Browsing by Author "Wetzstein, Branimir"

Now showing 1 - 2 of 2

Open Access
An event model for WS-BPEL 2.0
(2011) Kopp, Oliver; Henke, Sebastian; Karastoyanova, Dimka; Khalaf, Rania; Leymann, Frank; Sonntag, Mirko; Steinmetz, Thomas; Unger, Tobias; Wetzstein, Branimir
This report presents an engine-independent WS-BPEL 2.0 event model. It supports both passive monitoring and active control of process execution by external applications. Some of the assumptions in the presented event model are inspired by a particular implementation, e.g. fault handling and compensation; however they are kept as general as possible, so that they can be mapped on other engine-specific approaches to tackle faults and support compensation. In addition, the report draws on the experience of some of the authors in business process management and software development. The overall BPEL event model consists of a set of event models for the different types of BPEL entities that change their states: processes, process instances, general activities, scope activities, invoke activities, loops, links, variables, partner links, and correlation sets. The event model is used by the authors of the report in several projects, all utilizing process life cycle events in different scenarios.
Open Access
KPI-related monitoring, analysis, and adaptation of business processes
(2016) Wetzstein, Branimir; Leymann, Frank (Prof. Dr. Dr. h. c.)
In today's companies, business processes are increasingly supported by IT systems. They can be implemented as service orchestrations, for example in WS-BPEL, running on Business Process Management (BPM) systems. A service orchestration implements a business process by orchestrating a set of services. These services can be arbitrary IT functionality, human tasks, or again service orchestrations. Often, these business processes are implemented as part of business-to-business collaborations spanning several participating organizations. Service choreographies focus on modeling how processes of different participants interact in such collaborations. An important aspect in BPM is performance management. Performance is measured in terms of Key Performance Indicators (KPIs), which reflect the achievement towards business goals. KPIs are based on domain-specific metrics typically reflecting the time, cost, and quality dimensions. Dealing with KPIs involves several phases, namely monitoring, analysis, and adaptation. In a first step, KPIs have to be monitored in order to evaluate the current process performance. In case monitoring shows negative results, there is a need for analyzing and understanding the reasons why KPI targets are not reached. Finally, after identifying the influential factors of KPIs, the processes have to be adapted in order to improve the performance. %The goal thereby is to enable these phases in an automated manner. This thesis presents an approach how KPIs can be monitored, analyzed, and used for adaptation of processes. The concrete contributions of this thesis are: (i) an approach for monitoring of processes and their KPIs in service choreographies; (ii) a KPI dependency analysis approach based on classification learning which enables explaining how KPIs depend on a set of influential factors; (iii) a runtime adaptation approach which combines monitoring and KPI analysis in order to enable proactive adaptation of processes for improving the KPI performance; (iv) a prototypical implementation and experiment-based evaluation.