Interface to remote servers

The interface between the server and remote servers is handled by the Open Client software, Client-Library™. The Client-Library features that are used to implement the interface are dependent upon the class of server with which Component Integration Services is interacting.

For example, if the server class is direct_connect (access_server), a number of features such as cursor and dynamic requests are used. These features are not used by a server of class db2.

Before the server can interact with a remote server, you need to configure the following:

Directory services

Before accessing remote tables with Component Integration Services, you must either have access to LDAP directory services, or an interfaces file (sql.ini file on Windows NT). For information on setting up directory services, see the configuration documentation for your platform. You may wish to refer to Appendix A which serves as a basic tutorial for Component Integration Services users.

Remote server definition

Remote servers are defined by means of the stored procedure sp_addserver. This procedure is documented in the Adaptive Server Reference Manual. When using release 12.0 or greater you must name the local server.

Logging into remote servers

Once the remote server has been configured, login information must be provided. By default, the server uses the names and passwords of its clients whenever it connects to a remote server on behalf of those clients. However, this default can be overridden by the use of the stored procedure sp_addexternlogin. This procedure allows a system administrator to define the name and password for each user who connects to a remote server.

Using connect to server_name, you can verify that the server configuration is correct. This command establishes a passthrough mode connection to the remote server. Passthrough mode allows clients to communicate with remote servers in native syntax. This passthrough mode remains in effect until you issue a disconnect command.

Defining remote objects

Once a remote server has been properly configured, objects in that remote server cannot be accessed as tables until a mapping between them and a local object (proxy table) has been established.

You can create new tables on remote servers, and you can define the schema for an existing object in a remote server. The procedures for both are similar.

You can use one of two methods for defining the storage location of remote objects:

  1. Define the storage location of individual objects

  2. Define the default location of all objects in a database