Seamless Business Intelligence Migration for a Better Data Future

Utilizing migVisor™ to Quickly Migrate to Amazon Redshift, Powering Zalando with Improved Access to Business Data

At a glance

CLIENT

Zalando

STRATEGIC PARTNER

Amazon Web Services (AWS)

SERVICE

Cloud Data Migration

INDUSTRY

Retail

Zalando, a leading European fashion and lifestyle platform, has been a longstanding client of EPAM for several years. As part of our work in modernizing its data platforms and delivering advanced analytics, we worked together to migrate Zalando’s existing data warehouse (DWH) platform that fuels its business intelligence (BI) to Amazon Redshift. Leveraging the power of AWS and EPAM’s migVisor tool, we completed the migration seamlessly and ahead of schedule. With reduced query times and improved data flow efficiency, Zalando’s new modern BI platform can support its 3,000 users and 50+ data teams with enhanced scalability and improved data management.

MIGRATION STRATEGY

Accelerating Migration with AWS and migVisor

Zalando, a leading platform for fashion and lifestyle eCommerce, has been partnering with EPAM and AWS since 2022 on over 15 programs and 25 data initiatives. This collaboration focuses on modernizing data platforms, delivering advanced analytics and developing AI use cases and models.

As part of its efforts to evolve its data platform and transition technical solutions to AWS products, Zalando conducted an analysis of its existing DWH platform that fuels BI and identified several challenges:

Elasticity

Zalando constantly had to pay a high price for peak loads that were only needed for a few hours each week.

Scalability

The platform was unable to keep up with growing data demands, impacting performance.

Operational Risks

Fixed session limits and intricate operational routines posed risks, particularly during peak periods, like Cyber Week sales events.

Cross-Team Complexity

With over 50 data teams relying on it, the DWH platform was a critical component of Zalando's business infrastructure.

Code Conversion & Modernization

Modernizing the legacy code, including Lua-based logic and custom routines, required re-architecting processes and redesigning workflows, adding significant effort and risk to migration.

After analyzing the platform and developing a migration strategy to address these challenges, Zalando turned to EPAM to support executing this strategy effectively and securing AWS Migration Acceleration Program (MAP) funding to cover the cost. Leveraging AWS tools and solutions, as well as EPAM’s migVisor data migration tools, we helped the team migrate to AWS Redshift.

We synchronized the data between the Redshift cluster and the incumbent DWH with the DataHub and then migrated the workloads to produce the data. This approach enabled us to work in parallel and reduce dependencies between migration efforts of the teams as the data was already there — decoupling storage and compute to enhance scalability and cost attribution.

A phased conversion approach was adopted to minimize disruption and risk, where teams from both EPAM and Zalando created objects and transformation logic in Redshift, using the legacy DWH platform as a data source. Finally, individual data teams from Zalando verified and gave final acceptance to all Redshift objects, ensuring complete accuracy.

Our migVisor suite of tools helped us achieve a successful migration: migVisor Analytics streamlined the migration scope through detailed data lineage, migVisor Convertor improved efficiency and migration time, and migVisor Reconciler ensured data accuracy and reduced validation efforts.

The migration was completed on budget and ahead of schedule within the 16-month planned timeline. As a result, Zalando was able to achieve better query performance, improved data flow efficiency and enhanced data management. With increased scalability, Zalando is now able to keep up with the demands required by the more than 3,000 business users and 50 data teams that leverage the platform.

RESULTS

3,000+

end users leveraging the new platform

50+

data teams successfully migrated

43%

reduction in reporting query execution times

60%

improvement in extensive report execution time

TECH STACK