Towards flexible event processing in distributed data streams

The FERARI project aims to develop a highly scalable distributed streaming architecture supporting complex event processing in a communication-efficient manner. Two key requirements for our system are that its architecture is not tied to the underlying streaming platform used in its implementation and that it allows the easy definition of communication-efficient methods for monitoring a global condition over a distributed set of states. In this paper we present the architecture of our system and explain how these key requirements are met. Concerning the actual implementation of our system in a scalable distributed streaming platform, it is reasonable not to re-invent the wheel but to use one of the actual Big Data Streaming platforms as a starting point. For this reason, we evaluate some popular platforms and discuss whether they meet our requirements.