zur Startseite


Adaptive Load Shedding in Complex Event Processing
Betreuer M. Sc. Ahmad Slo
Prüfer Prof. Dr. rer. nat. Dr. h. c. Kurt Rothermel

Thesis Description

The tremendous increase in data volume and the need to interpret this data in real- time, to extract useful information have motivated many research communities to develop technologies that process such huge data online. Complex event processing (CEP) is one effective system to process such stream of data. CEP is used in many domains such as IoT, social media, E-commerce, etc.

However in burst situations, the input stream volume may exceed the system ca- pacity. This increases the processing latency of events or it even may break down the whole system. One way to handle burst situations is to drop a part of input data, also known as load shedding.

We have developed a load shedding strategy to drop unimportant events from the input event stream. We built a utility model which probabilistically learns about the importance of events in the input event stream. However, if the input event distribu- tion changes over time and the impact of the utility model on the quality of results is negatively increased, we must re-train the model to adapt for those changes. In this thesis, our goal is to develop a strategy that realizes the changes in the data and the need to re-train the model. Moreover, our load shedding strategy has a parameter (called f ) that controls when to drop events and the period in which the events are dropped. This parameter has a considerable impact on the quality of results and it must be tuned on-line.


  • Understand our available prototype CEP framework.
  • Develop a strategy to trigger the need for re-training the utility model.
  • Investigate and develop a strategy to dynamically tune f on-line.
  • Implement and integrate the proposed algorithm in the framework.
  • Evaluate the developed model and algorithm extensively.
  • Document the developed concepts, algorithms and the evaluations in written form.
  • Present your results in VS colloquium.


  • Very good programming knowledge in Java.
  • Good background in probability theory and/or Machine Learning.
Download the description