Chapter 3: What is Open Data Spaces (ODS)?
3.1 Concepts of Open Dataspaces
Open Dataspaces is neither an extension of an enterprise's internal data infrastructure nor a shared platform controlled by a single company. What matters is not "collecting all data in one place (Push and Ingest)," but rather "handling data in a distributed state — where it is, what it means, and who can use it and how (Serving and Pull)."
As an approach to distributed data management, "Data Mesh" has already attracted significant attention in the United States and elsewhere, with a growing number of companies adopting it. However, data mesh focuses on how to manage data scattered within an organization while keeping it distributed, and does not address data management that crosses organizations and national boundaries. Open Dataspaces is an approach designed to fill this gap.
There are various challenges in cross-organizational data management, and Open Dataspaces provides solutions addressing primarily the following three:
"Where to Get" Problem: Addresses the existence, identity, and discoverability of data. (e.g., Where does that data exist in the first place? Does that data refer to the same thing as this data?)
"What to Mean" Problem: Addresses the meaning, vocabulary, and semantic consistency of data. (e.g., What does that data mean? Is the meaning of that data consistent with the meaning of this data?)
"Who and How to Use" Problem: Addresses subject trust, access, and terms of use. (e.g., Who is attempting to access the data? Who is allowed to access that data? How must that data be used?)
These are organized as three pillars accordingly:
（1）DAD（Data Addressability and Discoverability）
（2）OSI（Ontology and Semantic Interoperability）
（3）IUC（Identity and Usage Control）
The design philosophy, principles, and architectural implications of these three pillars are documented in Design Philosophy and ODS-RAM. This document limits its coverage to mapping the role of each pillar to the entry opportunities available to users.
3.2 Differences from Conventional Data Management
Conventional data management has relied on two approaches:
consolidating data internally for management and analysis
establishing individual point-to-point connections with each required counterpart.
Open Dataspaces takes a different approach — neither consolidation nor individual negotiation — by handling existence, meaning, and terms of use while keeping data distributed. Table 2 presents a comparison with conventional data management approaches. For a more detailed comparison, please refer to "Design Philosophy".
Table 2 Comparison with Conventional Data Management Approaches
Data placement
Handled individually per connection
Consolidated internally
Handled in a distributed state
Discovery
Assumes prior knowledge of connections and specifications
Assumes an internal catalog
Assumes addressability and discoverability (DAD)
Handling of meaning
Dependent on individual implementation
Managed under internal schema; meaning and structure tend to be tightly coupled
Meaning (semantics/ontology) is separated from structure (OSI)
Identity and usage control
Dependent on individual contracts and implementation
Centered on internal access management
Identity, access, and terms of use are handled explicitly (IUC)
3.3 Value Provided by Open Dataspaces
3.3.1 Value for Data Providers
The reasons data providers hesitate to share data externally go beyond the risk of information leakage alone. Compounding problems include: "We cannot accurately communicate what our data means to external parties," "Our data may be conflated with other companies' data and misused," and "It is difficult to operationalize terms of use." Open Dataspaces enables the separate handling of issues related to data existence and identification, data meaning and vocabulary, and access and terms of use. This separation makes it easier for data providing enterprises to design external data sharing within the necessary scope and conditions, without needing to unify their internal systems or internal identifiers.
3.3.2 Value for Data Users
By separately handling discoverability, semantic organization, and the management of entities and terms of use, Open Dataspaces reduces the operational burden of "interpreting meaning, ensuring consistency, and confirming terms of use" after acquiring external data. This makes it easier for utilizing enterprises to incorporate external data into BI/DI analysis, business applications, and AI initiatives.
3.4 Examples of Business Models Achievable with ODS
Entry opportunities for users in Open Dataspaces can be organized as shown in Table 4. These represent general business model perspectives applicable across industries and use cases.
Table 4 Entry Pattern to Open Dataspaces
Collection and utilization of external data
Acquire data from multiple providers and apply it to BI/DI/AI analysis and operational improvement
Reduction of individual API coordination costs, faster decision-making (Save) / Improved AI accuracy, new service creation (Make)
External provision and monetization of internal data
Provide internally held data to external parties under defined conditions, generating revenue
Acquisition of new revenue streams, expansion of partnerships (Make)
Ecosystem participation through mutual data exchange
Exchange data mutually with multiple parties to enable supply chain management, quality control, and traceability
Operational efficiency and risk reduction (Save) / Compliance with industry standards, enhanced brand value (Make)
Adoption of ODP-implemented services as a user
Use ODP-implemented managed services and applications as a user
Reduced initial investment, faster time to value (Save)
It should be noted that in Open Dataspaces, N:N connectivity becomes a realistic option through a degree of discipline and interoperability. The scale of business value that cannot be achieved through one-to-one individual integrations represents the fundamental motivation for participation in Open Dataspaces.
3.5 The Relationship Between AI and Open Dataspaces
Finally, the relationship between the AI industry and the Open Dataspaces market is examined. In AI applications — particularly in analytics based on operational data and the use of Agentic AI — the ability to continuously handle "data context, meaning, terms of use, and currency" is more important than simply having a large volume of data. The Design Philosophy also highlights that in the Agrntic AI era, unique context is the essence of data, and emphasizes the importance of domain owners explicitly providing that context as meaning (ontology). Open Dataspaces technology provides the technical foundation for addressing these challenges through endpoint discoverability, semantics and ontology, and identity and usage control.
As a result, the following kinds of benefits can be expected:
It becomes easier to perform reasoning and interpretation that presupposes relationships and context between data points
It becomes easier to construct AI pipelines that account for data provenance and currency
It becomes easier to incorporate external data into AI applications without relying solely on individual integration efforts
The next chapter builds on this foundation to examine, from the perspective of participation structure, the positions that users can occupy.
Last updated