The analysis of complex event processing frameworks

OData support
Dr. Vörös András
Department of Measurement and Information Systems

Nowadays the IoT (Internet of Things) has become more and more popular, be-cause the costs of sensors and other tools have rapidly declined . In many aspects of life, we connect with smart devices, which are continuously collecting data from their envi-ronment.

Unfortunately, systems lack the ability to process and analyse the data effectively that are collected from these IoT devices to detect valuable and interesting events. Be-cause of this problem, a demand appeared for processing systems based on event-driven architecture, so-called complex-event processing systems (CEP). Many concrete tools have been developed to solve these problems, but if we wanted to buy a CEP system, we couldn’t compare between these tools, because of the lack of a formal and widely accepted definition about the concepts of CEP systems.

In my thesis, I examine some tools, and try to compare them with each other. I want to know, what are the main differences in their languages and their expressiveness.

I used a test framework which consists of a LeapMotion sensor, a ZeroMQ and a Google Protobuf implementation. I had to integrate this benchmark system to each CEP systems. They provide events for all my selected CEP systems.

I tried out the Apache Flink: CEP library, the Drools framework, with its Fusion extension, and finally, the WSO2 Complex Event Processor, more precisely its event processing language, called Siddhi Query Language.

After this comparison I close my thesis with my gathered experience about the topic and explain the possibilities and ways on how to continue this research.


Please sign in to download the files of this thesis.