Skip to main content

Data Mesh Architecture

Unlock the true potential of enterprise data.

What is Data Mesh Architecture?

While most organisations realise the value of data in informing decision-making and creating more tailored customer experiences, centralised data platform architectures often hinder the ability to provide timely insights at scale.

Data Mesh Architecture is a decentralised approach to data management that aims to unlock the true potential of enterprise data by assigning ownership and governance to the business domains that create the data.

Data Mesh is founded on four principles.

Domain-oriented data ownership

Data Mesh architecture embraces the concept of domain-driven design. It gives responsibility and ownership to the business domains who are most familiar with it, usually this is the domain that produces the data. The domain teams have access to tools to publish data and control the governance to share the data and metadata. Similar to domain driven development, contracts are used to control the interfaces between producers and consumers of data. The focus is on decentralising data hosting, making it easily consumable by different users across the organisation in a standardised manner

Data as a product

Data Mesh treats data as a product and applies product management and domain driven design principles to data. Data products can have SLAs around various aspects, including data quality, refresh rates, and even user satisfaction. By moving accountability from a centralised data engineering to the domain team that is closer to the data, the quality and trust of the data increases.

Self-service data platform

Data Mesh architecture enables domain autonomy by providing self-service capabilities to business teams. Data engineers build the necessary infrastructure to empower domain experts without requiring them to have data engineering skills. This infrastructure supports the self-service provisioning of data products.

Federated data governance

Federated data governance describes how the organisation can define central data governance, but still allow data producers to manage and control their data products locally. This is implemented using a data platform and / or tools that allow data producers to create, share and manage access to their data products. Each domain is responsible for specific decisions, such as data models, quality assurance and data governance (i.e. who can access their data), whilst also complying with organisation wide governance rules. The data platform also allows data consumers to use data from multiple domains, once they have been granted access to them.

Get in touch

Empowering RenaissanceRe’s Data-Driven Future

RenaissanceRe, one of the world’s leading reinsurance companies, specialises in managing risks associated with climate change, natural hazards, cyber threats, and societal upheaval. With 11 global offices across EMEA, APAC and the Americas, in 2022 gross premiums were $9.2B with around $17B of capital under management. The key differentiator compared to the competition is their ability to use leading-edge technology to help them analyse risk and understand their portfolio better than anyone else.

Following the successful migration of the rollup process to AWS, RenaissanceRe chose to modernise its Software Solution estate, which has been built over the past 30 years. This included some legacy monolithic applications with inconsistent integration patterns, implicit dependencies at the data layer, inconsistent use of APIs and Message Queues, and multiple front-end technologies.

RenaissanceRe’s modernisation goals included supporting significant business scale-up and enhanced efficiency, fostering agility through composability, developing new business capabilities and the democratisation of data.

The result? A new data platform service, named Iris, designed to offer a common platform for all domain teams within RenaissanceRe, enabling them to produce and consume data consistently and in a standardised manner. Iris enables RenaissanceRe to remove gatekeepers using an ‘Inner Source’ model for controlled access to cross-domain datasets. It promotes open development of shared platform components and encourages contributions via pull requests. Iris cultivates a self-service engineering culture, prevents obstacles, and supports event-driven structures for adaptable workflows aligned with evolving business needs.

The Benefits

Why use Data Mesh Architecture?

Data Accessibility

Self-service capabilities provided by data mesh platforms make it easier for users to access, discover and combine data products. This means quicker data insights and promotes knowledge sharing across different domains.

Improved Governance

Separating datasets into distinct domains based on their purpose helps ensure all relevant regulations are met at each stage of the process – making it easier to maintain compliance.

Empowered Domain Experts

Decentralised data ownership moves ownership of the data from centralised data teams to the domain teams that created it. This enhances data quality and reliability of the data, making the entire organisation more data-driven.

Increased Scalability

Due to its distributed nature, data mesh architecture can efficiently scale up or down, surpassing the capabilities of centralised data platforms.

AWS Advanced Tier Consulting Partner

We are AWS MAP-compliant.
With Serverless expertise.

Blog & News

The future is always changing. We'll keep you there.

Check out our latest observations and insights into where it’s all going.