(Excerpted from my research paper “Data Mesh and Data Space: A Comparative Analysis with a Focus on Governance.” – Attila Papp, et al.)
As organizations move towards data-driven business models, one key question emerges: how should data be shared — within the enterprise, or across an entire ecosystem?
Data architectures continue evolving, and the latest trend seems to be to enable data sharing intra- and inter-company. It is the oxygen of modern companies: it unlocks valuable insights, faster innovation and more resilient, data-driven decision making – while also supporting transparency mandates such and ESG and upcoming data-sovereignty regulations. In short, shared data turns isolated facts into collective intelligence.
While data sharing within a company is mainly fueled by analytics, efficiency, and data-driven decision making, it is primarily regulatory motivated between companies. Nevertheless, the two emergent approaches – data mesh and data space – seem quite similar, and it is worth taking a closer look. This post provides an excerpt of our forthcoming publication, “Data Mesh and Data Space: A Comparative Analysis with a Focus on Governance”.
Data Mesh (as per Dehghani 2020) is a domain-oriented, self-serve architecture that treats data as a product and governs it through federated, automated policies.
Data Space (as per Data Spaces Support Centre) is a cross-organizational framework that enables sovereign, trusted data exchange via a shared rulebook and open interoperability standards.
In both cases, their goal is to enable decentralized data sharing. However, they reach it differently. While data meshes emerged in large U.S. enterprises aiming to fix bottlenecks introduced by data lakes; data spaces emerged in the European Union to enable cross-organizational data sharing. A data mesh therefore is driven by efficiency and increased autonomy for domains and an overall goal of providing analytics in time to support data-driven decision making. A data space on the other hand is fueled by sovereignty, transparency and fairness and its main value proposition is to enable trusted data exchange to foster joint innovation and compliance.
Their commonalities mainly lie in interoperability and a data-product mindset, as they both treat data as a product. A data mesh mainly sees it as governed analytics asset, and monetization is not important – in sharp contrast to data spaces. Policies, as units of describing access – or in the case of data spaces: data sharing agreements – can be a point of convergence amongst many.
The following is a comparison table with more details:
| Dimension | Data Mesh | Data Space |
| Origin | Born inside large U.S. enterprises to fix bottlenecks in centralized data lakes. | Emerged in the European Union to enable cross-organizational sharing with data sovereignty. |
| Typical Scope | Intra-company—multiple business domains in one organization. | Inter-company—ecosystem or industry-wide. |
| Primary Driver | Efficiency & domain autonomy. | Sovereignty, transparency & fairness. |
| Key Problem | Centralized silos slow analytics: teams need ownership of their data products. | Centralized platforms can’t satisfy trust, legal, and competitive constraints across firms. |
| Value Proposition | Faster delivery, agile analytics, and data democratization. | Trusted data exchange, joint innovation, compliance with regulations (e.g., EU Data Act, CSRD). |
| Must-Haves to Start | Domain-driven culture, platform team, self-serve tooling. | Common rulebook, mature internal governance at each participant, and interoperability standards. |
| Decentralization Focus | Empowers teams inside one company. | Empowers organizations (and even individuals) across an ecosystem. |
| Interoperability Lens | Schemas, contracts, and metadata are agreed at an enterprise level. | Legal, organizational, semantic & technical layers are agreed upon at a sector level. |
| Data-Product Mindset | Data product = governed analytic dataset; business value comes second. | Data product = economic unit for monetization; technical/metadata specs follow. |
| Tech Stack Trends | Often cloud-native “boxed” services (e.g., Databricks, AWS DataZone); OSS on the rise (Iceberg, Polaris). | Strong OSS culture (Gaia-X, Eclipse Dataspace Components) plus vendor-neutral connectors. |
(Adapted from A. Papp et al., forthcoming)
3 Key Takeaways
- Same destination, different starting point. Both paradigms aim at repeatable, reliable, secure data-sharing. Data Mesh starts inside the enterprise; Data space starts between enterprises.
- Governance is key. Data Mesh relies on federated computational governance: local teams own policies but comply with enterprise-wide guardrails baked into the platform. Data spaces formalize trust in a rulebook enforced by an ecosystem-level Governance Authority.
- Convergence is coming. Open table formats, shared vocabularies, and policy engines (e.g., IDS Contracts, OPA) are shrinking the gap. Expect hybrid architectures where an internal mesh plugs straight into an external data space.
What’s next?
In the second part of this series, we compare their distinct governance approaches. A complete scientific publication with more details is forthcoming. Follow us to be in the loop.
Which paradigm are you exploring—and why?
