From Server-centric to Client-centric Data Integration over Decentralized Knowledge Graphs with Personal Query Engines

Knowledge Graph technologies provide us with the ability to interlink data across organizational boundaries, which lead to implicit Decentralized Knowledge Graphs (DKGs). This interlinking ability makes DKGs a natural foundation for public and private data, as all data can be stored and managed by the people or authorities owning this data. When one wants to build applications or services on top of DKGs, this decentralized nature leads to major technical challenges regarding data integration. Today, most of these challenges are being tackled in a server-centric manner, where decentralized data is integrated and centralized within servers that run somewhere in the cloud, far from the control of the people that actually own the data. Consequentially, this server-centricity lies at the root of many privacy concerns surrounding personal data. In this article, I argue that the ability to store data in a decentralized manner is insufficient for solving issues surrounding privacy and trustworthiness. Additionally, we need to rethink how and where we integrate this data, and perform computations over it. Concretely, this article focuses on a paradigm shift from server-centric to client-centric data integration, by providing users with reusable and personal query engines. These are client-side engines that are in the driving seat for integrating decentralized data, as a basis for offering users with full algorithmic control over how their data is processed, in a transparent manner, which is private by design. This article discusses three open research challenges that are fundamental to client-centric query processing for driving data integration within decentralized applications. Namely, the challenges of heterogeneity, personalization, and performance. While server-centric data integration maximizes performance by sacrificing user control, transparency, and privacy, a client-centric approach does not require this sacrifice, by shifting focus to the client. This offers a new paradigm for application domains where decentralization is an inherent property, such as healthcare, social media, and personal genomics.

@inproceedings{taelman_semanticsbluesky_clientcentricqueryengines_2026,
  author = {Taelman, Ruben},
  title = {From Server-centric to Client-centric Data Integration over Decentralized Knowledge Graphs with Personal Query Engines},
  month = sep,
  booktitle = {Joint Proceedings of Posters, Demos, Blue Sky, Workshops, and Tutorials of the 22nd International Conference on Semantic Systems},
  year = {2026},
  url = {https://www.rubensworks.net/raw/publications/2026/taelman_semanticsbluesky_clientcentricqueryengines_2026.pdf}
}