Retail

Leaving Hadoop for the Cloud

#Cloud

#Dataiku

#Migration

#Ambition

Ambition
of the mission

Hadoop platforms have long been the key solution for bringing data to scale... but at the cost of severe technical constraints! Administrative complexity, lack of flexibility and scalability, technological incompatibilities... In order to address these difficulties, our client finally made the decision to migrate its data platform to a technologically different Azure environment. So how to carry out a large scope of BI and AI projects, without having to reimplement everything?

#Method

Our
approach

Think

Rethinking the ecosystem through a resolutely cloud approach

How do you get out of the Hadoop ecosystem? It was necessary to rethink the technological stack, with the strong constraint of covering the same functional perimeter, while gaining in flexibility, simplicity and performance! This thinking led to the deployment of a platform centered on Snowflake, Kubernetes, and Dataiku.

Make

Reconciling Run and Technical Migration

The migration of such a broad scope, based on a total change in technological paradigm, generally means a complete overhaul of what already exists. In order to save time and reliability, a migration method to automate redundant tasks through the development of dedicated tools has been implemented in order to guarantee service continuity for business teams.

Scale

Gain development capacity

This cloud migration was an opportunity to extend the capabilities of the customer's platform, through the introduction of cloud technologies that were previously inaccessible. Our team thus trained the group's architects in Azure technologies, expanded the functionalities of the initial platform, and then supported the datalab business teams on best practices for using these tools.

#Benefits

Indicators
of success

Technical solution more flexible and efficient

Migration more than 100 projects

Scalability gains and stability

#DataStories

Discover our
other use cases

See all our use cases