Designing a Reproducible Measurement Framework for MQTT-to-Kafka Performance Evaluation
| Typ | Bachelorarbeit | |
|---|---|---|
| Aushang | ||
| Betreuer | Wenden Sie sich bei Interesse oder Fragen bitte an: Maximilian Hummel (E-Mail: maximilian.hummel@kit.edu) |
Motivation
IoT and microservice systems depend on scalable messaging pipelines to process massive data streams. MQTT brokers such as EMQX, integrated with Kafka, are widely used, but reproducible benchmarking setups are scarce. Establishing such an environment and producing a structured performance dataset would enable systematic research in microservice performance modeling.
Tasks
The thesis focuses on designing and implementing a reproducible measurement setup for MQTT–Kafka integration. Your tasks include: Reviewing existing benchmarking approaches.
Designing a parameterizable testbed for throughput and latency experiments.
Automating deployment and execution of benchmark runs.
Conducting measurements under different workloads and configurations.
Compiling results into a reusable performance dataset for research.
You will gain experience in distributed systems benchmarking, reproducibility in experimental research, and performance data collection. Knowledge of container-based deployments, scripting (Python, Bash), and distributed systems is recommended.
Additional Info
Optional collaboration with an industry research partner (ABB) is possible, offering access to further resources & expertise during your thesis.
Supervision is available in German or English.
Email: maximilian.hummel@kit.edu