Socio-Physical Advanced Research Cloud Infrastructure (SPARCI)

• Subject Area: Computer Science
• Term: Funded in 2019
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 432399058

The research project entitled “Socio-Physical Advanced Research” is a joint project initiated by multiple research groups in the Faculty of Computer Science at the University of Koblenz-Landau. The research infrastructure and equipment described in the project proposal provides a “Socio-Physical Advanced Research Cloud Infrastructure” (SPARCI) and consists of a high-performance computing cluster, which is complemented with a long-range wide area network infrastructure (LoRaWAN infrastructure). The research project involves the collection, storage, analysis and use of large volumes of data (Big Data) generated in, and from, socio-technical (human-machine) and cyber-physical (mechanical and electronic things) systems. Data providers (sources) and data recipients (consumers) are the parties conducting the research, which is focused on data-intensive applications in three empirical domains: business, public administration and the public world wide web. The project addresses the intelligent connection of real, physical objects with their virtual environment in these three domains.The proposed research applies artificial intelligence and machine learning methods which generate extremely high demand for computing power and storage capacity. In particular, the application of deep learning algorithms and neural networks generates a need for GPU-based computing power, which reduces the processing times of general-purpose computation on existing large data sets to a practical level by using graphical processing units. To collect sensor data, the major research equipment includes a LoRaWAN infrastructure that will be installed both permanently on the campuses of the university and temporarily in off-campus research projects. This low power wide area network infrastructure enables wireless, energy-efficient, bidirectional transmission of small data packets, such as sensor data, over long distances (up to 15 km) and allows the geolocation of transmitters by applying multilateration and time difference of arrival methods on the arrival times of the data packets at multiple receivers.The research groups involved have significant expertise and have conducted extensive preliminary work in the areas of machine learning, innovative approaches to Web Science, collaboration systems, the Internet of Things and IT concepts in public administration. The combination of the research groups’ strengths and specializations in the different, but complementary research areas brings the necessary interdisciplinary and holistic approach.To address compliance issues, the project participants will develop a concept for storing large amounts of heterogeneous sensor data, considering data privacy, data security and personal privacy issues. One challenge of collecting and storing sensor data will be the technical and conceptual prevention of possible assertions about the identity or the movement profile of individuals.

• DFG Programme: Major Research Instrumentation
• Major Instrumentation: Socio-Physical Advanced Research Cloud Infrastructure (SPARCI)
• Instrumentation Group: 7000 Datenverarbeitungsanlagen, zentrale Rechenanlagen
• Applicant Institution: Universität Koblenz
• Leader: Professorin Dr. Petra Schubert