Performance modelling for distributed transaction processing systems

Automated business processes running on distributed transaction processing (DTP) systems are the IT backbone of services industries. New web services standards such as BPEL have increased the importance of DTP systems in business practice. Understanding the performance of such infrastructures is essential for IT departments as they are forced to meet pre-defined quality-of-service metrics. The complexity of multiple interacting services running on multiple hardware resources as well as the volatility in the demand for these services can make performance prediction challenging. While business process automation has been a dominant topic in the recent years, surprisingly little has been published on performance modelling of large-scale DTP systems. Applications of queueing network models are largely focused on rather small systems. In this thesis, we evaluate different types of queueing network models and discrete event simulations for predicting performance metrics of largescale DTP systems, based on log data from three real-world DTP systems. The results show that queueing network models yield accurate and reliable predictions under a variety of demand mix scenarios, but also highlight some of the problems and modelling challenges in this domain.