encryption system and is  
   zero octets in length, with keytype zero (0).

6.3.2. DES in CBC mode with a CRC-32 checksum (des-cbc-crc)

The des-cbc-crc encryption mode encrypts information under the Data  
   Encryption Standard \[11\] using the cipher block chaining mode \[12\].  
   A CRC-32 checksum (described in ISO 3309 \[14\]) is applied to the  
   confounder and message sequence (msg-seq) and placed in the cksum  
   field.  DES blocks are 8 bytes.  As a result, the data to be  
   encrypted (the concatenation of confounder, checksum, and message)  
   must be padded to an 8 byte boundary before encryption.  The details  
   of the encryption of this data are identical to those for the des-  
   cbc-md5 encryption mode.

Note that, since the CRC-32 checksum is not collisionproof, an

Kohl & Neuman                                                  \[Page 71\]  
  
RFC 1510                        Kerberos                  September 1993

attacker could use a probabilistic chosenplaintext attack to generate  
   a valid message even if a confounder is used \[13\]. The use of  
   collision-proof checksums is recommended for environments where such  
   attacks represent a significant threat.  The use of the CRC-32 as the  
   checksum for ticket or authenticator is no longer mandated as an  
   interoperability requirement for Kerberos Version 5 Specification 1  
   (See section 9.1 for specific details).

6.3.3. DES in CBC mode with an MD4 checksum (des-cbc-md4)

The des-cbc-md4 encryption mode encrypts information under the Data  
   Encryption Standard \[11\] using the cipher block chaining mode \[12\].  
   An MD4 checksum (described in \[15\]) is applied to the confounder and  
   message sequence (msg-seq) and placed in the cksum field.  DES blocks  
   are 8 bytes.  As a result, the data to be encrypted (the  
   concatenation of confounder, checksum, and message) must be padded to  
   an 8 byte boundary before encryption.  The details of the encryption  
   of this data are identical to those for the descbc-md5 encryption  
   mode.

6.3.4. DES in CBC mode with an MD5 checksum (des-cbc-md5)

The des-cbc-md5 encryption mode encrypts information under the Data  
   Encryption Standard \[11\] using the cipher block chaining mode \[12\].  
   An MD5 checksum (described in \[16\]) is applied to the confounder and  
   message sequence (msg-seq) and placed in the cksum field.  DES blocks  
   are 8 bytes.  As a result, the data to be encrypted (the  
   concatenation of confounder, checksum, and message) must be padded to  
   an 8 byte boundary before encryption.

Plaintext and DES ciphtertext are encoded as 8-octet blocks which are  
   concatenated to make the 64-bit inputs for the DES algorithms.  The  
   first octet supplies the 8 most significant bits (with the octet's  
   MSbit used as the DES input block's MSbit, etc.), the second octet  
   the next 8 bits, ..., and the eighth octet supplies the 8 least  
   significant bits.

Encryption under DES using cipher block chaining requires an  
   additional input in the form of an initialization vector.  Unless  
   otherwise specified, zero should be used as the initialization  
   vector.  Kerberos' use of DES requires an 8-octet confounder.

The DES specifications identify some "weak" and "semiweak" keys;  
   those keys shall not be used for encrypting messages for use in  
   Kerberos.  Additionally, because of the way that keys are derived for  
   the encryption of checksums, keys shall not be used that yield "weak"  
   or "semi-weak" keys when eXclusive-ORed with the constant  
   F0F0F0F0F0F0F0F0.

Kohl & Neuman                                                  \[Page 72\]  
  
RFC 1510                        Kerberos                  September 1993

A DES key is 8 octets of data, with keytype one (1).  This consists  
   of 56 bits of key, and 8 parity bits (one per octet).  The key is  
   encoded as a series of 8 octets written in MSB-first order. The bits  
   within the key are also encoded in MSB order.  For example, if the  
   encryption key is:  
   (B1,B2,...,B7,P1,B8,...,B14,P2,B15,...,B49,P7,B50,...,B56,P8) where  
   B1,B2,...,B56 are the key bits in MSB order, and P1,P2,...,P8 are the  
   parity bits, the first octet of the key would be B1,B2,...,B7,P1  
   (with B1 as the MSbit).  \[See the FIPS 81 introduction for  
   reference.\]

To generate a DES key from a text string (password), the text string  
   normally must have the realm and each component of the principal's  
   name appended(In some cases, it may be necessary to use a different  
   "mix-in" string for compatibility reasons; see the discussion of  
   padata in section 5.4.2.), then padded with ASCII nulls to an 8 byte  
   boundary.  This string is then fan-folded and eXclusive-ORed with  
   itself to form an 8 byte DES key.  The parity is corrected on the  
   key, and it is used to generate a DES CBC checksum on the initial  
   string (with the realm and name appended).  Next, parity is corrected  
   on the CBC checksum.  If the result matches a "weak" or "semiweak"  
   key as described in the DES specification, it is eXclusive-ORed with  
   the constant 00000000000000F0.  Finally, the result is returned as  
   the key.  Pseudocode follows:

string\_to\_key(string,realm,name) \{  
             odd = 1;  
             s = string + realm;  
             for(each component in name) \{  
                  s = s + component;  
             \}  
             tempkey = NULL;  
             pad(s); /\* with nulls to 8 byte boundary \*/  
             for(8byteblock in s) \{  
                  if(odd == 0)  \{  
                      odd = 1;  
                      reverse(8byteblock)  
                  \}  
                  else odd = 0;  
                  tempkey = tempkey XOR 8byteblock;  
             \}  
             fixparity(tempkey);  
             key = DES-CBC-check(s,tempkey);  
             fixparity(key);  
             if(is\_weak\_key\_key(key))  
                  key = key XOR 0xF0;  
             return(key);  
        \}

Kohl & Neuman                                                  \[Page 73\]  
  
RFC 1510                        Kerberos                  September 1993

6.4.  Checksums

The following is the ASN.1 definition used for a checksum:

Checksum ::=   SEQUENCE \{  
                           cksumtype\[0\]   INTEGER,  
                           checksum\[1\]    OCTET STRING  
            \}

cksumtype This field indicates the algorithm used to generate the  
             accompanying checksum.

checksum  This field contains the checksum itself, encoded  
             as an octet string.

Detailed specification of selected checksum types appear later in  
   this section.  Negative values for the checksum type are reserved for  
   local use.  All non-negative values are reserved for officially  
   assigned type fields and interpretations.

Checksums used by Kerberos can be classified by two properties:  
   whether they are collision-proof, and whether they are keyed.  It is  
   infeasible to find two plaintexts which generate the same checksum  
   value for a collision-proof checksum.  A key is required to perturb  
   or initialize the algorithm in a keyed checksum.  To prevent  
   message-stream modification by an active attacker, unkeyed checksums  
   should only be used when the checksum and message will be  
   subsequently encrypted (e.g., the checksums defined as part of the  
   encryption algorithms covered earlier in this section).  Collision-  
   proof checksums can be made tamper-proof as well if the checksum  
   value is encrypted before inclusion in a message.  In such cases, the  
   composition of the checksum and the encryption algorithm must be  
   considered a separate checksum algorithm (e.g., RSA-MD5 encrypted  
   using DES is a new checksum algorithm of type RSA-MD5-DES).  For most  
   keyed checksums, as well as for the encrypted forms of collisionproof  
   checksums, Kerberos prepends a confounder before the checksum is  
   calculated.

6.4.1. The CRC-32 Checksum (crc32)

The CRC-32 checksum calculates a checksum based on a cyclic  
   redundancy check as described in ISO 3309 \[14\].  The resulting  
   checksum is four (4) octets in length.  The CRC-32 is neither keyed  
   nor collision-proof.  The use of this checksum is not recommended.  
   An attacker using a probabilistic chosen-plaintext attack as  
   described in \[13\] might be able to generate an alternative message  
   that satisfies the checksum.  The use of collision-proof checksums is  
   recommended for environments where such attacks represent a

Kohl & Neuman                                                  \[Page 74\]  
  
RFC 1510                        Kerberos                  September 1993

significant threat.

6.4.2. The RSA MD4 Checksum (rsa-md4)

The RSA-MD4 checksum calculates a checksum using the RSA MD4  
   algorithm \[15\].  The algorithm takes as input an input message of  
   arbitrary length and produces as output a 128-bit (16 octet)  
   checksum.  RSA-MD4 is believed to be collision-proof.

6.4.3. RSA MD4 Cryptographic Checksum Using DES (rsa-md4des)

The RSA-MD4-DES checksum calculates a keyed collisionproof checksum  
   by prepending an 8 octet confounder before the text, applying the RSA  
   MD4 checksum algorithm, and encrypting the confounder and the  
   checksum using DES in cipher-block-chaining (CBC) mode using a  
   variant of the key, where the variant is computed by eXclusive-ORing  
   the key with the constant F0F0F0F0F0F0F0F0 (A variant of the key is  
   used to limit the use of a key to a particular function, separating  
   the functions of generating a checksum from other encryption  
   performed using the session key.  The constant F0F0F0F0F0F0F0F0 was  
   chosen because it maintains key parity.  The properties of DES  
   precluded the use of the complement.  The same constant is used for  
   similar purpose in the Message Integrity Check in the Privacy  
   Enhanced Mail standard.).  The initialization vector should be zero.  
   The resulting checksum is 24 octets long (8 octets of which are  
   redundant).  This checksum is tamper-proof and believed to be  
   collision-proof.

The DES specifications identify some "weak keys"; those keys shall  
   not be used for generating RSA-MD4 checksums for use in Kerberos.

The format for the checksum is described in the following diagram:

+--+--+--+--+--+--+--+--  
      |  des-cbc(confounder  
      +--+--+--+--+--+--+--+--

+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+  
                        rsa-md4(confounder+msg),key=var(key),iv=0)  |  
                    +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+

The format cannot be described in ASN.1, but for those who prefer an  
   ASN.1-like notation:

rsa-md4-des-checksum ::=   ENCRYPTED       UNTAGGED SEQUENCE \{  
                              confounder\[0\]   UNTAGGED OCTET STRING(8),  
                              check\[1\]        UNTAGGED OCTET STRING(16)  
   \}

Kohl & Neuman                                                  \[Page 75\]  
  
RFC 1510                        Kerberos                  September 1993

6.4.4. The RSA MD5 Checksum (rsa-md5)

The RSA-MD5 checksum calculates a checksum using the RSA MD5  
   algorithm \[16\].  The algorithm takes as input an input message of  
   arbitrary length and produces as output a 128-bit (16 octet)  
   checksum.  RSA-MD5 is believed to be collision-proof.

6.4.5. RSA MD5 Cryptographic Checksum Using DES (rsa-md5des)

The RSA-MD5-DES checksum calculates a keyed collisionproof checksum  
   by prepending an 8 octet confounder before the text, applying the RSA  
   MD5 checksum algorithm, and encrypting the confounder and the  
   checksum using DES in cipher-block-chaining (CBC) mode using a  
   variant of the key, where the variant is computed by eXclusive-ORing  
   the key with the constant F0F0F0F0F0F0F0F0.  The initialization  
   vector should be zero.  The resulting checksum is 24 octets long (8  
   octets of which are redundant).  This checksum is tamper-proof and  
   believed to be collision-proof.

The DES specifications identify some "weak keys"; those keys shall  
   not be used for encrypting RSA-MD5 checksums for use in Kerberos.

The format for the checksum is described in the following diagram:

+--+--+--+--+--+--+--+--  
      |  des-cbc(confounder  
      +--+--+--+--+--+--+--+--

+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+  
                         rsa-md5(confounder+msg),key=var(key),iv=0)  |  
                     +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+

The format cannot be described in ASN.1, but for those who prefer an  
   ASN.1-like notation:

rsa-md5-des-checksum ::=   ENCRYPTED       UNTAGGED SEQUENCE \{  
                              confounder\[0\]   UNTAGGED OCTET STRING(8),  
                              check\[1\]        UNTAGGED OCTET STRING(16)  
   \}

6.4.6. DES cipher-block chained checksum (des-mac)

The DES-MAC checksum is computed by prepending an 8 octet confounder  
   to the plaintext, performing a DES CBC-mode encryption on the result  
   using the key and an initialization vector of zero, taking the last  
   block of the ciphertext, prepending the same confounder and  
   encrypting the pair using DES in cipher-block-chaining (CBC) mode  
   using a a variant of the key, where the variant is computed by

Kohl & Neuman                                                  \[Page 76\]  
  
RFC 1510                        Kerberos                  September 1993

eXclusive-ORing the key with the constant F0F0F0F0F0F0F0F0.  The  
   initialization vector should be zero.  The resulting checksum is 128  
   bits (16 octets) long, 64 bits of which are redundant. This checksum  
   is tamper-proof and collision-proof.

The format for the checksum is described in the following diagram:

+--+--+--+--+--+--+--+--  
      |   des-cbc(confounder  
      +--+--+--+--+--+--+--+--

+-----+-----+-----+-----+-----+-----+-----+-----+  
                       des-mac(conf+msg,iv=0,key),key=var(key),iv=0) |  
                     +-----+-----+-----+-----+-----+-----+-----+-----+

The format cannot be described in ASN.1, but for those who prefer an  
   ASN.1-like notation:

des-mac-checksum ::=    ENCRYPTED       UNTAGGED SEQUENCE \{  
                           confounder\[0\]   UNTAGGED OCTET STRING(8),  
                           check\[1\]        UNTAGGED OCTET STRING(8)  
   \}

The DES specifications identify some "weak" and "semiweak" keys;  
   those keys shall not be used for generating DES-MAC checksums for use  
   in Kerberos, nor shall a key be used whose veriant is "weak" or  
   "semi-weak".

6.4.7. RSA MD4 Cryptographic Checksum Using DES alternative  
       (rsa-md4-des-k)

The RSA-MD4-DES-K checksum calculates a keyed collision-proof  
   checksum by applying the RSA MD4 checksum algorithm and encrypting  
   the results using DES in cipherblock-chaining (CBC) mode using a DES  
   key as both key and initialization vector. The resulting checksum is  
   16 octets long. This checksum is tamper-proof and believed to be  
   collision-proof.  Note that this checksum type is the old method for  
   encoding the RSA-MD4-DES checksum and it is no longer recommended.

6.4.8. DES cipher-block chained checksum alternative (desmac-k)

The DES-MAC-K checksum is computed by performing a DES CBC-mode  
   encryption of the plaintext, and using the last block of the  
   ciphertext as the checksum value. It is keyed with an encryption key  
   and an initialization vector; any uses which do not specify an  
   additional initialization vector will use the key as both key and  
   initialization vector.  The resulting checksum is 64 bits (8 octets)  
   long. This checksum is tamper-proof and collision-proof.  Note that

Kohl & Neuman                                                  \[Page 77\]  
  
RFC 1510                        Kerberos                  September 1993

this checksum type is the old method for encoding the DESMAC checksum  
   and it is no longer recommended.

The DES specifications identify some "weak keys"; those keys shall  
   not be used for generating DES-MAC checksums for use in Kerberos.

7.  Naming Constraints

7.1.  Realm Names

Although realm names are encoded as GeneralStrings and although a  
   realm can technically select any name it chooses, interoperability  
   across realm boundaries requires agreement on how realm names are to  
   be assigned, and what information they imply.

To enforce these conventions, each realm must conform to the  
   conventions itself, and it must require that any realms with which  
   inter-realm keys are shared also conform to the conventions and  
   require the same from its neighbors.

There are presently four styles of realm names: domain, X500, other,  
   and reserved.  Examples of each style follow:

domain:   host.subdomain.domain (example)  
          X500:   C=US/O=OSF (example)  
         other:   NAMETYPE:rest/of.name=without-restrictions (example)  
      reserved:   reserved, but will not conflict with above

Domain names must look like domain names: they consist of components  
   separated by periods (.) and they contain neither colons (:) nor  
   slashes (/).

X.500 names contain an equal (=) and cannot contain a colon (:)  
   before the equal.  The realm names for X.500 names will be string  
   representations of the names with components separated by slashes.  
   Leading and trailing slashes will not be included.

Names that fall into the other category must begin with a prefix that  
   contains no equal (=) or period (.) and the prefix must be followed  
   by a colon (:) and the rest of the name. All prefixes must be  
   assigned before they may be used.  Presently none are assigned.

The reserved category includes strings which do not fall into the  
   first three categories.  All names in this category are reserved. It  
   is unlikely that names will be assigned to this category unless there  
   is a very strong argument for not using the "other" category.

These rules guarantee that there will be no conflicts between the

Kohl & Neuman                                                  \[Page 78\]  
  
RFC 1510                        Kerberos                  September 1993

various name styles.  The following additional constraints apply to  
   the assignment of realm names in the domain and X.500 categories: the  
   name of a realm for the domain or X.500 formats must either be used  
   by the organization owning (to whom it was assigned) an Internet  
   domain name or X.500 name, or in the case that no such names are  
   registered, authority to use a realm name may be derived from the  
   authority of the parent realm.  For example, if there is no domain  
   name for E40.MIT.EDU, then the administrator of the MIT.EDU realm can  
   authorize the creation of a realm with that name.

This is acceptable because the organization to which the parent is  
   assigned is presumably the organization authorized to assign names to  
   its children in the X.500 and domain name systems as well.  If the  
   parent assigns a realm name without also registering it in the domain  
   name or X.500 hierarchy, it is the parent's responsibility to make  
   sure that there will not in the future exists a name identical to the  
   realm name of the child unless it is assigned to the same entity as  
   the realm name.

7.2.  Principal Names

As was the case for realm names, conventions are needed to ensure  
   that all agree on what information is implied by a principal name.  
   The name-type field that is part of the principal name indicates the  
   kind of information implied by the name.  The name-type should be  
   treated as a hint.  Ignoring the name type, no two names can be the  
   same (i.e., at least one of the components, or the realm, must be  
   different).  This constraint may be eliminated in the future.  The  
   following name types are defined:

name-type      value   meaning  
      NT-UNKNOWN       0     Name type not known  
      NT-PRINCIPAL     1     Just the name of the principal as in  
                             DCE, or for users  
      NT-SRV-INST      2     Service and other unique instance (krbtgt)  
      NT-SRV-HST       3     Service with host name as instance  
                             (telnet, rcommands)  
      NT-SRV-XHST      4     Service with host as remaining components  
      NT-UID           5     Unique ID

When a name implies no information other than its uniqueness at a  
   particular time the name type PRINCIPAL should be used.  The  
   principal name type should be used for users, and it might also be  
   used for a unique server.  If the name is a unique machine generated  
   ID that is guaranteed never to be reassigned then the name type of  
   UID should be used (note that it is generally a bad idea to reassign  
   names of any type since stale entries might remain in access control  
   lists).

Kohl & Neuman                                                  \[Page 79\]  
  
RFC 1510                        Kerberos                  September 1993

If the first component of a name identifies a service and the  
   remaining components identify an instance of the service in a server  
   specified manner, then the name type of SRV-INST should be used.  An  
   example of this name type is the Kerberos ticket-granting ticket  
   which has a first component of krbtgt and a second component  
   identifying the realm for which the ticket is valid.

If instance is a single component following the service name and the  
   instance identifies the host on which the server is running, then the  
   name type SRV-HST should be used. This type is typically used for  
   Internet services such as telnet and the Berkeley R commands.  If the  
   separate components of the host name appear as successive components  
   following the name of the service, then the name type SRVXHST should  
   be used.  This type might be used to identify servers on hosts with  
   X.500 names where the slash (/) might otherwise be ambiguous.

A name type of UNKNOWN should be used when the form of the name is  
   not known. When comparing names, a name of type UNKNOWN will match  
   principals authenticated with names of any type.  A principal  
   authenticated with a name of type UNKNOWN, however, will only match  
   other names of type UNKNOWN.

Names of any type with an initial component of "krbtgt" are reserved  
   for the Kerberos ticket granting service.  See section 8.2.3 for the  
   form of such names.

7.2.1. Name of server principals

The principal identifier for a server on a host will generally be  
   composed of two parts: (1) the realm of the KDC with which the server  
   is registered, and (2) a two-component name of type NT-SRV-HST if the  
   host name is an Internet domain name or a multi-component name of  
   type NT-SRV-XHST if the name of the host is of a form such as X.500  
   that allows slash (/) separators.  The first component of the two- or  
   multi-component name will identify the service and the latter  
   components will identify the host.  Where the name of the host is not  
   case sensitive (for example, with Internet domain names) the name of  
   the host must be lower case.  For services such as telnet and the  
   Berkeley R commands which run with system privileges, the first  
   component will be the string "host" instead of a service specific  
   identifier.

8.  Constants and other defined values

8.1.  Host address types

All negative values for the host address type are reserved for local  
   use.  All non-negative values are reserved for officially assigned

Kohl & Neuman                                                  \[Page 80\]  
  
RFC 1510                        Kerberos                  September 1993

type fields and interpretations.

The values of the types for the following addresses are chosen to  
   match the defined address family constants in the Berkeley Standard  
   Distributions of Unix.  They can be found in <sys/socket.h> with  
   symbolic names AF\_xxx (where xxx is an abbreviation of the address  
   family name).

Internet addresses

Internet addresses are 32-bit (4-octet) quantities, encoded in MSB  
      order.  The type of internet addresses is two (2).

CHAOSnet addresses

CHAOSnet addresses are 16-bit (2-octet) quantities, encoded in MSB  
      order.  The type of CHAOSnet addresses is five (5).

ISO addresses

ISO addresses are variable-length.  The type of ISO addresses is  
      seven (7).

Xerox Network Services (XNS) addresses

XNS addresses are 48-bit (6-octet) quantities, encoded in MSB  
      order.  The type of XNS addresses is six (6).

AppleTalk Datagram Delivery Protocol (DDP) addresses

AppleTalk DDP addresses consist of an 8-bit node number and a 16-  
      bit network number.  The first octet of the address is the node  
      number; the remaining two octets encode the network number in MSB  
      order. The type of AppleTalk DDP addresses is sixteen (16).

DECnet Phase IV addresses

DECnet Phase IV addresses are 16-bit addresses, encoded in LSB  
      order.  The type of DECnet Phase IV addresses is twelve (12).

8.2.  KDC messages

8.2.1. IP transport

When contacting a Kerberos server (KDC) for a KRB\_KDC\_REQ request  
   using IP transport, the client shall send a UDP datagram containing  
   only an encoding of the request to port 88 (decimal) at the KDC's IP

Kohl & Neuman                                                  \[Page 81\]  
  
RFC 1510                        Kerberos                  September 1993

address; the KDC will respond with a reply datagram containing only  
   an encoding of the reply message (either a KRB\_ERROR or a  
   KRB\_KDC\_REP) to the sending port at the sender's IP address.

8.2.2. OSI transport

During authentication of an OSI client to and OSI server, the mutual  
   authentication of an OSI server to an OSI client, the transfer of  
   credentials from an OSI client to an OSI server, or during exchange  
   of private or integrity checked messages, Kerberos protocol messages  
   may be treated as opaque objects and the type of the authentication  
   mechanism will be:

OBJECT IDENTIFIER ::= \{iso (1), org(3), dod(5),internet(1),  
                          security(5), kerberosv5(2)\}

Depending on the situation, the opaque object will be an  
   authentication header (KRB\_AP\_REQ), an authentication reply  
   (KRB\_AP\_REP), a safe message (KRB\_SAFE), a private message  
   (KRB\_PRIV), or a credentials message (KRB\_CRED).  The opaque data  
   contains an application code as specified in the ASN.1 description  
   for each message.  The application code may be used by Kerberos to  
   determine the message type.

8.2.3. Name of the TGS

The principal identifier of the ticket-granting service shall be  
   composed of three parts: (1) the realm of the KDC issuing the TGS  
   ticket (2) a two-part name of type NT-SRVINST, with the first part  
   "krbtgt" and the second part the name of the realm which will accept  
   the ticket-granting ticket.  For example, a ticket-granting ticket  
   issued by the ATHENA.MIT.EDU realm to be used to get tickets from the  
   ATHENA.MIT.EDU KDC has a principal identifier of "ATHENA.MIT.EDU"  
   (realm), ("krbtgt", "ATHENA.MIT.EDU") (name).  A ticket-granting  
   ticket issued by the ATHENA.MIT.EDU realm to be used to get tickets  
   from the MIT.EDU realm has a principal identifier of "ATHENA.MIT.EDU"  
   (realm), ("krbtgt", "MIT.EDU") (name).

8.3.  Protocol constants and associated values

The following tables list constants used in the protocol and defines  
   their meanings.

Kohl & Neuman                                                  \[Page 82\]  
  
RFC 1510                        Kerberos                  September 1993

\---------------+-----------+----------+----------------+---------------  
Encryption type|etype value|block size|minimum pad size|confounder size  
\---------------+-----------+----------+----------------+---------------  
NULL                0            1              0              0  
des-cbc-crc         1            8              4              8  
des-cbc-md4         2            8              0              8  
des-cbc-md5         3            8              0              8

\-------------------------------+-------------------+-------------  
Checksum type                  |sumtype value      |checksum size  
\-------------------------------+-------------------+-------------  
CRC32                           1                   4  
rsa-md4                         2                   16  
rsa-md4-des                     3                   24  
des-mac                         4                   16  
des-mac-k                       5                   8  
rsa-md4-des-k                   6                   16  
rsa-md5                         7                   16  
rsa-md5-des                     8                   24

\-------------------------------+-----------------  
padata type                    |padata-type value  
\-------------------------------+-----------------  
PA-TGS-REQ                      1  
PA-ENC-TIMESTAMP                2  
PA-PW-SALT                      3

\-------------------------------+-------------  
authorization data type        |ad-type value  
\-------------------------------+-------------  
reserved values                 0-63  
OSF-DCE                         64  
SESAME                          65

\-------------------------------+-----------------  
alternate authentication type  |method-type value  
\-------------------------------+-----------------  
reserved values                 0-63  
ATT-CHALLENGE-RESPONSE          64

\-------------------------------+-------------  
transited encoding type        |tr-type value  
\-------------------------------+-------------  
DOMAIN-X500-COMPRESS            1  
reserved values                 all others

Kohl & Neuman                                                  \[Page 83\]  
  
RFC 1510                        Kerberos                  September 1993

\--------------+-------+-----------------------------------------  
Label         |Value  |Meaning or MIT code  
\--------------+-------+-----------------------------------------

pvno             5     current Kerberos protocol version number

message types

KRB\_AS\_REQ      10     Request for initial authentication  
KRB\_AS\_REP      11     Response to KRB\_AS\_REQ request  
KRB\_TGS\_REQ     12     Request for authentication based on TGT  
KRB\_TGS\_REP     13     Response to KRB\_TGS\_REQ request  
KRB\_AP\_REQ      14     application request to server  
KRB\_AP\_REP      15     Response to KRB\_AP\_REQ\_MUTUAL  
KRB\_SAFE        20     Safe (checksummed) application message  
KRB\_PRIV        21     Private (encrypted) application message  
KRB\_CRED        22     Private (encrypted) message to forward  
                       credentials  
KRB\_ERROR       30     Error response

name types

KRB\_NT\_UNKNOWN   0   Name type not known  
KRB\_NT\_PRINCIPAL 1   Just the name of the principal as in DCE, or  
                     for users  
KRB\_NT\_SRV\_INST  2   Service and other unique instance (krbtgt)  
KRB\_NT\_SRV\_HST   3   Service with host name as instance (telnet,  
                     rcommands)  
KRB\_NT\_SRV\_XHST  4   Service with host as remaining components  
KRB\_NT\_UID       5   Unique ID

error codes

KDC\_ERR\_NONE                   0   No error  
KDC\_ERR\_NAME\_EXP               1   Client's entry in database has  
                                   expired  
KDC\_ERR\_SERVICE\_EXP            2   Server's entry in database has  
                                   expired  
KDC\_ERR\_BAD\_PVNO               3   Requested protocol version number  
                                   not supported  
KDC\_ERR\_C\_OLD\_MAST\_KVNO        4   Client's key encrypted in old  
                                   master key  
KDC\_ERR\_S\_OLD\_MAST\_KVNO        5   Server's key encrypted in old  
                                   master key  
KDC\_ERR\_C\_PRINCIPAL\_UNKNOWN    6   Client not found in Kerberos database  
KDC\_ERR\_S\_PRINCIPAL\_UNKNOWN    7   Server not found in Kerberos database  
KDC\_ERR\_PRINCIPAL\_NOT\_UNIQUE   8   Multiple principal entries in  
                                   database

Kohl & Neuman                                                  \[Page 84\]  
  
RFC 1510                        Kerberos                  September 1993

KDC\_ERR\_NULL\_KEY               9   The client or server has a null key  
KDC\_ERR\_CANNOT\_POSTDATE       10   Ticket not eligible for postdating  
KDC\_ERR\_NEVER\_VALID           11   Requested start time is later than  
                                   end time  
KDC\_ERR\_POLICY                12   KDC policy rejects request  
KDC\_ERR\_BADOPTION             13   KDC cannot accommodate requested  
                                   option  
KDC\_ERR\_ETYPE\_NOSUPP          14   KDC has no support for encryption  
                                   type  
KDC\_ERR\_SUMTYPE\_NOSUPP        15   KDC has no support for checksum type  
KDC\_ERR\_PADATA\_TYPE\_NOSUPP    16   KDC has no support for padata type  
KDC\_ERR\_TRTYPE\_NOSUPP         17   KDC has no support for transited type  
KDC\_ERR\_CLIENT\_REVOKED        18   Clients credentials have been revoked  
KDC\_ERR\_SERVICE\_REVOKED       19   Credentials for server have been  
                                   revoked  
KDC\_ERR\_TGT\_REVOKED           20   TGT has been revoked  
KDC\_ERR\_CLIENT\_NOTYET         21   Client not yet valid - try again  
                                   later  
KDC\_ERR\_SERVICE\_NOTYET        22   Server not yet valid - try again  
                                   later  
KDC\_ERR\_KEY\_EXPIRED           23   Password has expired - change  
                                   password to reset  
KDC\_ERR\_PREAUTH\_FAILED        24   Pre-authentication information  
                                   was invalid  
KDC\_ERR\_PREAUTH\_REQUIRED      25   Additional pre-authentication  
                                   required\*  
KRB\_AP\_ERR\_BAD\_INTEGRITY      31   Integrity check on decrypted field  
                                   failed  
KRB\_AP\_ERR\_TKT\_EXPIRED        32   Ticket expired  
KRB\_AP\_ERR\_TKT\_NYV            33   Ticket not yet valid  
KRB\_AP\_ERR\_REPEAT             34   Request is a replay  
KRB\_AP\_ERR\_NOT\_US             35   The ticket isn't for us  
KRB\_AP\_ERR\_BADMATCH           36   Ticket and authenticator don't match  
KRB\_AP\_ERR\_SKEW               37   Clock skew too great  
KRB\_AP\_ERR\_BADADDR            38   Incorrect net address  
KRB\_AP\_ERR\_BADVERSION         39   Protocol version mismatch  
KRB\_AP\_ERR\_MSG\_TYPE           40   Invalid msg type  
KRB\_AP\_ERR\_MODIFIED           41   Message stream modified  
KRB\_AP\_ERR\_BADORDER           42   Message out of order  
KRB\_AP\_ERR\_BADKEYVER          44   Specified version of key is not  
                                   available  
KRB\_AP\_ERR\_NOKEY              45   Service key not available  
KRB\_AP\_ERR\_MUT\_FAIL           46   Mutual authentication failed  
KRB\_AP\_ERR\_BADDIRECTION       47   Incorrect message direction  
KRB\_AP\_ERR\_METHOD             48   Alternative authentication method  
                                   required\*  
KRB\_AP\_ERR\_BADSEQ             49   Incorrect sequence number in message  
KRB\_AP\_ERR\_INAPP\_CKSUM        50   Inappropriate type of checksum in

Kohl & Neuman                                                  \[Page 85\]  
  
RFC 1510                        Kerberos                  September 1993

message  
KRB\_ERR\_GENERIC               60   Generic error (description in e-text)  
KRB\_ERR\_FIELD\_TOOLONG         61   Field is too long for this  
                                   implementation

\*This error carries additional information in the e-data field.  The  
   contents of the e-data field for this message is described in section  
   5.9.1.

9.  Interoperability requirements

Version 5 of the Kerberos protocol supports a myriad of options.  
   Among these are multiple encryption and checksum types, alternative  
   encoding schemes for the transited field, optional mechanisms for  
   pre-authentication, the handling of tickets with no addresses,  
   options for mutual authentication, user to user authentication,  
   support for proxies, forwarding, postdating, and renewing tickets,  
   the format of realm names, and the handling of authorization data.

In order to ensure the interoperability of realms, it is necessary to  
   define a minimal configuration which must be supported by all  
   implementations.  This minimal configuration is subject to change as  
   technology does. For example, if at some later date it is discovered  
   that one of the required encryption or checksum algorithms is not  
   secure, it will be replaced.

9.1.  Specification 1

This section defines the first specification of these options.  
   Implementations which are configured in this way can be said to  
   support Kerberos Version 5 Specification 1 (5.1).

Encryption and checksum methods

The following encryption and checksum mechanisms must be supported.  
   Implementations may support other mechanisms as well, but the  
   additional mechanisms may only be used when communicating with  
   principals known to also support them: Encryption: DES-CBC-MD5  
   Checksums: CRC-32, DES-MAC, DES-MAC-K, and DES-MD5

Realm Names

All implementations must understand hierarchical realms in both the  
   Internet Domain and the X.500 style.  When a ticket granting ticket  
   for an unknown realm is requested, the KDC must be able to determine  
   the names of the intermediate realms between the KDCs realm and the  
   requested realm.

Kohl & Neuman                                                  \[Page 86\]  
  
RFC 1510                        Kerberos                  September 1993

Transited field encoding

DOMAIN-X500-COMPRESS (described in section 3.3.3.1) must be  
   supported.  Alternative encodings may be supported, but they may be  
   used only when that encoding is supported by ALL intermediate realms.

Pre-authentication methods

The TGS-REQ method must be supported.  The TGS-REQ method is not used  
   on the initial request. The PA-ENC-TIMESTAMP method must be supported  
   by clients but whether it is enabled by default may be determined on  
   a realm by realm basis. If not used in the initial request and the  
   error KDC\_ERR\_PREAUTH\_REQUIRED is returned specifying PA-ENCTIMESTAMP  
   as an acceptable method, the client should retry the initial request  
   using the PA-ENC-TIMESTAMP preauthentication method. Servers need not  
   support the PAENC-TIMESTAMP method, but if not supported the server  
   should ignore the presence of PA-ENC-TIMESTAMP pre-authentication in  
   a request.

Mutual authentication

Mutual authentication (via the KRB\_AP\_REP message) must be supported.

Ticket addresses and flags

All KDC's must pass on tickets that carry no addresses (i.e.,  if a  
   TGT contains no addresses, the KDC will return derivative tickets),  
   but each realm may set its own policy for issuing such tickets, and  
   each application server will set its own policy with respect to  
   accepting them. By default, servers should not accept them.

Proxies and forwarded tickets must be supported.  Individual realms  
   and application servers can set their own policy on when such tickets  
   will be accepted.

All implementations must recognize renewable and postdated tickets,  
   but need not actually implement them.  If these options are not  
   supported, the starttime and endtime in the ticket shall specify a  
   ticket's entire useful life.  When a postdated ticket is decoded by a  
   server, all implementations shall make the presence of the postdated  
   flag visible to the calling server.

User-to-user authentication

Support for user to user authentication (via the ENC-TKTIN-SKEY KDC  
   option) must be provided by implementations, but individual realms  
   may decide as a matter of policy to reject such requests on a per-  
   principal or realm-wide basis.

Kohl & Neuman                                                  \[Page 87\]  
  
RFC 1510                        Kerberos                  September 1993

Authorization data

Implementations must pass all authorization data subfields from  
   ticket-granting tickets to any derivative tickets unless directed to  
   suppress a subfield as part of the definition of that registered  
   subfield type (it is never incorrect to pass on a subfield, and no  
   registered subfield types presently specify suppression at the KDC).

Implementations must make the contents of any authorization data  
   subfields available to the server when a ticket is used.  
   Implementations are not required to allow clients to specify the  
   contents of the authorization data fields.

9.2.  Recommended KDC values

Following is a list of recommended values for a KDC implementation,  
   based on the list of suggested configuration constants (see section  
   4.4).

minimum lifetime                5 minutes

maximum renewable lifetime      1 week

maximum ticket lifetime         1 day

empty addresses                 only when suitable restrictions appear  
                                   in authorization data

proxiable, etc.                 Allowed.

10.  Acknowledgments

Early versions of this document, describing version 4 of the  
   protocol, were written by Jennifer Steiner (formerly at Project  
   Athena); these drafts provided an excellent starting point for this  
   current version 5 specification.  Many people in the Internet  
   community have contributed ideas and suggested protocol changes for  
   version 5. Notable contributions came from Ted Anderson, Steve  
   Bellovin and Michael Merritt \[17\], Daniel Bernstein, Mike Burrows,  
   Donald Davis, Ravi Ganesan, Morrie Gasser, Virgil Gligor, Bill  
   Griffeth, Mark Lillibridge, Mark Lomas, Steve Lunt, Piers McMahon,  
   Joe Pato, William Sommerfeld, Stuart Stubblebine, Ralph Swick, Ted  
   T'so, and Stanley Zanarotti.  Many others commented and helped shape  
   this specification into its current form.

Kohl & Neuman                                                  \[Page 88\]  
  
RFC 1510                        Kerberos                  September 1993

11.  References

\[1\]  Miller, S., Neuman, C., Schiller, J., and  J. Saltzer, "Section  
        E.2.1: Kerberos  Authentication and Authorization System",  
        M.I.T. Project Athena, Cambridge, Massachusetts, December 21,  
        1987.

\[2\]  Steiner, J., Neuman, C., and J. Schiller, "Kerberos: An  
        Authentication Service for Open Network Systems", pp. 191-202 in  
        Usenix Conference Proceedings, Dallas, Texas, February, 1988.

\[3\]  Needham, R., and M. Schroeder, "Using Encryption for  
        Authentication in Large Networks of Computers", Communications  
        of the ACM, Vol. 21 (12), pp. 993-999, December 1978.

\[4\]  Denning, D., and G. Sacco, "Time stamps in Key Distribution  
        Protocols", Communications of the ACM, Vol. 24 (8), pp. 533-536,  
        August 1981.

\[5\]  Kohl, J., Neuman, C., and T. Ts'o, "The Evolution of the  
        Kerberos Authentication Service", in an IEEE Computer Society  
        Text soon to be published, June 1992.

\[6\]  Davis, D., and R. Swick, "Workstation Services and Kerberos  
        Authentication at Project Athena", Technical Memorandum TM-424,  
        MIT Laboratory for Computer Science, February 1990.

\[7\]  Levine, P., Gretzinger, M, Diaz, J., Sommerfeld, W., and K.  
        Raeburn, "Section E.1: Service Management System, M.I.T.  
        Project Athena, Cambridge, Mas sachusetts (1987).

\[8\]  CCITT, Recommendation X.509: The Directory Authentication  
        Framework, December 1988.

\[9\]  Neuman, C., "Proxy-Based Authorization and Accounting for  
        Distributed Systems," in Proceedings of the 13th International  
        Conference on Distributed Computing Systems", Pittsburgh, PA,  
        May 1993.

\[10\] Pato, J., "Using Pre-Authentication to Avoid Password Guessing  
        Attacks", Open Software Foundation DCE Request for Comments 26,  
        December 1992.

\[11\] National Bureau of Standards, U.S. Department of Commerce, "Data  
        Encryption Standard", Federal Information Processing Standards  
        Publication 46, Washington, DC (1977).

Kohl & Neuman                                                  \[Page 89\]  
  
RFC 1510                        Kerberos                  September 1993

\[12\] National Bureau of Standards, U.S. Department of Commerce, "DES  
        Modes of Operation", Federal Information Processing Standards  
        Publication 81, Springfield, VA, December 1980.

\[13\] Stubblebine S., and V. Gligor, "On Message Integrity in  
        Cryptographic Protocols", in Proceedings of the IEEE Symposium  
        on Research in Security and Privacy, Oakland, California, May  
        1992.

\[14\] International Organization for Standardization, "ISO Information  
        Processing Systems - Data Communication High-Level Data Link  
        Control Procedure - Frame Structure", IS 3309, October 1984, 3rd  
        Edition.

\[15\] Rivest, R., "The MD4 Message Digest Algorithm", RFC 1320, MIT  
        Laboratory for Computer Science, April 1992.

\[16\] Rivest, R., "The MD5 Message Digest Algorithm", RFC 1321, MIT  
        Laboratory for Computer Science, April 1992.

\[17\] Bellovin S., and M. Merritt, "Limitations of the Kerberos  
        Authentication System", Computer Communications Review, Vol.  
        20(5), pp. 119-132, October 1990.

12.  Security Considerations

Security issues are discussed throughout this memo.

13.  Authors' Addresses

John Kohl  
   Digital Equipment Corporation  
   110 Spit Brook Road, M/S ZKO3-3/U14  
   Nashua, NH  03062

Phone: 603-881-2481  
   EMail: [jtkohl@zk3.dec.com][jtkohl_zk3.dec.com]

B. Clifford Neuman  
   USC/Information Sciences Institute  
   4676 Admiralty Way \#1001  
   Marina del Rey, CA 90292-6695

Phone: 310-822-1511  
   EMail: [bcn@isi.edu][bcn_isi.edu]

Kohl & Neuman                                                  \[Page 90\]  
  
RFC 1510                        Kerberos                  September 1993

A.  Pseudo-code for protocol processing

This appendix provides pseudo-code describing how the messages are to  
   be constructed and interpreted by clients and servers.

A.1.  KRB\_AS\_REQ generation  
        request.pvno := protocol version; /\* pvno = 5 \*/  
        request.msg-type := message type; /\* type = KRB\_AS\_REQ \*/

if(pa\_enc\_timestamp\_required) then  
                request.padata.padata-type = PA-ENC-TIMESTAMP;  
                get system\_time;  
                padata-body.patimestamp,pausec = system\_time;  
                encrypt padata-body into request.padata.padata-value  
                        using client.key; /\* derived from password \*/  
        endif

body.kdc-options := users's preferences;  
        body.cname := user's name;  
        body.realm := user's realm;  
        body.sname := service's name; /\* usually "krbtgt",  
                                         "localrealm" \*/  
        if (body.kdc-options.POSTDATED is set) then  
                body.from := requested starting time;  
        else  
                omit body.from;  
        endif  
        body.till := requested end time;  
        if (body.kdc-options.RENEWABLE is set) then  
                body.rtime := requested final renewal time;  
        endif  
        body.nonce := random\_nonce();  
        body.etype := requested etypes;  
        if (user supplied addresses) then  
                body.addresses := user's addresses;  
        else  
                omit body.addresses;  
        endif  
        omit body.enc-authorization-data;  
        request.req-body := body;

kerberos := lookup(name of local kerberos server (or servers));  
        send(packet,kerberos);

wait(for response);  
        if (timed\_out) then  
                retry or use alternate server;  
        endif

Kohl & Neuman                                                  \[Page 91\]  
  
RFC 1510                        Kerberos                  September 1993

A.2.  KRB\_AS\_REQ verification and KRB\_AS\_REP generation  
        decode message into req;

client := lookup(req.cname,req.realm);  
        server := lookup(req.sname,req.realm);  
        get system\_time;  
        kdc\_time := system\_time.seconds;

if (!client) then  
                /\* no client in Database \*/  
                error\_out(KDC\_ERR\_C\_PRINCIPAL\_UNKNOWN);  
        endif  
        if (!server) then  
                /\* no server in Database \*/  
                error\_out(KDC\_ERR\_S\_PRINCIPAL\_UNKNOWN);  
        endif

if(client.pa\_enc\_timestamp\_required and  
           pa\_enc\_timestamp not present) then  
                error\_out(KDC\_ERR\_PREAUTH\_REQUIRED(PA\_ENC\_TIMESTAMP));  
        endif

if(pa\_enc\_timestamp present) then  
                decrypt req.padata-value into decrypted\_enc\_timestamp  
                        using client.key;  
                        using auth\_hdr.authenticator.subkey;  
                if (decrypt\_error()) then  
                        error\_out(KRB\_AP\_ERR\_BAD\_INTEGRITY);  
                if(decrypted\_enc\_timestamp is not within allowable  
                        skew) then error\_out(KDC\_ERR\_PREAUTH\_FAILED);  
                endif  
                if(decrypted\_enc\_timestamp and usec is replay)  
                        error\_out(KDC\_ERR\_PREAUTH\_FAILED);  
                endif  
                add decrypted\_enc\_timestamp and usec to replay cache;  
        endif

use\_etype := first supported etype in req.etypes;

if (no support for req.etypes) then  
                error\_out(KDC\_ERR\_ETYPE\_NOSUPP);  
        endif

new\_tkt.vno := ticket version; /\* = 5 \*/  
        new\_tkt.sname := req.sname;  
        new\_tkt.srealm := req.srealm;  
        reset all flags in new\_tkt.flags;

Kohl & Neuman                                                  \[Page 92\]  
  
RFC 1510                        Kerberos                  September 1993

/\* It should be noted that local policy may affect the  \*/  
        /\* processing of any of these flags.  For example, some \*/  
        /\* realms may refuse to issue renewable tickets         \*/

if (req.kdc-options.FORWARDABLE is set) then  
                set new\_tkt.flags.FORWARDABLE;  
        endif  
        if (req.kdc-options.PROXIABLE is set) then  
                set new\_tkt.flags.PROXIABLE;  
        endif  
        if (req.kdc-options.ALLOW-POSTDATE is set) then  
                set new\_tkt.flags.ALLOW-POSTDATE;  
        endif  
        if ((req.kdc-options.RENEW is set) or  
            (req.kdc-options.VALIDATE is set) or  
            (req.kdc-options.PROXY is set) or  
            (req.kdc-options.FORWARDED is set) or  
            (req.kdc-options.ENC-TKT-IN-SKEY is set)) then  
                error\_out(KDC\_ERR\_BADOPTION);  
        endif

new\_tkt.session := random\_session\_key();  
        new\_tkt.cname := req.cname;  
        new\_tkt.crealm := req.crealm;  
        new\_tkt.transited := empty\_transited\_field();

new\_tkt.authtime := kdc\_time;

if (req.kdc-options.POSTDATED is set) then  
           if (against\_postdate\_policy(req.from)) then  
                error\_out(KDC\_ERR\_POLICY);  
           endif  
           set new\_tkt.flags.INVALID;  
           new\_tkt.starttime := req.from;  
        else  
           omit new\_tkt.starttime; /\* treated as authtime when  
                                      omitted \*/  
        endif  
        if (req.till = 0) then  
                till := infinity;  
        else  
                till := req.till;  
        endif

new\_tkt.endtime := min(till,  
                              new\_tkt.starttime+client.max\_life,  
                              new\_tkt.starttime+server.max\_life,  
                              new\_tkt.starttime+max\_life\_for\_realm);

Kohl & Neuman                                                  \[Page 93\]  
  
RFC 1510                        Kerberos                  September 1993

if ((req.kdc-options.RENEWABLE-OK is set) and  
            (new\_tkt.endtime < req.till)) then  
                /\* we set the RENEWABLE option for later processing \*/  
                set req.kdc-options.RENEWABLE;  
                req.rtime := req.till;  
        endif

if (req.rtime = 0) then  
                rtime := infinity;  
        else  
                rtime := req.rtime;  
        endif

if (req.kdc-options.RENEWABLE is set) then  
                set new\_tkt.flags.RENEWABLE;  
                new\_tkt.renew-till := min(rtime,  
                new\_tkt.starttime+client.max\_rlife,  
                new\_tkt.starttime+server.max\_rlife,  
                new\_tkt.starttime+max\_rlife\_for\_realm);  
        else  
                omit new\_tkt.renew-till; /\* only present if RENEWABLE \*/  
        endif

if (req.addresses) then  
                new\_tkt.caddr := req.addresses;  
        else  
                omit new\_tkt.caddr;  
        endif

new\_tkt.authorization\_data := empty\_authorization\_data();

encode to-be-encrypted part of ticket into OCTET STRING;  
        new\_tkt.enc-part := encrypt OCTET STRING  
            using etype\_for\_key(server.key), server.key, server.p\_kvno;

/\* Start processing the response \*/

resp.pvno := 5;  
        resp.msg-type := KRB\_AS\_REP;  
        resp.cname := req.cname;  
        resp.crealm := req.realm;  
        resp.ticket := new\_tkt;

resp.key := new\_tkt.session;  
        resp.last-req := fetch\_last\_request\_info(client);  
        resp.nonce := req.nonce;  
        resp.key-expiration := client.expiration;

Kohl & Neuman                                                  \[Page 94\]  
  
RFC 1510                        Kerberos                  September 1993

resp.flags := new\_tkt.flags;

resp.authtime := new\_tkt.authtime;  
        resp.starttime := new\_tkt.starttime;  
        resp.endtime := new\_tkt.endtime;

if (new\_tkt.flags.RENEWABLE) then  
                resp.renew-till := new\_tkt.renew-till;  
        endif

resp.realm := new\_tkt.realm;  
        resp.sname := new\_tkt.sname;

resp.caddr := new\_tkt.caddr;

encode body of reply into OCTET STRING;

resp.enc-part := encrypt OCTET STRING  
                         using use\_etype, client.key, client.p\_kvno;  
        send(resp);

A.3.  KRB\_AS\_REP verification  
        decode response into resp;

if (resp.msg-type = KRB\_ERROR) then  
                if(error = KDC\_ERR\_PREAUTH\_REQUIRED(PA\_ENC\_TIMESTAMP))  
                        then set pa\_enc\_timestamp\_required;  
                        goto KRB\_AS\_REQ;  
                endif  
                process\_error(resp);  
                return;  
        endif

/\* On error, discard the response, and zero the session key \*/  
        /\* from the response immediately \*/

key = get\_decryption\_key(resp.enc-part.kvno, resp.enc-part.etype,  
                                 resp.padata);  
        unencrypted part of resp := decode of decrypt of resp.enc-part  
                                using resp.enc-part.etype and key;  
        zero(key);

if (common\_as\_rep\_tgs\_rep\_checks fail) then  
                destroy resp.key;  
                return error;  
        endif

if near(resp.princ\_exp) then

Kohl & Neuman                                                  \[Page 95\]  
  
RFC 1510                        Kerberos                  September 1993

print(warning message);  
        endif  
        save\_for\_later(ticket,session,client,server,times,flags);

A.4.  KRB\_AS\_REP and KRB\_TGS\_REP common checks  
        if (decryption\_error() or  
            (req.cname != resp.cname) or  
            (req.realm != resp.crealm) or  
            (req.sname != resp.sname) or  
            (req.realm != resp.realm) or  
            (req.nonce != resp.nonce) or  
            (req.addresses != resp.caddr)) then  
                destroy resp.key;  
                return KRB\_AP\_ERR\_MODIFIED;  
        endif

/\* make sure no flags are set that shouldn't be, and that  \*/  
        /\* all that should be are set                              \*/  
        if (!check\_flags\_for\_compatability(req.kdc-options,resp.flags))  
                then destroy resp.key;  
                return KRB\_AP\_ERR\_MODIFIED;  
        endif

if ((req.from = 0) and  
            (resp.starttime is not within allowable skew)) then  
                destroy resp.key;  
                return KRB\_AP\_ERR\_SKEW;  
        endif  
        if ((req.from != 0) and (req.from != resp.starttime)) then  
                destroy resp.key;  
                return KRB\_AP\_ERR\_MODIFIED;  
        endif  
        if ((req.till != 0) and (resp.endtime > req.till)) then  
                destroy resp.key;  
                return KRB\_AP\_ERR\_MODIFIED;  
        endif

if ((req.kdc-options.RENEWABLE is set) and  
            (req.rtime != 0) and (resp.renew-till > req.rtime)) then  
                destroy resp.key;  
                return KRB\_AP\_ERR\_MODIFIED;  
        endif  
        if ((req.kdc-options.RENEWABLE-OK is set) and  
            (resp.flags.RENEWABLE) and  
            (req.till != 0) and  
            (resp.renew-till > req.till)) then  
                destroy resp.key;  
                return KRB\_AP\_ERR\_MODIFIED;

Kohl & Neuman                                                  \[Page 96\]  
  
RFC 1510                        Kerberos                  September 1993

endif

A.5.  KRB\_TGS\_REQ generation  
        /\* Note that make\_application\_request might have to     \*/  
        /\* recursivly call this routine to get the appropriate  \*/  
        /\* ticket-granting ticket                               \*/

request.pvno := protocol version; /\* pvno = 5 \*/  
        request.msg-type := message type; /\* type = KRB\_TGS\_REQ \*/

body.kdc-options := users's preferences;  
        /\* If the TGT is not for the realm of the end-server  \*/  
        /\* then the sname will be for a TGT for the end-realm \*/  
        /\* and the realm of the requested ticket (body.realm) \*/  
        /\* will be that of the TGS to which the TGT we are    \*/  
        /\* sending applies                                    \*/  
        body.sname := service's name;  
        body.realm := service's realm;

if (body.kdc-options.POSTDATED is set) then  
                body.from := requested starting time;  
        else  
                omit body.from;  
        endif  
        body.till := requested end time;  
        if (body.kdc-options.RENEWABLE is set) then  
                body.rtime := requested final renewal time;  
        endif  
        body.nonce := random\_nonce();  
        body.etype := requested etypes;  
        if (user supplied addresses) then  
                body.addresses := user's addresses;  
        else  
                omit body.addresses;  
        endif

body.enc-authorization-data := user-supplied data;  
        if (body.kdc-options.ENC-TKT-IN-SKEY) then  
                body.additional-tickets\_ticket := second TGT;  
        endif

request.req-body := body;  
        check := generate\_checksum (req.body,checksumtype);

request.padata\[0\].padata-type := PA-TGS-REQ;  
        request.padata\[0\].padata-value := create a KRB\_AP\_REQ using  
                                      the TGT and checksum

Kohl & Neuman                                                  \[Page 97\]  
  
RFC 1510                        Kerberos                  September 1993

/\* add in any other padata as required/supplied \*/

kerberos := lookup(name of local kerberose server (or servers));  
        send(packet,kerberos);

wait(for response);  
        if (timed\_out) then  
                retry or use alternate server;  
        endif

A.6.  KRB\_TGS\_REQ verification and KRB\_TGS\_REP generation  
        /\* note that reading the application request requires first  
        determining the server for which a ticket was issued, and  
        choosing the correct key for decryption.  The name of the  
        server appears in the plaintext part of the ticket. \*/

if (no KRB\_AP\_REQ in req.padata) then  
                error\_out(KDC\_ERR\_PADATA\_TYPE\_NOSUPP);  
        endif  
        verify KRB\_AP\_REQ in req.padata;

/\* Note that the realm in which the Kerberos server is  
        operating is determined by the instance from the  
        ticket-granting ticket.  The realm in the ticket-granting  
        ticket is the realm under which the ticket granting ticket was  
        issued.  It is possible for a single Kerberos server to  
        support more than one realm. \*/

auth\_hdr := KRB\_AP\_REQ;  
        tgt := auth\_hdr.ticket;

if (tgt.sname is not a TGT for local realm and is not  
                req.sname) then error\_out(KRB\_AP\_ERR\_NOT\_US);

realm := realm\_tgt\_is\_for(tgt);

decode remainder of request;

if (auth\_hdr.authenticator.cksum is missing) then  
                error\_out(KRB\_AP\_ERR\_INAPP\_CKSUM);  
        endif  
        if (auth\_hdr.authenticator.cksum type is not supported) then  
                error\_out(KDC\_ERR\_SUMTYPE\_NOSUPP);  
        endif  
        if (auth\_hdr.authenticator.cksum is not both collision-proof  
            and keyed)  then  
                error\_out(KRB\_AP\_ERR\_INAPP\_CKSUM);  
        endif

Kohl & Neuman                                                  \[Page 98\]  
  
RFC 1510                        Kerberos                  September 1993

set computed\_checksum := checksum(req);  
        if (computed\_checksum != auth\_hdr.authenticatory.cksum) then  
                error\_out(KRB\_AP\_ERR\_MODIFIED);  
        endif

server := lookup(req.sname,realm);

if (!server) then  
                if (is\_foreign\_tgt\_name(server)) then  
                        server := best\_intermediate\_tgs(server);  
                else  
                        /\* no server in Database \*/  
                        error\_out(KDC\_ERR\_S\_PRINCIPAL\_UNKNOWN);  
                endif  
        endif

session := generate\_random\_session\_key();

use\_etype := first supported etype in req.etypes;

if (no support for req.etypes) then  
                error\_out(KDC\_ERR\_ETYPE\_NOSUPP);  
        endif

new\_tkt.vno := ticket version; /\* = 5 \*/  
        new\_tkt.sname := req.sname;  
        new\_tkt.srealm := realm;  
        reset all flags in new\_tkt.flags;

new\_tkt.caddr := tgt.caddr;  
        resp.caddr := NULL; /\* We only include this if they change \*/  
        if (req.kdc-options.FORWARDABLE is set) then  
                if (tgt.flags.FORWARDABLE is reset) then  
                        error\_out(KDC\_ERR\_BADOPTION);  
                endif  
                set new\_tkt.flags.FORWARDABLE;  
        endif  
        if (req.kdc-options.FORWARDED is set) then  
                if (tgt.flags.FORWARDABLE is reset) then  
                        error\_out(KDC\_ERR\_BADOPTION);  
                endif  
                set new\_tkt.flags.FORWARDED;  
                new\_tkt.caddr := req.addresses;

Kohl & Neuman                                                  \[Page 99\]  
  
RFC 1510                        Kerberos                  September 1993

resp.caddr := req.addresses;  
        endif  
        if (tgt.flags.FORWARDED is set) then  
                set new\_tkt.flags.FORWARDED;  
        endif

if (req.kdc-options.PROXIABLE is set) then  
                if (tgt.flags.PROXIABLE is reset)  
                        error\_out(KDC\_ERR\_BADOPTION);  
                endif  
                set new\_tkt.flags.PROXIABLE;  
        endif  
        if (req.kdc-options.PROXY is set) then  
                if (tgt.flags.PROXIABLE is reset) then  
                        error\_out(KDC\_ERR\_BADOPTION);  
                endif  
                set new\_tkt.flags.PROXY;  
                new\_tkt.caddr := req.addresses;  
                resp.caddr := req.addresses;  
        endif

if (req.kdc-options.POSTDATE is set) then  
                if (tgt.flags.POSTDATE is reset)  
                        error\_out(KDC\_ERR\_BADOPTION);  
                endif  
                set new\_tkt.flags.POSTDATE;  
        endif  
        if (req.kdc-options.POSTDATED is set) then  
                if (tgt.flags.POSTDATE is reset) then  
                        error\_out(KDC\_ERR\_BADOPTION);  
                endif  
                set new\_tkt.flags.POSTDATED;  
                set new\_tkt.flags.INVALID;  
                if (against\_postdate\_policy(req.from)) then  
                        error\_out(KDC\_ERR\_POLICY);  
                endif  
                new\_tkt.starttime := req.from;  
        endif

if (req.kdc-options.VALIDATE is set) then  
                if (tgt.flags.INVALID is reset) then  
                        error\_out(KDC\_ERR\_POLICY);  
                endif  
                if (tgt.starttime > kdc\_time) then  
                        error\_out(KRB\_AP\_ERR\_NYV);  
                endif  
                if (check\_hot\_list(tgt)) then

Kohl & Neuman                                                 \[Page 100\]  
  
RFC 1510                        Kerberos                  September 1993

error\_out(KRB\_AP\_ERR\_REPEAT);  
                endif  
                tkt := tgt;  
                reset new\_tkt.flags.INVALID;  
        endif

if (req.kdc-options.(any flag except ENC-TKT-IN-SKEY, RENEW,  
                             and those already processed) is set) then  
                error\_out(KDC\_ERR\_BADOPTION);  
        endif

new\_tkt.authtime := tgt.authtime;

if (req.kdc-options.RENEW is set) then  
          /\* Note that if the endtime has already passed, the ticket \*/  
          /\* would have been rejected in the initial authentication  \*/  
          /\* stage, so there is no need to check again here          \*/  
                if (tgt.flags.RENEWABLE is reset) then  
                        error\_out(KDC\_ERR\_BADOPTION);  
                endif  
                if (tgt.renew-till >= kdc\_time) then  
                        error\_out(KRB\_AP\_ERR\_TKT\_EXPIRED);  
                endif  
                tkt := tgt;  
                new\_tkt.starttime := kdc\_time;  
                old\_life := tgt.endttime - tgt.starttime;  
                new\_tkt.endtime := min(tgt.renew-till,  
                                       new\_tkt.starttime + old\_life);  
        else  
                new\_tkt.starttime := kdc\_time;  
                if (req.till = 0) then  
                        till := infinity;  
                else  
                        till := req.till;  
                endif  
                new\_tkt.endtime := min(till,  
                                   new\_tkt.starttime+client.max\_life,  
                                   new\_tkt.starttime+server.max\_life,  
                                   new\_tkt.starttime+max\_life\_for\_realm,  
                                   tgt.endtime);

if ((req.kdc-options.RENEWABLE-OK is set) and  
                    (new\_tkt.endtime < req.till) and  
                    (tgt.flags.RENEWABLE is set) then  
                        /\* we set the RENEWABLE option for later  \*/  
                        /\* processing                             \*/  
                        set req.kdc-options.RENEWABLE;  
                        req.rtime := min(req.till, tgt.renew-till);

Kohl & Neuman                                                 \[Page 101\]  
  
RFC 1510                        Kerberos                  September 1993

endif  
        endif

if (req.rtime = 0) then  
                rtime := infinity;  
        else  
                rtime := req.rtime;  
        endif

if ((req.kdc-options.RENEWABLE is set) and  
            (tgt.flags.RENEWABLE is set)) then  
                set new\_tkt.flags.RENEWABLE;  
                new\_tkt.renew-till := min(rtime,  
                new\_tkt.starttime+client.max\_rlife,  
                new\_tkt.starttime+server.max\_rlife,  
                new\_tkt.starttime+max\_rlife\_for\_realm,  
                tgt.renew-till);  
        else  
                new\_tkt.renew-till := OMIT;  
                              /\* leave the renew-till field out \*/  
        endif  
        if (req.enc-authorization-data is present) then  
                decrypt req.enc-authorization-data  
                        into    decrypted\_authorization\_data  
                        using auth\_hdr.authenticator.subkey;  
                if (decrypt\_error()) then  
                        error\_out(KRB\_AP\_ERR\_BAD\_INTEGRITY);  
                endif  
        endif  
        new\_tkt.authorization\_data :=  
        req.auth\_hdr.ticket.authorization\_data +  
                                 decrypted\_authorization\_data;

new\_tkt.key := session;  
        new\_tkt.crealm := tgt.crealm;  
        new\_tkt.cname := req.auth\_hdr.ticket.cname;

if (realm\_tgt\_is\_for(tgt) := tgt.realm) then  
                /\* tgt issued by local realm \*/  
                new\_tkt.transited := tgt.transited;  
        else  
                /\* was issued for this realm by some other realm \*/  
                if (tgt.transited.tr-type not supported) then  
                        error\_out(KDC\_ERR\_TRTYPE\_NOSUPP);  
                endif  
                new\_tkt.transited  
                   := compress\_transited(tgt.transited + tgt.realm)  
        endif

Kohl & Neuman                                                 \[Page 102\]  
  
RFC 1510                        Kerberos                  September 1993

encode encrypted part of new\_tkt into OCTET STRING;  
        if (req.kdc-options.ENC-TKT-IN-SKEY is set) then  
                if (server not specified) then  
                        server = req.second\_ticket.client;  
                endif  
                if ((req.second\_ticket is not a TGT) or  
                    (req.second\_ticket.client != server)) then  
                        error\_out(KDC\_ERR\_POLICY);  
                endif

new\_tkt.enc-part := encrypt OCTET STRING using  
                        using etype\_for\_key(second-ticket.key),  
                                                      second-ticket.key;  
        else  
                new\_tkt.enc-part := encrypt OCTET STRING  
                        using etype\_for\_key(server.key), server.key,  
                                                      server.p\_kvno;  
        endif

resp.pvno := 5;  
        resp.msg-type := KRB\_TGS\_REP;  
        resp.crealm := tgt.crealm;  
        resp.cname := tgt.cname;  
        resp.ticket := new\_tkt;

resp.key := session;  
        resp.nonce := req.nonce;  
        resp.last-req := fetch\_last\_request\_info(client);  
        resp.flags := new\_tkt.flags;

resp.authtime := new\_tkt.authtime;  
        resp.starttime := new\_tkt.starttime;  
        resp.endtime := new\_tkt.endtime;

omit resp.key-expiration;

resp.sname := new\_tkt.sname;  
        resp.realm := new\_tkt.realm;

if (new\_tkt.flags.RENEWABLE) then  
                resp.renew-till := new\_tkt.renew-till;  
        endif

encode body of reply into OCTET STRING;

if (req.padata.authenticator.subkey)  
                resp.enc-part := encrypt OCTET STRING using use\_etype,

Kohl & Neuman                                                 \[Page 103\]  
  
RFC 1510                        Kerberos                  September 1993

req.padata.authenticator.subkey;  
        else resp.enc-part := encrypt OCTET STRING  
                              using use\_etype, tgt.key;

send(resp);

A.7.  KRB\_TGS\_REP verification  
        decode response into resp;

if (resp.msg-type = KRB\_ERROR) then  
                process\_error(resp);  
                return;  
        endif

/\* On error, discard the response, and zero the session key from  
        the response immediately \*/

if (req.padata.authenticator.subkey)  
                unencrypted part of resp :=  
                        decode of decrypt of resp.enc-part  
                        using resp.enc-part.etype and subkey;  
        else unencrypted part of resp :=  
                        decode of decrypt of resp.enc-part  
                        using resp.enc-part.etype and tgt's session key;  
        if (common\_as\_rep\_tgs\_rep\_checks fail) then  
                destroy resp.key;  
                return error;  
        endif

check authorization\_data as necessary;  
        save\_for\_later(ticket,session,client,server,times,flags);

A.8.  Authenticator generation  
        body.authenticator-vno := authenticator vno; /\* = 5 \*/  
        body.cname, body.crealm := client name;  
        if (supplying checksum) then  
                body.cksum := checksum;  
        endif  
        get system\_time;  
        body.ctime, body.cusec := system\_time;  
        if (selecting sub-session key) then  
                select sub-session key;  
                body.subkey := sub-session key;  
        endif  
        if (using sequence numbers) then  
                select initial sequence number;  
                body.seq-number := initial sequence;  
        endif

Kohl & Neuman                                                 \[Page 104\]  
  
RFC 1510                        Kerberos                  September 1993

A.9.  KRB\_AP\_REQ generation  
        obtain ticket and session\_key from cache;

packet.pvno := protocol version; /\* 5 \*/  
        packet.msg-type := message type; /\* KRB\_AP\_REQ \*/

if (desired(MUTUAL\_AUTHENTICATION)) then  
                set packet.ap-options.MUTUAL-REQUIRED;  
        else  
                reset packet.ap-options.MUTUAL-REQUIRED;  
        endif  
        if (using session key for ticket) then  
                set packet.ap-options.USE-SESSION-KEY;  
        else  
                reset packet.ap-options.USE-SESSION-KEY;  
        endif  
        packet.ticket := ticket; /\* ticket \*/  
        generate authenticator;  
        encode authenticator into OCTET STRING;  
        encrypt OCTET STRING into packet.authenticator  
                             using session\_key;

A.10.  KRB\_AP\_REQ verification  
        receive packet;  
        if (packet.pvno != 5) then  
                either process using other protocol spec  
                or error\_out(KRB\_AP\_ERR\_BADVERSION);  
        endif  
        if (packet.msg-type != KRB\_AP\_REQ) then  
                error\_out(KRB\_AP\_ERR\_MSG\_TYPE);  
        endif  
        if (packet.ticket.tkt\_vno != 5) then  
                either process using other protocol spec  
                or error\_out(KRB\_AP\_ERR\_BADVERSION);  
        endif  
        if (packet.ap\_options.USE-SESSION-KEY is set) then  
                retrieve session key from ticket-granting ticket for  
                 packet.ticket.\{sname,srealm,enc-part.etype\};  
        else  
           retrieve service key for  
           packet.ticket.\{sname,srealm,enc-part.etype,enc-part.skvno\};  
        endif  
        if (no\_key\_available) then  
                if (cannot\_find\_specified\_skvno) then  
                        error\_out(KRB\_AP\_ERR\_BADKEYVER);  
                else  
                        error\_out(KRB\_AP\_ERR\_NOKEY);  
                endif

Kohl & Neuman                                                 \[Page 105\]  
  
RFC 1510                        Kerberos                  September 1993

endif  
        decrypt packet.ticket.enc-part into decr\_ticket  
                                       using retrieved key;  
        if (decryption\_error()) then  
                error\_out(KRB\_AP\_ERR\_BAD\_INTEGRITY);  
        endif  
        decrypt packet.authenticator into decr\_authenticator  
                using decr\_ticket.key;  
        if (decryption\_error()) then  
                error\_out(KRB\_AP\_ERR\_BAD\_INTEGRITY);  
        endif  
        if (decr\_authenticator.\{cname,crealm\} !=  
            decr\_ticket.\{cname,crealm\}) then  
                error\_out(KRB\_AP\_ERR\_BADMATCH);  
        endif  
        if (decr\_ticket.caddr is present) then  
                if (sender\_address(packet) is not in decr\_ticket.caddr)  
                        then error\_out(KRB\_AP\_ERR\_BADADDR);  
                endif  
        elseif (application requires addresses) then  
                error\_out(KRB\_AP\_ERR\_BADADDR);  
        endif  
        if (not in\_clock\_skew(decr\_authenticator.ctime,  
                              decr\_authenticator.cusec)) then  
                error\_out(KRB\_AP\_ERR\_SKEW);  
        endif  
        if (repeated(decr\_authenticator.\{ctime,cusec,cname,crealm\}))  
                then error\_out(KRB\_AP\_ERR\_REPEAT);  
        endif  
        save\_identifier(decr\_authenticator.\{ctime,cusec,cname,crealm\});  
        get system\_time;  
        if ((decr\_ticket.starttime-system\_time > CLOCK\_SKEW) or  
            (decr\_ticket.flags.INVALID is set)) then  
                /\* it hasn't yet become valid \*/  
                error\_out(KRB\_AP\_ERR\_TKT\_NYV);  
        endif  
        if (system\_time-decr\_ticket.endtime > CLOCK\_SKEW) then  
                error\_out(KRB\_AP\_ERR\_TKT\_EXPIRED);  
        endif  
        /\* caller must check decr\_ticket.flags for any pertinent \*/  
        /\* details \*/  
        return(OK, decr\_ticket, packet.ap\_options.MUTUAL-REQUIRED);

A.11.  KRB\_AP\_REP generation  
        packet.pvno := protocol version; /\* 5 \*/  
        packet.msg-type := message type; /\* KRB\_AP\_REP \*/  
        body.ctime := packet.ctime;  
        body.cusec := packet.cusec;

Kohl & Neuman                                                 \[Page 106\]  
  
RFC 1510                        Kerberos                  September 1993

if (selecting sub-session key) then  
                select sub-session key;  
                body.subkey := sub-session key;  
        endif  
        if (using sequence numbers) then  
                select initial sequence number;  
                body.seq-number := initial sequence;  
        endif

encode body into OCTET STRING;

select encryption type;  
        encrypt OCTET STRING into packet.enc-part;

A.12.  KRB\_AP\_REP verification  
        receive packet;  
        if (packet.pvno != 5) then  
                either process using other protocol spec  
                or error\_out(KRB\_AP\_ERR\_BADVERSION);  
        endif  
        if (packet.msg-type != KRB\_AP\_REP) then  
                error\_out(KRB\_AP\_ERR\_MSG\_TYPE);  
        endif  
        cleartext := decrypt(packet.enc-part)  
                     using ticket's session key;  
        if (decryption\_error()) then  
                error\_out(KRB\_AP\_ERR\_BAD\_INTEGRITY);  
        endif  
        if (cleartext.ctime != authenticator.ctime) then  
                error\_out(KRB\_AP\_ERR\_MUT\_FAIL);  
        endif  
        if (cleartext.cusec != authenticator.cusec) then  
                error\_out(KRB\_AP\_ERR\_MUT\_FAIL);  
        endif  
        if (cleartext.subkey is present) then  
                save cleartext.subkey for future use;  
        endif  
        if (cleartext.seq-number is present) then  
                save cleartext.seq-number for future verifications;  
        endif  
        return(AUTHENTICATION\_SUCCEEDED);

A.13.  KRB\_SAFE generation  
        collect user data in buffer;

/\* assemble packet: \*/  
        packet.pvno := protocol version; /\* 5 \*/  
        packet.msg-type := message type; /\* KRB\_SAFE \*/

Kohl & Neuman                                                 \[Page 107\]  
  
RFC 1510                        Kerberos                  September 1993

body.user-data := buffer; /\* DATA \*/  
        if (using timestamp) then  
                get system\_time;  
                body.timestamp, body.usec := system\_time;  
        endif  
        if (using sequence numbers) then  
                body.seq-number := sequence number;  
        endif  
        body.s-address := sender host addresses;  
        if (only one recipient) then  
                body.r-address := recipient host address;  
        endif  
        checksum.cksumtype := checksum type;  
        compute checksum over body;  
        checksum.checksum := checksum value; /\* checksum.checksum \*/  
        packet.cksum := checksum;  
        packet.safe-body := body;

A.14.  KRB\_SAFE verification  
        receive packet;  
        if (packet.pvno != 5) then  
                either process using other protocol spec  
                or error\_out(KRB\_AP\_ERR\_BADVERSION);  
        endif  
        if (packet.msg-type != KRB\_SAFE) then  
                error\_out(KRB\_AP\_ERR\_MSG\_TYPE);  
        endif  
        if (packet.checksum.cksumtype is not both collision-proof  
                                             and keyed) then  
                error\_out(KRB\_AP\_ERR\_INAPP\_CKSUM);  
        endif  
        if (safe\_priv\_common\_checks\_ok(packet)) then  
                set computed\_checksum := checksum(packet.body);  
                if (computed\_checksum != packet.checksum) then  
                        error\_out(KRB\_AP\_ERR\_MODIFIED);  
                endif  
                return (packet, PACKET\_IS\_GENUINE);  
        else  
                return common\_checks\_error;  
        endif

A.15.  KRB\_SAFE and KRB\_PRIV common checks  
        if (packet.s-address != O/S\_sender(packet)) then  
            /\* O/S report of sender not who claims to have sent it \*/  
            error\_out(KRB\_AP\_ERR\_BADADDR);  
        endif  
        if ((packet.r-address is present) and  
            (packet.r-address != local\_host\_address)) then

Kohl & Neuman                                                 \[Page 108\]  
  
RFC 1510                        Kerberos                  September 1993

/\* was not sent to proper place \*/  
                error\_out(KRB\_AP\_ERR\_BADADDR);  
        endif  
        if (((packet.timestamp is present) and  
             (not in\_clock\_skew(packet.timestamp,packet.usec))) or  
            (packet.timestamp is not present and timestamp expected))  
                then error\_out(KRB\_AP\_ERR\_SKEW);  
        endif  
        if (repeated(packet.timestamp,packet.usec,packet.s-address))  
                then error\_out(KRB\_AP\_ERR\_REPEAT);  
        endif  
        if (((packet.seq-number is present) and  
             ((not in\_sequence(packet.seq-number)))) or  
            (packet.seq-number is not present and sequence expected))  
                then error\_out(KRB\_AP\_ERR\_BADORDER);  
        endif  
        if (packet.timestamp not present and  
            packet.seq-number not present) then  
                error\_out(KRB\_AP\_ERR\_MODIFIED);  
        endif

save\_identifier(packet.\{timestamp,usec,s-address\},  
                        sender\_principal(packet));

return PACKET\_IS\_OK;

A.16.  KRB\_PRIV generation  
        collect user data in buffer;

/\* assemble packet: \*/  
        packet.pvno := protocol version; /\* 5 \*/  
        packet.msg-type := message type; /\* KRB\_PRIV \*/

packet.enc-part.etype := encryption type;

body.user-data := buffer;  
        if (using timestamp) then  
                get system\_time;  
                body.timestamp, body.usec := system\_time;  
        endif  
        if (using sequence numbers) then  
                body.seq-number := sequence number;  
        endif  
        body.s-address := sender host addresses;  
        if (only one recipient) then  
                body.r-address := recipient host address;  
        endif

Kohl & Neuman                                                 \[Page 109\]  
  
RFC 1510                        Kerberos                  September 1993

encode body into OCTET STRING;

select encryption type;  
        encrypt OCTET STRING into packet.enc-part.cipher;

A.17.  KRB\_PRIV verification  
        receive packet;  
        if (packet.pvno != 5) then  
                either process using other protocol spec  
                or error\_out(KRB\_AP\_ERR\_BADVERSION);  
        endif  
        if (packet.msg-type != KRB\_PRIV) then  
                error\_out(KRB\_AP\_ERR\_MSG\_TYPE);  
        endif

cleartext := decrypt(packet.enc-part) using negotiated key;  
        if (decryption\_error()) then  
                error\_out(KRB\_AP\_ERR\_BAD\_INTEGRITY);  
        endif

if (safe\_priv\_common\_checks\_ok(cleartext)) then  
            return(cleartext.DATA, PACKET\_IS\_GENUINE\_AND\_UNMODIFIED);  
        else  
                return common\_checks\_error;  
        endif

A.18.  KRB\_CRED generation  
        invoke KRB\_TGS; /\* obtain tickets to be provided to peer \*/

/\* assemble packet: \*/  
        packet.pvno := protocol version; /\* 5 \*/  
        packet.msg-type := message type; /\* KRB\_CRED \*/

for (tickets\[n\] in tickets to be forwarded) do  
                packet.tickets\[n\] = tickets\[n\].ticket;  
        done

packet.enc-part.etype := encryption type;

for (ticket\[n\] in tickets to be forwarded) do  
                body.ticket-info\[n\].key = tickets\[n\].session;  
                body.ticket-info\[n\].prealm = tickets\[n\].crealm;  
                body.ticket-info\[n\].pname = tickets\[n\].cname;  
                body.ticket-info\[n\].flags = tickets\[n\].flags;  
                body.ticket-info\[n\].authtime = tickets\[n\].authtime;  
                body.ticket-info\[n\].starttime = tickets\[n\].starttime;  
                body.ticket-info\[n\].endtime = tickets\[n\].endtime;  
                body.ticket-info\[n\].renew-till = tickets\[n\].renew-till;

Kohl & Neuman                                                 \[Page 110\]  
  
RFC 1510                        Kerberos                  September 1993

body.ticket-info\[n\].srealm = tickets\[n\].srealm;  
                body.ticket-info\[n\].sname = tickets\[n\].sname;  
                body.ticket-info\[n\].caddr = tickets\[n\].caddr;  
        done

get system\_time;  
        body.timestamp, body.usec := system\_time;

if (using nonce) then  
                body.nonce := nonce;  
        endif

if (using s-address) then  
                body.s-address := sender host addresses;  
        endif  
        if (limited recipients) then  
                body.r-address := recipient host address;  
        endif

encode body into OCTET STRING;

select encryption type;  
        encrypt OCTET STRING into packet.enc-part.cipher  
        using negotiated encryption key;

A.19.  KRB\_CRED verification  
        receive packet;  
        if (packet.pvno != 5) then  
                either process using other protocol spec  
                or error\_out(KRB\_AP\_ERR\_BADVERSION);  
        endif  
        if (packet.msg-type != KRB\_CRED) then  
                error\_out(KRB\_AP\_ERR\_MSG\_TYPE);  
        endif

cleartext := decrypt(packet.enc-part) using negotiated key;  
        if (decryption\_error()) then  
                error\_out(KRB\_AP\_ERR\_BAD\_INTEGRITY);  
        endif  
        if ((packet.r-address is present or required) and  
           (packet.s-address != O/S\_sender(packet)) then  
            /\* O/S report of sender not who claims to have sent it \*/  
            error\_out(KRB\_AP\_ERR\_BADADDR);  
        endif  
        if ((packet.r-address is present) and  
            (packet.r-address != local\_host\_address)) then  
                /\* was not sent to proper place \*/  
                error\_out(KRB\_AP\_ERR\_BADADDR);

Kohl & Neuman                                                 \[Page 111\]  
  
RFC 1510                        Kerberos                  September 1993

endif  
        if (not in\_clock\_skew(packet.timestamp,packet.usec)) then  
                error\_out(KRB\_AP\_ERR\_SKEW);  
        endif  
        if (repeated(packet.timestamp,packet.usec,packet.s-address))  
                then error\_out(KRB\_AP\_ERR\_REPEAT);  
        endif  
        if (packet.nonce is required or present) and  
           (packet.nonce != expected-nonce) then  
                error\_out(KRB\_AP\_ERR\_MODIFIED);  
        endif

for (ticket\[n\] in tickets that were forwarded) do  
                save\_for\_later(ticket\[n\],key\[n\],principal\[n\],  
                               server\[n\],times\[n\],flags\[n\]);  
        return

A.20.  KRB\_ERROR generation

/\* assemble packet: \*/  
        packet.pvno := protocol version; /\* 5 \*/  
        packet.msg-type := message type; /\* KRB\_ERROR \*/

get system\_time;  
        packet.stime, packet.susec := system\_time;  
        packet.realm, packet.sname := server name;

if (client time available) then  
                packet.ctime, packet.cusec := client\_time;  
        endif  
        packet.error-code := error code;  
        if (client name available) then  
                packet.cname, packet.crealm := client name;  
        endif  
        if (error text available) then  
                packet.e-text := error text;  
        endif  
        if (error data available) then  
                packet.e-data := error data;  
        endif

Kohl & Neuman                                                 \[Page 112\]

转载于:https://www.cnblogs.com/MaxWoods/archive/2007/08/14/855037.html

[kerberos_MIT.EDU]: mailto:kerberos@MIT.EDU
[kerberos-request_MIT.EDU]: mailto:kerberos-request@MIT.EDU
[info-kerberos_MIT.EDU]: mailto:info-kerberos@MIT.EDU
[jtkohl_zk3.dec.com]: mailto:jtkohl@zk3.dec.com
[bcn_isi.edu]: mailto:bcn@isi.edu