FIXME This page has been moved to a new location. Below, you'll find an excerpt from the new page.


Performance Sequence Diagram

The purpose of the performance sequence diagram plug-in is to provide you with a means to assess the performance of your processes. This plug-in can especially be of help if you want to know what behavior in your processes is common, what behavior is rare and what behavior may result in extreme situations (e.g. instances with extremely high throughput times). What is more, it can assist you in finding the causes of such behavior.

In the logs of today's information systems, a lot of information is recorded. This plug-in will focus on information that is stored on the audit trail entry level of the logs. In an audit trail entry, information about one event is recorded.

Common information stored at the audit trail entry level includes:

  • TaskID: Name of the task involved in the event.
  • Event type: Type of the event, e.g. schedule, start, complete.
  • Timestamp: Timestamp at which the event took place.
  • Originator: Name of the originator involved in the event.

Though one could also think of other information that may be stored here, such as:

  • Department: Name of the department at which the event was handled.
  • Part: Name of the part of the case that is currently being handled. For instance, suppose the process under consideration is the production line of a car. A case is then the production of one car. Parts currently handled may be: doors, wheels, windows etc.
  • Country: Name of the country in which the event took place.

This plug-in allows you to focus on a certain data-element (such as taskID, originator, department) and see how transfer of work between instances of the selected data-element (for instance in case of originator this can be Nick or John) takes place for each case. The transfer of work is displayed in a sequence diagram, in which you can easily see which data-elements cooperate. What is more, this plug-in takes the sequences that visualize the transfer of work, and compares them against each other to see if they follow the same pattern. The found patterns are then displayed in a separate diagram: the pattern diagram. Here the patterns are displayed, sorted based on the frequency of sequences that follow the pattern (the pattern with the highest frequency if displayed at the top). Furthermore, information such as the mean throughput time of the patterns is available as well here. Using this information, you can easily determine which patterns appear often, and thus seem to be common behavior, which not (rare, maybe even unwanted, behavior) and which patterns result in a high throughput time, and thus may indicate unwanted behavior. With help of the patterns, you may be able to understand the causes of this unwanted behavior (a lot of time spend at a certain data-element instance perhaps). This can help you with preventing the behavior from appearing in the future.

→ Read more...