Federated Data Integration in Data Spaces

Abstract

Data Spaces form a network for sovereign data sharing. In this chapter, we explore the implications that the IDS reference architecture will have on typical scenarios of federated data integration and question answering processes. After a classification of data integration scenarios and their special requirements, we first present a workflow-based solution for integrated data materialization that has been used in several IDS use cases. We then discuss some limitations of such approaches and propose an additional approach based on logic formalisms and machine learning methods that promise to reduce data traffic, security, and privacy risks while helping users to select more meaningful data sources.

Towards Auditable and Intelligent Privacy-Preserving Record Linkage

“…A field that can leverage the PPRL techniques is the Federated Data Linkage application. Federated Data Linkage is a data integration task that intends to link data of multiple institutions (e.g., a hospital) that integrates a large network/federation (e.g., all federal hospitals of a country) [Jarke and Quix 2022]. In the Federated Data Linkage context, laws (e.g., GDPR and LGPD) and regulations (e.g., HIPAA) require that the privacy of the individuals are preserved during the task execution [Boyd et al 2012].…”