In a former article, we introduced Business Intelligence Monitoring within SAP Solution Manager. We chose in this post to explain in more details how SAP Business Intelligence system provides historical and predictive views of data to enable more effective decision making.
From Star Schema to Info Cube
The objective of data warehousing is to gather structured data from diverse sources to facilitate data retrieval for analytical processing and decision making. In SAP Business Intelligence system, we can find a large amount of Info Cubes organized into Info Areas just like files organized into directories in an operating system; for example, specific cube for workload analysis and specific cube for exception analysis.
Technically, Info Cubes are designed according to a special database design technique called a Star Schema which consists of a fact table to reference multiple dimensions (see figure below). Indeed, the fact table contains all primary keys of associated dimension tables; each of which has a single primary key that uniquely identifies each member record (row). Another important concept is Info Object (characteristics and key figures) which represents the smallest units of BI. That is, Info Cubes are composed of Info Objects divided into characteristics, units, and time characteristics (forming the basis of the key fields) and the key figures which form the data part of the fact table of Info Cubes.
By adopting the star schema design, querying large data sets are optimized and used in data warehouses to support business intelligence and analytic applications. Namely, you can analyze the dataset of the Business Information Warehouse by creating queries for our Info Providers using the BEx Query Designer that can select and combine Info Objects or reusable structures in a query. This desktop tool enables you to choose haw to navigate and evaluate the data in the selected Info Provider. Finally, you can use the Web Template (an HTML document that is used to define the structure of a Web application) as a starting point for creating a Web application.
The system monitoring can also provide metrics about system performance classified into three categories: availability, performance and exceptions; where performance specifies the average response time and exceptions provide more specific error patterns that can be customized by the customer based on customer need or experience.
In a word, the key success of BI monitoring remains in its capability to perform monitoring at two levels: system level monitoring for the involved technical systems and monitoring of important BI objects like queries, templates and Business Object jobs. All this is about ensuring that the involved technical systems and components are working as expected.
Author- Salma MATOUSSI