hyperlakeDiscuss a deployment ↗
Videos · 2:09

What is a Data Mesh

Data mesh distributes data ownership to business domains while using product thinking, self-service infrastructure, and federated governance at scale.

In the full article

  1. What is data mesh, and what does the architecture change?
  2. Why do centralized data platforms stall as organizations grow?
  3. What are the four data mesh principles?
  4. How do you implement a data mesh operating model?
  5. How Hyperlake helps
  6. Frequently asked questions
Read the transcript

Auto-generated captions; wording may contain recognition errors.

Most large organizations build their data platforms the same way. A central team owned all the data. Business domains sent their data to a central lake or warehouse, and when analysts needed something, they submitted a request and waited. This pattern works until the organization grows past a certain size, at which point it becomes the primary reason data initiatives stall. The central team cannot understand every domain deeply enough to serve it well. Request queues grow longer and longer. Data quality degrades in translation from domain to central platform. The teams closest to the data lose ownership and context. By the time an insight reaches someone who can act on it, the moment has often passed. Data mesh is an architectural and organizational approach that treats this as a distribution problem, not a technology problem. Rather than moving data to a central team, it moves ownership of data to the teams that generate and understand it. The approach rests on four principles. Domain-oriented ownership means business domains own and serve their own data, rather than delegating that responsibility to a central engineering team. Data as a product means domains treat their data outputs with the same care as software products, with clear interfaces, quality guarantees, and accountability to consumers. A self-serve data platform provides shared infrastructure that lets domain teams build and operate data products without needing to be infrastructure specialists. Federated computational governance applies shared standards across all domains without requiring a central authority to enforce them manually. Data mesh does not replace your storage layer. It changes who is responsible for what sits on it, and how that data is offered to the rest of the organization. The organizational change is harder than the technical one, and most implementations that struggle do so for that reason.

Start with a workload. Build the environment around it.

Explore example deployments, or see how the platform assembles, deploys, governs and operates the stack.