Energy Measurement System for Data Lake: An Initial Approach

Data Lakes are increasingly deployed as a solution for Big Data analytics. Recent improvements in Data Lake technology have focused on improving data access, governance, and discoverability. However, the energy consumption of data operations, a non-trivial issue for eco-conscious organizations, is currently overlooked. Furthermore, existing monitoring tools do not adequately address the complexities of Data Lake architectures. This paper presents the initial phase of developing a system for measuring energy in Data Lake pipeline operations. The novelty of our solution lies in the fact that we define four measures to assess the power usage of crucial hardware components in a Data Lake context: CPU, RAM, NIC, and storage devices. To validate our approach, we developed a monitoring tool grounded in real-world datasets from a Data Lake benchmark.