Scalability Benchmarking of Apache Flink

Biernat, Nico Alexander (2020) Scalability Benchmarking of Apache Flink (Bachelor thesis), Kiel University, Kiel, 53 pp


Download (663Kb) | Preview


In a time where more and more data is generated, collected and processed, the ability to process Big Data in real-time has become more relevant for faster reaction to important events in several application domains. This led to the development of numerous distributed stream processing engines that support the development of large streaming applications. An important characteristic of a stream processing engine is its ability to scale in order to handle an increasing workload.
In this thesis, we extend the first work on benchmarking the scalability of distributed stream processing engines by providing the necessary components to execute scalability benchmarks for the stream processing engine Apache Flink. For this, we migrate common use cases for stream processing, deriving from an Industrial Internet of Things application,
to Apache Flink. Furthermore, we integrate those implementations into an existing framework for scalability benchmarking. The scalability benchmarks are executed in a public cloud environment and the effect of different configurations on the scalability of Apache Flink are evaluated. Our results show that while some configurations positively impact the scalability of Apache Flink for one use case, they might be less beneficial for or even negatively impact the scalability for another use case.

Document Type: Thesis (Bachelor thesis)
Thesis Advisors: Hasselbring, Wilhelm and Henning, Sören
Keywords: Stream Processing, Scalability, Benchmarking, Flink,
Research affiliation: Kiel University > Software Engineering
Kiel University
Related URLs:
Projects: TITAN
Date Deposited: 19 Oct 2020 15:21
Last Modified: 17 Nov 2020 17:31

Actions (login required)

View Item View Item

Document Downloads

More statistics for this item...