12 Directory Security Concepts

12.1 Data Integrity and Oracle Internet Directory

Oracle Internet Directory ensures that data has not been modified, deleted, or replayed during transmission by using Secure Sockets Layer (SSL). SSL generates a cryptographically secure message digest—through cryptographic checksums using either the MD5 algorithm or the Secure Hash Algorithm (SHA)—and includes it with each packet sent across the network.

12.2 Data Privacy and Oracle Internet Directory

Oracle Internet Directory ensures that data is not disclosed during transmission by using public key encryption available with SSL. In public-key encryption, the sender of a message encrypts the message with the public key of the recipient. Upon delivery, the recipient decrypts the message using the recipient's private key. Specifically, Oracle Internet Directory supports two levels of encryption available through SSL:

• DES40

  The DES40 algorithm, available internationally, is a variant of DES in which the secret key is preprocessed to provide 40 effective key bits. It is designed for use by customers outside the USA and Canada who want to use a DES-based encryption algorithm. This feature gives commercial customers a choice in the algorithm they use, regardless of their geographic location.

• RC4_40

  Oracle has obtained license to export the RC4 data encryption algorithm with a 40-bit key size to virtually all destinations where other Oracle products are available. This makes it possible for international corporations to safeguard their entire operations with fast cryptography.

  
  

  See Also: Chapter 13, "Secure Sockets Layer (SSL) and the Directory" for more information about SSL

  
  

12.3 Authorization in Oracle Internet Directory

Authorization is the permission given to a user, program, or process to access an object or set of objects. When directory operations are attempted within a directory session, the directory server ensures that the user has the permissions to perform those operations. If the user does not have the permissions, then the directory server disallows the operation. The directory server protects directory data from unauthorized operations by directory users by using access control information.

Access control information is the directory metadata that captures the administrative policies relating to access control. This information is stored in Oracle Internet Directory as user-modifiable operational attributes, each of which is called an access control item (ACI).

Typically, a list of these ACI attribute values, called an access control list (ACL), is associated with directory objects. The attribute values on that list represent the permissions that various directory user entities (or subjects) have on a given object.

An ACI consists of:

• The object to which you are granting access

• The entities or subjects to whom you are granting access

• The kind of access you are granting

Access control policies can be prescriptive, that is, their security directives can be set to apply downward to all entries at lower positions in the directory information tree (DIT). The point from which such an access control policy applies is called an access control policy point (ACP).

ACIs are represented and stored as text strings in the directory. These strings must conform to a well-defined format, called the ACI directive format. Each valid value of an ACI attribute represents a distinct access control policy.

The following features of directory access control can be used by applications running in a hosted environment.

• Prescriptive access control

  Enables the service provider to specify access control lists (ACLs) for a collection of directory objects, instead of having to state the policies for each individual object. This feature simplifies the administration of access control, especially in large directories where many objects are governed by identical or similar policies.

• Hierarchical access control administration model

  Enables the service provider to delegate directory administration to hosted companies. The realm could in turn delegate further if necessary.

• Administrative override control for delegated domains

  Enables the service provider to perform diagnosis and recovery from unintentional account lockout or accidental security exposure.

• Dynamic evaluation of access control entities

  Enables subtree administrators to identify both subjects and objects in terms of their namespace and their association with other objects in the directory. For example, the administrator of one realm can allow only a user's manager to update that user's salary attribute. The administrator of another realm can establish and enforce a different policy regarding salary attributes.

12.4 Authentication in Oracle Internet Directory

Authentication is the process by which the directory server establishes the true identity of the user connecting to the directory. It occurs when an LDAP session is established by means of the ldapbind operation. Thus every session has an associated user identity.

To verify the identities of users, hosts, and clients, Oracle Internet Directory enables three general kinds of authentication, and these are described in these topics:

• Direct Authentication

• Indirect Authentication

• External Authentication

12.4.2 Indirect Authentication

Indirect authentication occurs through any entity that has credentials in the directory—for example, an application such as the Oracle Internet Directory Self-Service Console, or a middle tier such as a firewall or a RADIUS server. The application or middle tier becomes a proxy user. A proxy user has the privilege to impersonate an end user, performing on that user's behalf those operations for which that user has privileges.

Figure 12-1 and the accompanying text explain how indirect authentication takes place.

Figure 12-1 Indirect Authentication

Indirect authentication takes place as follows:

1. The end user sends to the application or middle tier a request containing a query to Oracle Internet Directory. The application or middle tier authenticates the end user.

2. The application or middle tier binds to the directory.

3. The application or middle tier performs a second bind, this time using the DN of the end user. It does not enter the end user's password.

4. The directory server recognizes this second bind as an attempt by the application or middle tier to switch to the end user's identity. It trusts the authentication granted to the end user by the application or middle tier, but must verify that the application or middle tier has the right to be the proxy for this user. It checks to see whether the ACP governing the end user entry gives this application or middle tier the proxy right for this end user.

   • If the end user entry does give the application or middle tier the necessary proxy right, then the directory server changes the authorization identity to that of the end user. All subsequent operations occur as if that end user had connected directly to the server and had been directly authenticated.

   • If the end user entry does not give the application or middle tier the necessary proxy right, then the directory server returns an "Insufficient Access" error message.

     
     

     See Also: Operations: What Access Are You Granting?

     
     

The directory server can, in the same session, authenticate and authorize other end users. It can also switch the session from the end user to the application or middle tier that opened the session.

To close the session, the application or middle tier sends an unbind request to the directory server.

For example, suppose you have:

• A middle tier that binds to the directory as cn=User1, which has proxy access on the entire directory

• An end user that can bind to the directory as cn=User2

When this end user sends to the application or middle tier a request containing a query to the directory, the application or middle tier authenticates the end user. The middle tier service then binds to the directory by using its own identity, cn=User1, then performs a second bind, this time by using only the DN of the end user, cn=User2. The Oracle directory server recognizes this second bind as an attempt by the proxy user to impersonate the end user. After the directory server verifies that cn=user1 has proxy access, it allows this second bind to succeed. It does not require any further validation of the end-user DN, such as a password. For the rest of the session, all LDAP operations are access-controlled as if cn=User2 were performing them.

If one user is being serviced by an application, and another user subsequently requests a service of that same application, then the application can establish a new connection and proceed as previously described without disrupting that prior session. If, however, no prior user is still being serviced, then the existing established connection can be re-used again and again without the need for a new connection.

12.5 Protection of User Passwords for Directory Authentication

Oracle Internet Directory can protect a user's directory password by storing it in the userPassword attribute as a one-way hashed value. You select the hashing algorithm you want to use. Storing passwords as one-way hashed values—rather than as encrypted values—more fully secures them because a malicious user can neither read nor decrypt them.

12.6 Password Policies in Oracle Internet Directory

A password policy is a set of rules governing how passwords are used. When a user attempts to bind to the directory, the directory server ensures that the password meets the various requirements set in the password policy.

When you establish a password policy, you set the following types of rules, to mention just a few:

• The maximum length of time a given password is valid

• The minimum number of characters a password must contain

• The number of numeric characters required in a password

  
  

  See Also: Chapter 15, "Password Policies in Oracle Internet Directory" for a fuller description of the rules you set when establishing password policies

  
  

12.7 Authentication by Using Simple Authentication and Security Layer (SASL)

The section "Direct Authentication" introduced the use of SASL within an Oracle Internet Directory environment. This section describes more fully how SASL works. It contains these topics:

• How a SASL-Enabled Client Authenticates to a Directory Server by Using Digest-MD5

• How a SASL-Enabled Client Authenticates to a Directory Server by Using External Authentication

How a SASL-Enabled Client Authenticates to a Directory Server by Using Digest-MD5

When a SASL-enabled client seeks Digest-MD5 authentication to a server, the authentication process is as follows:

1. The directory server sends to the LDAP client a digest-challenge that includes various Digest-MD5 authentication options that it supports and a special token.

2. The client selects an authentication option, then sends a digest-response to the server indicating the option it has selected. The response includes some secure tokens and a client credential in encrypted format. This allows it to authenticate itself to the server.

3. The directory server then decrypts and verifies the client credential from the response.

How a SASL-Enabled Client Authenticates to a Directory Server by Using External Authentication

Oracle Internet Directory provides SASL-external authentication over an SSL connection in which both client and server authenticate themselves to each other by providing certificates. The DN is derived from the client certificate used in the SSL network negotiation.

When a client seeks authentication to a directory server by using an external authentication mechanism such as SSL, the authentication process is as follows:

1. The client sends an initial message with the authorization identity.

2. The directory server uses information external to SASL to determine whether the client can validly authenticate as the authorization identity. If the client can validly authenticate, then the directory server indicates successful completion of the authentication exchange. Otherwise, the directory server indicates failure.

The system providing the external information may be IPsec or SSL/TLS. The authorization identity is derived as follows:

• In case of exact match, the authorization identity is derived from the client authentication credentials in the system providing external authentication—for example, the client SSL certificate.

• In case of exact match, the authorization identity is derived from the client authentication credentials in the system providing external authentication—for example, the client SSL certificate

If the client sends an empty string as the authorization identity, then the authorization identity is derived from the client authentication credentials in the system providing external authentication—for example, the SSL certificate.