Open Data Spaces and Distributed Data Management

Open Data Spaces (ODS) is an open, scalable technology concept for distributed data management that respects the diversity of countries and organizations. It is designed to avoid dependence on a specific cloud, platform, or company and to remain open beyond the institutional systems of any particular region.

ODS provides an open architecture and protocols for managing the business-specific data and context required to deploy Agentic AI reliably across domains and organizations while keeping them distributed. Users can understand the existence, location, and meaning of data, while providers can identify and control the users, target data, and purposes of use.

In ODS, “Open” does not mean unrestricted disclosure; it means opening data in a manageable way that combines transparency and controllability while preserving fairness.

For companies, this opens a path to using real data lying dormant internally as management capital connected to the circumstances and meaning behind its generation. Moving from centralized data management to a model in which operational domains manage and provide data themselves can facilitate data use not only across departments but also across organizations.

Agentic AI could shift from an answer engine that depends only on general information to an action engine that reasons from context specific to the operational site and leads to trustworthy actions. Stakeholders in wholesale, logistics, manufacturing, and other fields can provide the necessary portions to the necessary parties in the necessary manner.

Through a design that is not confined to the institutional systems of individual countries or regions, ODS can serve as a foundation for data management across organizations and national borders. Adoption across multiple fields indicates a movement to expand cross-industry data integration into practical operations.

ODS is designed to commercialize real data by adding domain context, move from centralized to distributed management, and extend the Data Mesh concept from interdepartmental to interorganizational use. Its approach is Serving and Pull rather than Push and Ingest: data providers offer data as a service, and users retrieve it.

The three pillars supporting distributed management are Data Addressability and Discoverability (DAD), which identifies and discovers data locations; Ontology and Semantic Interoperability (OSI), which aligns data meanings; and Identity and Usage Control (IUC), which manages who uses data and how.

The Double Product Quanta Model (DPQM) is a two-layer structure that treats data and domain context equally. It combines the open-world assumption, which accepts unknown information, with the closed-world assumption, which requires rigor at the time of use, achieving both safety and flexibility. Dynamic Ontology manages data structure and meaning separately to address changes in meaning and differences in definitions between organizations.

In terms of adoption, the automotive and batteries field has been operating and expanding on a full-scale basis since 2024; operations in unmanned aircraft traffic management are scheduled to expand from 2025; and implementation in chemicals and resource circulation is scheduled to begin from 2026. At present, however, certification systems and conformity verification mechanisms have not been established for ODS deliverables.