DeLorean: A Storage Layer to Analyze Physical Data at Scale
Michael Kußmann, Maximilian Berens, Ulrich Eitschberger, Ayse Kilic, Thomas Lindemann, Frank Meier, Ramon Niet, Margarete Schellenberg, Holger Stevens, Julian Wishahi, Bernhard Spaan and Jens Teubner
Proc. Datenbanksysteme für Business, Technologie und Web (BTW), Stuttgart, Germany, March 2017
Modern research in high energy physics depends on the ability to analyse massive volumes of data in short time. In this article, we report on DeLorean, which is a new system architecture for high-volume data processing in the domain of particle physics. DeLorean combines the simplicity and performance of relational database technology with the massive scalability of modern cloud execution platforms (Apache Drill for that matter). Experiments show a four-fold performance improvement over state-of-the-art solutions.