The completion of the new HPI Data Lab in 2021 will open up new perspectives, especially for research work in machine learning and data analysis. One focus will be on methods for processing and storing large and dynamic data sets as well as complex data analysis on parallel distributed infrastructures and modern database system architectures.

The Data Lab, an experimental platform for scientists from HPI's departments, is based on three main pillars: Hardware, Software, and Data.

The Data Lab provides the research projects at HPI with a wide range of robust and modern computer architectures. This includes high-performance servers with extensive main memory and powerful processors, computer clusters with many nodes and cores, GPU servers with many computationally optimized graphics cards, modern hardware systems with, for example, non-volatile RAM modules and power architectures, and SAN systems with sufficient storage capacity and high read/write speeds. The lab's hardware landscape is designed for expandability so that the research lab can grow with a growing demand for project capacities at HPI, and new hardware trends can be followed and realized in the lab.

On top of the powerful hardware, the Lab offers various state-of-the-art software systems for processing, transforming, analyzing, storing, and machine learning data. This off-the-shelf software includes runtime environments for major programming languages, frameworks for machine learning and distributed computing, and selected proprietary analysis software and databases.

The Lab will serve as a repository for research and industry project data that will be processed and analyzed in the Lab.

Upon completing the primary data center building at HPI Campus II, the Data Lab and its hardware will consolidate into a "Data Center" with other Labs at the Institute. The entire building will use the HPI Data Lab's waste heat as a supply heat in the future.