Data is everywhere these days, with the distribution of data and the proliferation of tools for every task. The opinion of our Data expert, Frédéric Périn, Business Developer - Data Management, sheds light on how the Data Mesh can respond to the problems of Data Management in companies.
What conclusions can you draw from the Data Mesh about data in organisations?
My first observation is that data is increasingly distributed across organisations. Today, organisations and businesses are looking to use the best tool for each task, whether it's a datalake, datawarehouse, machine learning or AI ..... Businesses, for their part, are looking to generate and identify more leads using tools such as PowerBI or Tableau, for example.
Large organisations use more than 400 datasources* to analyse their data, and 94% of AWS' best customers use at least 10 different database and analysis services**. A single system is not enough to meet the needs of organisations.
Finally, the data repository lacks flexibility: it is difficult to maintain this repository at IT level and difficult to use it on the business side. This leads to a loss of confidence in the KPIs provided by the repository. At the same time, business users are reporting a need for flexibility linked to business requirements, in search of rapid decision-making.
The Data Mesh organisation, supported by a relevant technological foundation, will provide a response to the observations made by organisations when it comes to managing their data.
How can the Data Mesh solve business needs in terms of data flexibility and agility?
I believe that the Data Mesh enables everyone to refocus on their core business and extract more value from data. The Data Mesh concept allows IT to focus on data collection, storage and security. Businesses, for their part, will focus on data consumption, analysis and strategic decision-making, which will also meet their need for flexibility.
Data Mesh consists of decentralising data management within the corporate eco-system. It is not a technical approach (although the technique contributes to it) or a business approach, but an organisational approach.
In short, datamesh aims to:
- Reduce friction between departments by decentralising business expertise linked to IT
- Decentralise data management
- Maximize value
- Reintegrate confidence in data
What are the fundamental requirements for setting up a data mesh organization?
First, you need to put in place a structure within which the business lines own the data. We're talking about data domains and APIs, for example. We then need to treat the data as a product. The aim is to be able to transform data into a product that can be used internally and/or externally, in other words to be a producer and consumer of data.
Then we are going to implement self-service data by eliminating the company's data silos, which are often multi-source and cross-functional. We are going to enable data interoperability from a technological and organizational point of view. Added to this is federated governance, which will provide the organization with the means to put in place a flexible and distributed organization to meet business needs, while guaranteeing IT compliance (RGPD / anonymization, etc.).
All these fundamentals will contribute to the implementation of data mesh in organizations. It is important to understand that data is society's most precious asset. Data Mesh will enable it to be exploited in the best possible way.
What approach does DEEP propose for implementing data mesh in organizations?
To set up a data mesh, I recommend starting with the following 3 essential pillars:
- Dataplatform
- Logical datafabric
- Federated governance
These 3 pillars of the Data Mesh enable the concept to be properly implemented in organisations.
* IDG, 2023
** AWS Analytics Data Strategy, Keynote
