Setting Up Security for BEA eLink TCP for Tuxedo

 

The eLink for Mainframe TCP for Tuxedo (hereafter referenced as eLink TCP for Tuxedo) component supports security features that allows a requester from Tuxedo to pass a user ID requirement through the OTMA or CICS server interfaces for verification through system security, such as RACF.

This document explains the following security topics:

• Security Checking from Tuxedo to Mainframe

• Security Checking from Mainframe to Tuxedo

• Setting Up Security

• Sample Security Files

• Data Area Security

 

---

Security Checking from Tuxedo to Mainframe

The following figure illustrates the process flow for security verifications from eLink TCP for Tuxedo to a mainframe.

Figure 4-1 Security Checking for Tuxedo to Mainframe Transactions

 

1. When the client program performs a tpinit(), the user's Tuxedo identity is validated against the tpusr file.

2. When the client program issues a tpcall() or tpacall(), Tuxedo verifies (against the tpacl file) that the user is authorized to invoke the gateway service.

3. When the gateway establishes the initial connection, connection security information (specified as RMTNAME and PASSWORD in the GWICONFIG file) is passed from the eLink TCP for Tuxedo gateway to the remote gateway. If the RMTNAME and PASSWORD values match the values configured on the remote gateway, the connection is established.

   With each request, the eLink TCP for Tuxedo gateway passes the user's Tuxedo identity to the remote gateway.

   Note: To pass authority checking, the user's Tuxedo identity must match the mainframe user ID exactly.

4. The remote mainframe gateway initiates a proxy to act on behalf of the specified user ID.

5. The proxy calls the specified service using system security to check authorization.

 

---

Security Checking from Mainframe to Tuxedo

The following figure illustrates the process flow for security verifications from a mainframe to eLink TCP for Tuxedo.

Figure 4-2 Security Checking for Mainframe to Tuxedo Transactions

 

1. The user ID, established at mainframe log in, is checked by system security to verify that the user has permission to start a client transaction.

2. The user ID is checked by system security to verify that the user has permission to send a request to the gateway.

3. With each request, the gateway passes the user ID to the Tuxedo gateway.

   Note: To pass authority checking, the user's Tuxedo identity must match the mainframe user ID exactly.

4. The eLink TCP for Tuxedo gateway maps the mainframe user ID to a Tuxedo user ID and issues the service request on behalf of that user.

5. The Tuxedo server performs access checks (based on the tpacl file) to verify that the user has access to the requested service.

 

---

Setting Up Security

The eLink TCP for Tuxedo product supports two methods for providing security:

• Tuxedo security plug-in

• Built-in Tuxedo security

Tuxedo Security Plug-in

The Tuxedo security plug-in enables the customization of the security functions, including the use of alternate implementations. The Tuxedo security plug-in is set up during Tuxedo plug-in configuration. Refer to the Tuxedo documentation for specific information about this feature.

Built-in Tuxedo Security

Built-in Tuxedo security is used when custom security plug-ins are not implemented.

To enable the built-in Tuxedo security feature, complete the following tasks.

1. Code SECURITY in the BEA Tuxedo UBBCONFIG file. Refer to the BEA Tuxedo Administration Guide for more information.

2. Set up user, group, and ACL files. Refer to the BEA Tuxedo Administrator's Guide for more information.

   Note: The user information in these files must match in the BEA Tuxedo and the mainframe environments or a security violation occurs.

3. Code the security parameter in your eLink TCP for Tuxedo configuration file (GWICONFIG). For GWICONFIG syntax and parameter definitions, refer to the Configuring BEA eLink TCP for Tuxedo section.

Sample Security Files

Part of the process for setting up security for eLink TCP requires you to have user, group, and ACL files. The following sections include these sample files.

User Files

The following sample is a user file that includes user names, encrypted passwords, a user ID number, group number, and a client name.

Listing 4-1 Sample User (tpusr) File

#illen:w2ZMOKeJmiU0M:1:0:TPCLTNM,someguy::
#illen:0YzvQeqzcNz56:1:0:TPCLTNM,*::
#eke:x3vG37eOqh0XE:2:0:TPCLTNM,*::
#illen:0YzvQeqzcNz56:1:1:TPCLTNM,*::
#illen:0YzvQeqzcNz56:1:2:TPCLTNM,*::
john:x3vG37eOqh0XE:2:1:TPCLTNM,*::
jim:0YzvQeqzcNz56:1:1:TPCLTNM,*::
richard:IxqosKHu5Q3BA:3:1:TPCLTNM,*::
JDOE:zBMWVUBNNBVgo:4:0:TPCLTNM,*::
smith:ULfRJzAeyGAD2:5:0:TPCLTNM,*::

Lines that begin with the pound sign (#) are users that have been changed or deleted by tpusrmod or tpusrdel.

Group File

The following sample is a group file that specifies the names and indexes of groups.

Note: The tpgrp file is only necessary when specifying ACL or MANDATORY_ACL modes for security. If you specify USER_AUTH for security, you can assign users to groups, but they do not correlate to the groups used for security by the remote system.

Listing 4-2 Sample Group (tpgrp) File

good::1:
bad::2:

ACL File

The tpacl file correlates a group and the services to which that group has access. In the tpacl file, the first field specifies what is protected, the second field specifies the type of object being protected (specified in the first field), and the third field specifies the group that has access to the object.

In the following example, only users in group 1 (john, jim, richard) can access TOLOWER, and only users in group 2 can access TOUPPER.

Note: The tpacl file is only necessary when specifying ACL or MANDATORY_ACL modes for security.

Listing 4-3 Sample ACL (tpacl) File

TOLOWER:SERVICE:1:
TOUPPER:SERVICE:2:

 

---

Data Area Security

The eLink TCP for Tuxedo gateway provides data area security which is a specialized security protocol for the following cases:

• User information is propagated across multiple Tuxedo domain boundaries

• A remote or local service requires a user's LTERM information

In these cases, a client's user ID, group name, and LTERM can be specified in the data area of a request. For Tuxedo clients, user information specified in the data area is verified by the remote gateway in the usual manner. For remote clients, remote user information is placed in the data area fields by the local gateway to be used by Tuxedo services. In this case, the remote client does not have to populate these fields, but must allocate space for them in the data area.

Enabling Data Area Security

Complete the following tasks to enable data area security.

1. Add fields to the user's data area on the local and remote hosts. These fields are passed to and from the mainframe host. For the field formats, refer to Listing 4-4.

2. Set WRAP=TPSD in the FOREIGN section corresponding to the remote host in the GWICONFIG file. For syntax and parameter definitions for the FOREIGN section of the GWICONFIG file, refer to the Defining the FOREIGN Section of the section.

3. Populate the data area with the user information before sending a request to a remote service.

4. The remote user's information is populated into the data area when a request is received for a local service.

   Note: If using a VIEW data format, allocate the extra fields before the application data as defined in Listing 4-4. If using the STRING data format, allocate 24 additional bytes at the beginning of the string to be used for the security fields.

Format

The user data area fields in C use the following format.

Listing 4-4 Syntax for C User Data Area Fields

struct da_security {
       char uname[8];  /*user name*/
       char group[8];  /*user group*/
       char lterm[8];  /*terminal id*/
       /*user data is appended here*/
}