Home | Glossary | Data Infrastructure | Understanding Data Mesh

Understanding Data Mesh

A data mesh is a modern approach to data architecture that emphasizes the decentralized management of data within an organization. In a data mesh, data is treated as a product that is owned and managed by individual teams, rather than as a centralized resource managed by a single data team or IT department. This approach is designed to enable greater agility, flexibility, and scalability in managing and analyzing data, while also promoting a culture of data ownership and accountability across the organization.

Group Of People Meeting With Technology. Group Of People Meeting With Technology.

The Concept of Data Mesh

The concept of data mesh is to shift the focus of data architecture from a centralized approach to a decentralized approach. In a traditional centralized approach, data is managed and governed by a centralized data team or IT department, which is responsible for maintaining and managing the organization's data infrastructure. This approach can lead to issues with data quality, ownership, and agility, as it can be slow and inflexible in responding to changing business needs.

Businessman Cooperating With Doctors While Going Through Reports On A Meeting. Mid adult businessman and healthcare workers analyzing documents during a meeting in the office. Businessman Cooperating With Doctors While Going Through Reports On A Meeting. Mid adult businessman and healthcare workers analyzing documents during a meeting in the office.
Hands of an unrecognizable woman sitting at her desk and doing financial analysis at work Hands of an unrecognizable woman sitting at her desk and doing financial analysis at work

What is data mesh architecture?

Data mesh architecture is a relatively new approach to data management that aims to address the challenges of scaling and democratizing data within large, complex organizations. In a data mesh architecture, data is treated as a product that is owned and managed by individual teams, rather than as a centralized resource. Each team is responsible for the data products they create, and they are expected to maintain the quality, reliability, and security of their data. These teams are typically cross-functional and include data engineers, data scientists, and domain experts.

Data Mesh Principles

The data mesh architecture is based on several principles, including:

  • Domain-oriented decentralized data ownership: In a data mesh, data is treated as a product that is owned and managed by individual teams or domains. Each domain is responsible for the management, quality, and governance of its own data products.
  • Self-serve data infrastructure: Each domain is responsible for building and managing its own data infrastructure, including data storage, processing, and analytics tools. This approach enables teams to build and deploy their own data products independently, while also promoting interoperability and data sharing across the organization.
  • Federated governance and mesh API: A federated governance model is used to ensure consistency and compliance across the organization, while a mesh API is used to enable data discovery, sharing, and collaboration across domains.
  • Product-centric thinking: Data is treated as a product, with a focus on delivering value to end users. This approach emphasizes the importance of understanding user needs and designing data products that meet those needs.
  • Platform thinking: A platform mindset is used to enable interoperability and data sharing across domains, with a focus on building modular and scalable data infrastructure that can be easily integrated with other components.
  • Data as a first-class citizen: Data is treated as a first-class citizen, with a focus on managing and governing data as a strategic asset that can drive business value.

These principles are designed to promote a culture of data ownership and accountability, while also enabling greater agility, flexibility, and scalability in managing and analyzing data within an organization.

Domain ownership

Domain ownership in the context of data mesh refers to the idea that each domain or business unit within an organization is responsible for the management, quality, and governance of its own data products. This means that each domain is responsible for building and managing its own data infrastructure, including data storage, processing, and analytics tools, as well as ensuring the quality and accuracy of its data.

Data as a product

Data as a product in the context of data mesh refers to the idea that data is treated as a valuable asset that can drive business value, rather than just a byproduct of software applications or IT systems. In this approach, data is managed and governed like a product, with a focus on delivering value to end-users.

Self-serve data platform

Self-serve data platform in the context of data mesh refers to the idea that each domain or business unit within an organization is responsible for building and managing its own data infrastructure, including data storage, processing, and analytics tools. This approach enables teams to build and deploy their own data products independently, without having to rely on a centralized data team or IT department.

Federated computational governance

Federated computational governance in the context of data mesh refers to the idea that a federated governance model is used to ensure consistency and compliance across the organization, while also enabling data discovery, sharing, and collaboration across domains.

In a federated computational governance model, each domain or business unit within an organization is responsible for managing its own data products, including data quality, access controls, and governance policies. However, there is also a centralized governance body that sets the overall policies, standards, and guidelines for data management and governance across the organization.

Data Mesh vs. Data Lake

Data mesh and data lake are both modern approaches to data architecture, but they differ in several key ways.

A data lake is a centralized repository that is used to store large volumes of structured, semi-structured, and unstructured data. The goal of a data lake is to provide a centralized source of truth for data within an organization, enabling teams to analyze and gain insights from large volumes of data. However, data lakes can be challenging to manage, as they require significant resources to ensure data quality, governance, and security.

In contrast, a data mesh is a decentralized approach to data architecture that emphasizes the ownership and management of data by individual teams or domains within an organization. In a data mesh, data is treated as a product, with each team responsible for building and managing its own data products. This approach promotes greater agility, flexibility, and scalability in managing and analyzing data, while also promoting a culture of data ownership and accountability.

Multi-ethnic serious businessmen looking at computer in office. Male executives are planning at desk. They are wearing smart casuals. Multi-ethnic serious businessmen looking at computer in office. Male executives are planning at desk. They are wearing smart casuals.
African-American woman sitting at the desk with laptop, reading, reviewing few documents. Female working from distance at the home office. Online business career. African-American woman sitting at the desk with laptop, reading, reviewing few documents. Female working from distance at the home office. Online business career.

Data Mesh vs. Data Fabric

A data fabric is a more centralized approach to data architecture that focuses on integrating and connecting disparate data sources and systems within an organization. A data fabric is designed to provide a unified view of data across the organization, enabling teams to access and analyze data in a more efficient and effective manner. This approach is often used in larger organizations with more complex data ecosystems.

The key difference between data mesh and data fabric is the approach to data ownership and management. In a data mesh, data is managed in a decentralized manner, while in a data fabric, data is managed in a more centralized manner. This has important implications for data governance, quality, and security, as well as the ability to scale and innovate with data.

How Data Mesh Works

To implement a data mesh, organizations typically follow a set of best practices and design patterns, including:

  • Identifying domains: The first step in implementing a data mesh is to identify the domains or business units within the organization that will be responsible for managing their own data products.
  • Defining APIs: Once domains have been identified, a set of APIs is defined that will enable data discovery, sharing, and collaboration across domains.
  • Building self-serve data infrastructure: Each domain is responsible for building and managing its own data infrastructure, including data storage, processing, and analytics tools.
  • Establishing federated governance: A federated governance model is established to ensure consistency and compliance across the organization, while also enabling data discovery, sharing, and collaboration across domains.
  • Emphasizing product-centric thinking: Each domain is responsible for delivering high-quality data products that meet the needs of its users, with a focus on delivering value to end users.

Benefits of a Data Mesh

Data mesh is a modern approach to data architecture that offers several benefits, including:

  • Greater agility and flexibility: By decentralizing data ownership and management, data mesh enables teams to build and deploy their own data products independently, without having to rely on a centralized data team or IT department. This approach enables greater agility and flexibility in managing and analyzing data within an organization.
  • Improved data quality and governance: Data mesh promotes a culture of data ownership and accountability, with each domain responsible for the quality and accuracy of its own data products. This approach helps to ensure that data is managed and governed in a consistent and effective manner, while also promoting greater collaboration and interoperability across domains.
  • Increased innovation and experimentation: By promoting a culture of data ownership and accountability, data mesh enables teams to experiment with new data products and analyze data in new ways, without being constrained by a centralized data team or IT department.
  • Improved scalability: By promoting self-serve data infrastructure and modular data architecture, data mesh enables organizations to scale their data infrastructure in a more efficient and cost-effective manner.
  • Enhanced data privacy and security: Data mesh promotes a federated governance model that ensures data privacy and security across the organization, while also enabling data sharing and collaboration across domains.
Personal perspective through reflective glass of smiling young businesswoman working on laptop at workstation near office window. Personal perspective through reflective glass of smiling young businesswoman working on laptop at workstation near office window.
Young businessman enjoying music and working on computer Young businessman enjoying music and working on computer

Data Mesh Use Cases

Data mesh is a modern approach to data architecture that has a wide range of use cases across industries and domains. Some of the common use cases of data mesh include:

  • Customer analytics: In industries such as retail, e-commerce, and financial services, data mesh can be used to analyze customer behavior and preferences, enabling organizations to deliver more personalized and targeted products and services.
  • Supply chain management: In industries such as manufacturing and logistics, data mesh can be used to track and analyze the movement of goods and materials, enabling organizations to optimize their supply chain operations and improve efficiency.
  • Healthcare analytics: In the healthcare industry, data mesh can be used to analyze patient data and clinical outcomes, enabling organizations to identify patterns and trends and deliver more personalized and effective treatments.
  • Fraud detection: In industries such as finance and insurance, data mesh can be used to detect and prevent fraud, by analyzing large volumes of transaction data and identifying suspicious activity.
  • Marketing and advertising: In industries such as advertising and media, data mesh can be used to analyze consumer behavior and preferences, enabling organizations to deliver more targeted and effective marketing campaigns.

By treating data as a product and promoting a culture of data ownership and accountability, organizations can better leverage their data assets to drive business value and competitive advantage.

Learn how Reltio can help.

UPDATED RELTIO FOOTER 2x UPDATED RELTIO FOOTER 2x