Performance of Industrial Sensor Data Persistence in Data Vault

Publication, 2018


F. Bachinger, J. Zenisek, L. Kammerer, M. Stimpfl, G. K. Kronberger - Performance of Industrial Sensor Data Persistence in Data Vault - Proceedings of the 30th European Modeling and Simulation Symposium EMSS2018, Budapest, Hungary, Hungary, 2018, pp. 226-233


Today manufacturing companies are facing important challenges from the market in terms of flexibility, ever growing product mixes, small lot sizes, high competition, etc. To meet these market conditions, digitalization and the use of data are offering a viable toolset considering the advances in the field throughout the last couple of years. The increasing use of sensor technology and the need for interconnecting data from different departments in smart production leads to a surge of recorded data. Persistence and integration of heterogeneous data, generated in a variety of software systems, is a key factor to gain value from data and its analysis. High flexibility in regards to the model is required to accommodate the data. Hence, application of the data vault modelling approach is a fitting candidate to design a data warehouse model. In this paper we present a data vault model for factory sensor data. We analyze the performance of the data warehouse in regards to bulk load of data and common analytic queries.