Main Content

Fine-Grained Energy Profiling of Stream Processing Systems

Author: Ruben Ehritt

Supervisors: Prof. Dr. Boris Koldehofe, M.Sc. Wenfei Huang

Format: Bachelor's thesis, TU Ilmenau

As computing infrastructure expands, the energy costs, both monetary and environmental, continue to rise. To counteract this trend, the energy efficiency of both hardware and software is continuously being optimized. However, effective optimization of a system's energy efficiency requires a comprehensive understanding of its energy footprint. This thesis proposes improved measurement methods specifically for stream processing systems. By focusing on this specific computing paradigm, we hope to better understand its particular characteristics. The primary contribution of this work is the development of a tool that facilitates the measurement and analysis of energy consumption in stream processing systems. In addition to the capabilities of existing tools, the proposed tool enables coordinated measurement across distributed systems. This approach aims to provide a comprehensive understanding of the entire system while maintaining the ability to perform low-level, fine-grained measurements.

References: https://bibliographie.tu-ilmenau.de/servlets/DozBibEntryServlet?mode=show&id=ilm_mods_00081055