Bookshelf v7.5.3: About Server Definitions

About Server Definitions

Siebel Marketing can access multiple servers (data sources) that hold customer data. When you create server definitions, you provide the ODBC connection information to access selected tables in each of your supported relational database management system (RDBMS) sources. Siebel Marketing Server also connects to the Marketing Repository tables in the same way that it accesses other data sources.

Because Siebel Marketing can extract information from multiple servers simultaneously and can join tables that are in separate RDBMSs, you can create as many data source server definitions as necessary. For example, if customer data and transactions data are in separate physical databases, you might create two Data Source server definitions. For more information, see Creating Server Definitions.

The Marketing Repository is a set of tables that record the information to support Marketing Server processes such as segment criteria, attributes, measures, and data mapping metadata. In addition, it handles processes for counts, snapshots, and lists. The Marketing Repository consists of a set of system tables in the Siebel database, thus it shares the same system DSN as the main Siebel Database. Although the Marketing Repository record can only be created once, the values for the reference name, DSN, user name, and password can be updated at any time.

As part of initialization, an administrator needs to set up a server of type Marketing Repository that points to the Siebel transactional database so that the Marketing Server can map and join temporary tables (attribute tables) that are created by the Marketing Server during data synchronization. There can be only one Marketing Repository server and the user name and password must match the Siebel OLTP table owner.

Marketing Server. Refers to the metadata contained in the Marketing Repository tables before performing tasks such as a records extract, count, or data retrieval.

Test Server. If you are testing new measures and attributes and your data warehouse has many millions of records, you may benefit from creating a special test server that loads only a fraction of your total records. Creating a test server can save time when designing and testing measures and attributes, as well as freeing the main server for other work.