New Framework Lets Companies Query Scattered Data Without Moving It
Researchers have developed a system that lets organizations search and integrate data across multiple databases in real time without consolidating everything into one place. The approach could cut costs and complexity for healthcare systems and other industries drowning in fragmented data sources.
Originaltitel: Architecting a Federated Semantic Knowledge Lake Framework: Knowledge Graph-as-a-Platform for Distributed Data Ecosystem
Organisationer kan sluta bygga dyra centraldatalager genom att integrera datasilor i realtid med en federerad semantisk kunskapssjö. Ramverket använder ontologibaserad dataåtkomst och FedX-federation för att koppla heterogena informationssystem utan att flytta data fysiskt. Forskarna skapar semantiska datapipelines som omvandlar ostrukturerad, semistrukturerad och strukturerad data till ontologiska kunskapsgrafer. En federerad virtuell kunskapsgraf (FVKG) möjliggör sökning över alla datakällor via SPARQL-endpoints—en plattformsstrategi som avlägsnar semantiska hinder mellan system. Ramverket testas inom sjukvården, där dataintegrering är kritisk för patientjournalsystemens interoperabilitet. Halmstad University och China Academy of Information and Communications Technology leder projektet. För leverantörer av dataplattformar och integrationsverktyg öppnas en marknad för federerade kunskapsgrafer som alternativ till traditionell ETL och datamigration.
With the exponential growth of heterogeneous data across distributed information systems (IS) and the influx of raw data with descriptions, organizations emphasized building virtual systems dealing with data silos across multiple data sources ranging from structured relational databases to unstructured data. Traditional data integration methods face scalability, freshness, and interoperability limitations. We propose a federated semantic knowledge lake (FSKL) framework built upon ontology-based data access (OBDA) and the FedX federation engines, enabling real-time integration and seamless interoperability across heterogeneous healthcare IS without requiring centralized data migration. The framework establishes semantic data pipelines for unstructured, semi-structured, and structured data sources, transforming them into ontological knowledge graphs. As a result, these are integrated into a federated virtual knowledge graph (FVKG) to enable seamless, real-time data access using SPARQL endpoints. The proposed knowledge-graph-as-a-platform (KGaaP) resolves semantic interoperability and supports service-oriented healthcare applications through SPARQL-based federated querying, promoting dynamic, scalable, and interoperable data ecosystems.