Data mesh is a concept that has gained significant attention in the realm of data architecture and management. Many people associate data mesh with being exclusively for analytical data, but this is a misconception. In reality, the principles of data mesh can be effectively applied to operational data as well. In this article, we will delve into the world of data mesh and explore how it can be utilized for both analytical and operational use cases.

Operational data is the lifeblood of a business, as it includes the real-time data that is used to run the day-to-day operations. This can include transactional data, customer data, and other data sets that are crucial for making real-time decisions within an organization. By applying data mesh principles to operational data, organizations can ensure that this data is of high-quality, easily accessible, and can be harnessed to drive business decisions on the fly.

Analytical data architecture is focused on processing large datasets and performing complex data analysis. It requires predictability and real-time processing of small datasets. On the other hand, operational data architecture focuses on real-time processing of small datasets and requires the ability to quickly retrieve and act upon the data. While there are differences in the processing requirements, there is also considerable overlap in the domain model, data treatment, and boundary setting between the two architectures.

The capabilities of an organization, whether viewed through an operational or analytical lens, remain the same. The applications that provide these capabilities have teams behind them that manage and maintain them. The language used for development is also consistent. Therefore, it makes sense to apply similar best practices for data management, event management, and API management, regardless of whether the data is used for analytical or operational purposes.

Building a data mesh architecture requires a solid foundation in the form of landing zones or infrastructure blueprints. These landing zones provide the necessary building blocks for a modern data platform. They ensure standardization across different domains and facilitate the deployment of both analytical and operational workloads. Landing zones help organizations achieve standardization and consistency in data distribution patterns and application integration patterns.

In an analytical data (mesh) pattern, each team owns and manages their own analytical data products. The data is made available to other teams through a data catalog, and each team is responsible for ensuring the quality and reliability of their data product. Different consumption patterns, such as event-carried state transfer pattern, API pattern, or lightweight virtualization query pattern, can be used to consume the data. In an operational pattern, each team owns and manages their own operational event and API products. They are responsible for ensuring the quality and reliability of their endpoints. APIs can be used for commands and strong consistent reads, while events can be used for establishing asynchronous communication. It is important to note that these patterns often overlap, and the process and integration layer can tap into different layers of the underlying data architecture.

Designing a robust application and data integration solution requires careful consideration and understanding of the business problem at hand. The reference diagram provides a consolidated overview of all the major patterns discussed in this article. It serves as a valuable resource for making informed tactical decisions, taking into account factors such as performance, maintainability, flexibility, cost, and resilience.

To ensure consistency and cohesion in data architecture and management, it is crucial to align all integration and data services. This can be achieved through the use of design guidelines, documentation, and standardized data and interface models. By connecting all interface attributes to the same set of elements or business terms within a catalog, organizations can create a unified and streamlined approach to data and interface management.

In conclusion, data mesh is not limited to analytical data alone. Operational use cases can also benefit greatly from the principles of data mesh. By treating data as a product and adopting a decentralized approach to data architecture, organizations can enhance the agility and efficiency of their data operations. This, in turn, leads to better business outcomes and improved decision-making capabilities.