Siebel Data Warehouse Installation and Administration Guide > DAC Quick Start >

Tailoring Execution Plans for Your Needs

Execution plans are groups of subject areas that are used to execute ETL processes. For more information about the repository objects that are the components that make up execution plans, see About DAC Repository Objects. For information about DAC functionality, see DAC Functional Reference. For detailed information about customizing execution plans, see Customizing the Siebel Data Warehouse.

In designing an execution plan, you should consider the following questions:

• Tables. Which tables need to be populated for the data warehouse. From which tables does your organization source data?
• Indices. Do the target tables have the correct indices defined?
• Tasks. Are the tasks that load this table defined?
• Dependencies. What does a task require to be run?
• Table groups. Do table groups need to be built?
• Subject areas. Do the subject areas cover all the relevant table groups?
• Execution plans. Which subject areas need to be refreshed and how often and at what time frequency?

When an execution plan is run, the DAC server uses the following logic to identify the tasks that need to be run and their order of execution:

1. Initial selection of tables.
   1. Find all the subject areas.
   2. Find all the table groups that belong to the subject areas.
   3. Find all the tables that belong to the table groups.
2. Recursive selection of related tables.

   Recursively find all the table groups and their tables with the above list of tables. For a description of this process, see About the DAC Table Groups Tab.

3. Initial selection of tasks.

   Find all the tasks that load into the tables selected above, that is, tasks whose target table is one of the selected tables (the tasks that are non-optional).

4. Recursive selection of all tasks.
   1. Depending on the non-optional dependency rules, recursively figure out the prerequisite tasks.
   2. Depending on the optional dependency rules, figure out the optional tasks.
5. Filter tasks based on the database connections that will participate in the ETL.
6. Selection of dependencies.

   Load dependencies to identify the execution order.

7. Identify the task details.

   Iterate through all the tasks selected and compute the task details based on whether the target table is loaded in full mode or incremental mode.

8. Identify the list of source tables for change capture, create tasks for change capture, and sync for Siebel source tables.
9. At the end of the ETL process, the DAC server updates refresh timestamps for all source tables (primary or auxiliary) and target tables. The refresh timestamps are the database timestamps. The ETL history timestamps (Start and End timestamps) are the DAC server timestamps.

In the DAC, the Design view allows you to view and modify the relationships among the repository objects discussed above. For more information about the Design view, see About the DAC Design View.