🕸️ Data Mesh: Decentralising Data Ownership

About this simulation

A data mesh treats data as a product owned by the business domain that understands it best — marketing owns its customer data product, logistics owns its shipment data product — instead of funnelling every dataset through one central warehouse team. This scene shows domain nodes arranged as a peer-to-peer network, each serving requests directly, wrapped in a translucent federated governance layer that enforces shared standards without centralising the data itself.

How it works

• Decentralized mesh — requests travel directly between domains. Latency stays low and roughly flat as you add more domains, because there is no single chokepoint.
• Centralized warehouse — every request is rerouted through one central hub. Latency climbs sharply as domain count and request rate increase, visualised by the longer, hub-routed flight path of each particle.
• Green particles passed their governance check; red particles were flagged as non-compliant. Governance strictness raises the compliance bar (and adds a small latency cost) in both architectures, but federated per-domain checks in the mesh catch violations more effectively than a single central gate.
• The glowing ring around each domain node pulses when a request lands on it.

Controls

• Architecture — switch between a decentralized mesh and a centralized warehouse.
• Domain count — number of independent data-owning domains (3–8).
• Request rate — how many data requests are issued per second.
• Governance strictness — how rigorously every request is checked against shared data-quality and access policies.
• ↺ Reset counters — clears the running stats. Drag to rotate, scroll to zoom.

Did you know?

The term "data mesh" was coined by Zhamak Dehghani in 2019. Its four core principles — domain ownership, data-as-a-product, self-serve infrastructure, and federated computational governance — are designed specifically so that decentralising ownership doesn't mean abandoning organisation-wide standards.