Detail

Scalable Business Process Automation

Modern software applications must manage automated business processes with millions of concurrent users. Existing application platforms fail on this scale, as demonstrated by the fact that successful globalized businesses such as Amazon or Google use their own hand-crafted systems to deploy applications.

In this talk, we will discuss strategies of successful systems for highly scalable business process automation, and explain why these strategies are not well supported by today’s application servers. Motivated by our observations, we introduce the Demaq project, which investigates a novel style of realizing highly scalable applications. The premise of Demaq is that distributed applications are just message management problems: Participants in automated business processes communicate using messages, and the execution of processes is the problem of determining which new messages to send, given a certain history of received messages. For this purpose, Demaq provides a powerful language and execution system for message management. Essentially, we turn large-scale process management into a data management problem, and leverage decades of research on database technology to solve this problem elegantly and efficiently. This boosts developer productivity and — as experimental results confirm — application scalability and performance.