Cyber Infrastructure and Beyond
Tuesday, 27 March 2018
11:00am - 12:30pm
Campus Stuttgart-Vaihingen, Pfaffenwaldring 47
Data Provisioning in Simulation Workflows
Peter Reimann, University of Stuttgart
The input data of computer-based simulations often come from diverse data sources that manage data in a multiplicity of proprietary formats. Corresponding simulation workflows thus have to carry out many complex data provisioning tasks. These tasks filter and transform heterogeneous input data in such a way that underlying calculation tools are able to ingest them. So, scientists have to spend much effort to implement many low-level data transformations. This talk covers a novel pattern-based approach that conquers the data complexity associated with simulation workflows and that completely removes the burden from scientists to implement lowlevel data transformations. Instead of designing many workflow tasks, scientists only need to select a small number of abstract patterns to describe a high-level simulation process. Furthermore, scientists are familiar with the parameters to be specified for the patterns, because these parameters are related to their domain-specific methodology.
Transparent Data Exchange in Simulation Choreographies
Michael Hahn, University of Stuttgart
Data provisioning, management and transformation are crucial tasks in scientific simulations, especially if multiple simulations are coupled as so-called multi-* simulations. The notion of choreographies will enable scientists to model such coupled simulations from a global perspective through the interconnection of different scientific workflows and simulation software without the need to directly provide complex technical details. To reflect data-related aspects on the level of such simulation choreographies, we introduced the notion of data-aware choreographies through our concepts for transparent data exchange (TraDE). Our main goals are to ease the modeling of coupled simulations and their data; provide continuous, uniform access to simulation data; automate and hide data management, exchange and transformation tasks from scientists wherever possible; and provide an integrated end-to-end tool support for the modeling and execution of data-aware choreographies.
Utilizing Networked Mobile Devices for Scientific Simulations
Christoph Dibak, University of Stuttgart
Emerging augmented reality devices enable novel, interactive simulations. Simulation results no longer need to be viewed on a computer screen but can be overlayed over real-world objects like complex machinery. This supports engineers in making better decisions in the field. However, having interactive simulation results available on mobile devices like augmented reality headsets is challenging since such battery-powered mobile devices are restricted in computational power, have limited energy resources, and are subjected to intermittent connectivity. Additionally, the execution of simulations is hard to distribute between heterogeneous devices. In our research, we developed different methods that can be used for the distribution of the computation between mobile device and remote server. Our approaches significantly improve latency, dynamically adapt quality, and provide simulation results even in harsh environments.