SAP APO Architecture

SAP APO™ consists technically of three parts: the database, the SAP BI™ data mart and the live cache. The SAP BI™ data mart consists of info cubes. The live cache is basically a huge main memory where the planning and the scheduling relevant data are kept to increase the performance for complex calculations. Though there is technically only one live cache per installation, the data is stored in three different ways depending on the application:

as a number per time period (month, week, day) and key figure (time series),

as an order with a category, date and exact time (hour: minute: second) and

as a quantity with a category and a date in the ATP time series.

We will refer to the according parts of the live cache as time series live cache, order live cache and ATP time series live cache.

Demand planning uses much of the SAP BI™ functionality and relies on the info cubes as data interface to any other system – SAP ERP™, SAP SEM™ or flat file. Therefore the historical data is always persistent in an info cube. For processing the data is written into the time series live cache. SNP and PP/DS use mainly the order live cache, though SNP is able to access the time series live cache as well, since there are many structural similarities between DP and SNP. ATP at last relies only on the ATP time series live cache. This way of data storage implies a certain redundancy, because orders are stored both in the order live cache and in the ATP time series live cache. The data model is however quite different, and the redundant data storage enables better performance for the applications. TP/VS finally uses the order live cache to reference other orders. Figure 3.1 shows how the SAP APO™ modules use the live cache and the data integration for transactional data from SAP ERP™ to SAP APO™.