This chapter provides an overview of the design of a Process Warehouse in PPM and then gives a brief explanation of each individual step. It is assumed that the necessary installation and configuration of the PPM system has been completed.
The design of a Process Warehouse comprises the following steps:
First, the actual data is extracted from the source system and provided in the form of process fragments for importing.
The process fragments are imported.
The next step involves searching all imported data for the fragments belonging to each business event and merging them into a process instance. Object attributes are copied to the process instance. When merging the process fragments, information about the actual user can be made anonymous.
The process instances generated are then classified: process instances of the same kind are assigned to process types, which are in turn summarized into process type groups.
For each process instance, the defined measures are calculated and stored in info cubes.
Exceeding of planned values is checked and, if necessary, signaled.
Designing a Process Warehouse therefore involves generating process fragments from sequences of transactions and document flows from the source system and merging them into process instances. These process instances are used as reference objects for analyses and assessments of process performance. They can be displayed as process models in PPM.
The event-driven process chain (EPC) consisting of a chain of objects is used to represent a process instance. Both objects and the process instance itself can have attributes, in which the instance data is saved. This data is used to classify the process instances into types and to calculate the measures. The calculated measures are in turn saved in attributes of the objects and process instances. The attributes are therefore the actual carriers of information in the PPM system. The available attributes are defined in the PPM configuration. The configuration is composed of default attributes, such as Process identifier and End time and freely definable system-specific attributes.
Additional information about EPCs:
An event-driven process chain (EPC) is a model type developed by Prof. Scheer to graphically describe the chronological sequence of a performance delivery process. It is based on the following assumptions:
Each activity within a process is triggered by a commercially relevant change of state of an information object. Each activity can result in a commercially relevant change of state of an information object.
The state of a business-relevant information object is defined graphically by an object of the Event type.
Objects of the Function type are used for the graphic representation of activities. Linking events and functions in series and connecting these objects with directed connections represents the control flow of the process graphically.
As an event can trigger several functions and, conversely, a function can have several events as its result, AND, OR or exclusive OR connections (rules) are inserted at these branches. They illustrate the logical relationship that exists between the sequenced objects.
Organizational units describe the groups of users executing a function.
See also
Information on the runppmimport command line program is available in the PPM Operation Guide.