Preventing Key Performance Indicators Violations Based on Proactive Runtime Adaptation in Service Oriented Environment

Key Performance Indicator (KPI) is a type of performance measurement that evaluates the success of an organization or a partial activity in which it engages. If during the running process instance the monitoring results show that the KPIs do not reach their target values, then the influential factors should be identified, and the appropriate adaptation strategies should be performed to prevent KPIs violations. In this paper, we propose an integrated monitoring, analysis, prediction and adaptation approach to prevent KPIs violations. We have considered more than one KPI for a specific process and have tried to reach their target values simultaneously by proactive runtime adaptation before the end of the running process. In order to identify the dependency between KPIs and lower-level influential factors, an analysis is done on the data collected from historical process executions. For this purpose, Data Mining techniques have been used. The result is used to predict the KPIs values of the running instance. If KPIs violations are detected, adaptation requirements and adaptation strategies are identified. Since it is possible to define several KPIs for one process, and each has its own importance, so in this paper we tried to satisfy several KPIs target values

Keywords:

Adaptation , Key Performance Indicator , Data Mining , Dependency Analysis , Decision Tree , Service-Oriented Environment , Supply Chain