Optimizing the Performance of Automated Business Processes Executed on Virtualized Infrastructure

With few exceptions, the opportunities cloud computing offers to business process management (BPM) technologies have been neglected so far. We investigate opportunities and challenges of implementing a BPM-aware cloud architecture for the benefit of process runtime optimization. Processes with predominantly automated tasks such as data transformation processes are key targets for this runtime optimization. In theory, off-the-shelf mechanisms offered by cloud providers, such as horizontal scaling, should already provide as much computational resources as necessary for a process to execute in a timely fashion. However, we show that making process data available to scaling decisions can significantly improve process turnaround time and better cater for the needs of BPM. We present a model and method of cloud-aware business process optimization which provides computational resources based on process knowledge. We describe a performance measurement experiment and evaluate it against the performance of a standard automatic horizontal scaling controller to demonstrate its potential.