1st Conference on Research Data Infrastructure (CoRDI)

The lecture will introduce participants to NFDI4Biodiversity, a consortium within the German National Research Data Infrastructure (NFDI) dedicated to data and services for biodiversity research and ecology. In the domain, well-developed international networks exist, with quite mature tools and data standards. These are used and disseminated in the NFDI4Biodiversity project in order to mobilise and publish data collected by national stakeholders according to the FAIR Guiding Principles for scientific data management and stewardship.The consortium partners provide methods and tools for archiving, publishing, searching and analysing data that are suitable for everyday use and have been tried and tested in practice. The consortium also functions as a forum for technical and legal matters of data and data-related workflows. In this way, the consortium is providing added value to the community regarding access to modern technologies and a comprehensive stock of biodiversity and environmental data.

Vortragende
Barbara Ebert, Judith Engel, Ivaylo Kostadinov, Anton Güntsch, Frank Oliver Glöckner

Zeit und Ort

12 September 2023, 14:00-15:30h, Audimax (Building 30.95)

Optimizing User Support through Synergistic Inte-gration of Helpdesk Systems: A Case Study of NFDI4Biodiversity and DataPLANT (Judith Sophie Engel, Ivaylo Kostadinov, Jimena Linares, Kevin Frey, Dirk von Suchodoletz, Cristina Martins Rodrigues)

Ort
Audimax Foyer (Building 30.95)

When Data Crosses Borders – Join Forces! Multidisciplinary Use Cases Within NFDI (Barbara Ebert, Sami Domisch, Christin Henzen, Jimena Linares, Kati Mozygemba, Bernhard Miller, Bernhard Seeger, Jörg Seegert)

Ort
Audimax Foyer (Building 30.95)

The talk will introduce the Aruna Object Storage (AOS), a fast, secure, and geo-redundant data storage that is part of the cloud layer of the Research Data Commons (RDC), the digital data infrastructure for biodiversity data developed in NFDI4Biodiversity.

AOS was developed because the exponential growth of scientific data has led to an increasing demand for effective data management and storage solutions. Academic computing infrastructures are often fragmented, which can make it challenging for researchers to leverage cloud-native principles and modern data analysis tools. To address this challenge, AOS was created: a cloud-native, scalable, and domain-agnostic object storage system that provides an S3-compatible interface for a variety of data analysis tools like Apache Spark, TensorFlow, and Pandas. The system uses an underlying distributed NewSQL database to manage detailed information about its resources and can be deployed across multiple data centers for geo-redundancy. AOS is designed to support modern DataOps practices, including the adoption of FAIR principles. Resources in AOS are organized into Objects, Datasets, Collections and Projects, which represent relations of data objects. Additionally, these can be further annotated with key-value pairs called Labels and Hooks to provide additional information about the data. The system's event-driven architecture makes it easy to automate actions and enforce data validation checks, significantly improving accessibility and reproducibility of scientific results. AOS is open source and freely available via aruna-storage.org.

Autor:innen
Sebastian Beyvers
Marius Alfred Dieckmann
Frank Förster
Alexander Goesmann
Jannis Hochmuth
Anna Rehm

Vortragender
Sebastian Beyvers

Zeit und Ort
13 September 2023, 14:00-15:30h, Audimax (Building 30.95)

Across many scientific domains, the ability to process large amounts of heterogeneous spatiotemporal data from various sources is crucial for solving challenging research questions. For example, researchers in NFDI4Biodiversity must combine observational data with satellite images to correlate biodiversity loss with climate change variables. In general, large datasets are not available on the system (called consumer) where the processing is performed, but first have to be retrieved from one or multiple external systems (called providers) that offer a corresponding service. Moreover, a consumer is often unaware of the datasets the providers offer. Ideally, a provider follows FAIR principles and thus supports mechanisms to simplify data exchange. However, in practice, multiple providers with valuable datasets are not as FAIR as desired or lack spatiotemporal-specific support for data exchange. Instead of improving each potential provider at the source, we propose an intermediary spatiotemporal data exchange layer (SDExL) that helps simplify data exchange so that domain experts can easily access valuable data with little technical know-how. Based on practical experience and guided by the FAIR principles, in our talk we postulate four requirements for building an SDExL. Then, we discuss two reference implementations within Geo Engine, a flexible analytical processing platform for spatiotemporal data used in projects like NFDI4Biodiversity and FAIR Data Spaces.

Autor:innen
C. Beilschmidt
D. Brandenstein
J. Drönner
N. Glombiewski
M. Mattig
B. Seeger

Zeit und Ort
13 September 2023, 14:00-15:30h, Audimax (Building 30.95)

Top-level ontologies (TLOs) and mid-level ontologies (MLOs) play a very important role in enabling semantic interoperability between domain-specific ontologies by providing a general structure and common high level entities and relationships for classifying and interlinking domain-specific concepts. A number of such ontologies have been proposed for different purposes. Unfortunately, due to different ontology design patterns, some of these ontologies are not interoperable out of the box. In order to increase the cross-domain interoperability of research data within NFDI and the EOSC, we need to harmonise the used TLO and MLO concepts to a common ground. The Section-Metadata working group Ontology Harmonisation & Mapping was formed to coordinate and guide such an alignment work, by recommending, providing and/or developing mappings, frameworks and tools [1]. We started to analyse which ontologies are used among many NFDI consortia and found that using only one specific TLO & MLO framework will not meet the different ontological requirements. Consequently, we need to provide formal mappings between many commonly used concepts, such as process, information, characteristic or method, defined in a variety of common TLOs and MLOs as well as SKOS vocabularies and other reference terminologies. Since this will be a complex and labour intensive process, we must also look at less complex solutions for interdisciplinary data integration. One approach could be to focus first on the observational part of research data. In our talk, we suggest I-ADOPT as a possible NFDI Wide interoperable variables description framework.

Autor:innen
R. Huber
N. Karam
O. Koepler
P. Strömert

Zeit und Ort
14 September 2023, 10:00-12:00h, Audimax (Building 30.95)

Auf der Marktplatz der NFDI-Konsortien kann man NFDI4Biodiversity während des Market of the Consortia an einem Posterstand im Audimax-Foyer (Gebäude 30.95) besuchen und kennenlernen.

Zeiten
Mittwoch, 13.09., 12:30-14:00h und 15:30-16:00h
Donnerstag, 14.09., 12:00-13:30h und 15:00-15:30h