SYNLAB group partnered with Expert Extend to migrate on-prem databases to Microsoft Fabric using medallion architecture, resulting into report refresh time reduced by 50% and giving data and governance teams a unified, self-service workspace.
SYNLAB Group, a leader in medical diagnostic services and specialty testing in Europe operating in more than 20 countries across 4 continents, ran its reporting platform entirely on-premises, with Dataflow Gen1 handling transformations directly against on-prem database tables, a setup that was becoming a bottleneck as data volumes grew. SYNLAB Group partnered with Expert Extend to migrate this platform to Microsoft Fabric, rebuilding the pipeline on a Bronze–Silver–Gold medallion architecture; by consolidating storage, transformation, and analytics onto OneLake, the migration eliminated the legacy bottleneck and gave SYNLAB’s data and governance teams a single, unified workspace, with self-service access to Gold-layer datasets and country-level access controls to support compliance across markets. The result was a roughly 50% reduction in report refresh times, alongside a materially stronger governance and data-access posture.
Repeat data-platform engagements for healthcare and laboratory-diagnostics clients across Europe, delivered under strict data-protection requirements.
Functional exposure to laboratory reporting, clinical operations KPIs, finance, country-level data governance and compliance-sensitive data handling.
Certified data professionals covering Microsoft Fabric, OneLake, Power BI, Azure data engineering and DevOps.
High customer satisfaction, follow-up engagements and recommendations constitute more than 90% of our projects.
SYNLAB is a clinical diagnostics provider whose reporting depended on data drawn directly from its on-prem source systems. Before the migration, transformation was handled in Dataflow Gen1, connected straight to the source tables, with no dedicated storage layer for the transformed output. Below is how the previous architecture compares with the current Microsoft Fabric medallion approach.
Previous Architecture
– Directly connected to the source data, putting a load on the transactional systems.
– Uses Dataflow Gen1 for all transformations.
– All transformations happen in one place.
– Repeated source-system load, as multiple dataflows use the same data source for different use cases.
– No destination storage system for the transformed dataset.
– No row-level security (RLS) or data masking on the reporting data.
Current Architecture
– Holds a copy of the source data, eliminating repeated source-system load.
– Uses a Warehouse as the Gold layer.
– Transformation can be done as and when required, in between the layers.
– Raw data is reusable, so other teams can transform it for their own use cases.
– Transformed datasets can be stored at any desired storage layer.
– Row-level security (RLS) and data masking control data access by users.
The core bottleneck was the Dataflow Gen1 refresh chain. Dataflow Gen1 pulled data from the on-prem database and applied transformations, and the semantic model then layered further transformations on top before refreshing the reports. This long, sequential chain made report refreshes slow, and the platform offered no way to control who saw what.
Dataflow Gen1 connected straight to the on-prem database, placing a heavy load on the transactional systems and slowing the movement of data from the on-prem database to the semantic model. With multiple dataflows hitting the same source for different use cases, that load was repeated again and again.
The semantic model applied further transformations on top of the data already transformed by the dataflows before refreshing the reports, adding still more time to each refresh. Every report refresh had to wait for this entire sequential chain to complete.
The reporting data carried no row-level security (RLS) and no data masking, so access could not be controlled at the user level, a significant gap for a diagnostics provider operating across many countries.
With transformation logic, source connections, and reporting data spread across the legacy chain, there was no single place to govern the platform, gaps that could only be addressed by consolidating all data in one environment.
Expert Extend adopted a medallion architecture (Bronze–Silver–Gold) approach for data orchestration into Microsoft Fabric: Bronze and Silver layers as lakehouses, and the Gold layer as a Warehouse. Each stage of the build removed a link of the legacy refresh chain while adding the governance SYNLAB needed.
Dataflow Gen1 was replaced with data pipelines that migrate data from the on-prem database into the Bronze layer. Shortcuts hold a copy of the source data directly in Bronze, which eliminates repeated source-system load; the raw data lands once and becomes reusable for every downstream team.
Transformations were distributed across the medallion layers rather than a single monolithic dataflow. Raw data lands from the on-prem database into Bronze; Spark notebooks apply aggregations, filters, and other required transformations as data moves from Bronze to Silver; and the selected, refined data is then promoted from Silver to the Gold layer.
Row-level security (RLS) and data masking were applied in the Gold-layer Warehouse, scoped by country: each country’s users see only the data for their own country. Governance is now centralised in one place, with self-service access to Gold-layer datasets for the teams that need them.
Semantic models now connect directly to the Warehouse tables and views, so reports always take the latest data from the Gold layer. The dataflow has been eliminated, refresh time is driven by data pipelines instead, which can be optimised to land the latest data in the Warehouse in minimal time.
By rebuilding the platform on a Fabric medallion architecture, SYNLAB turned a slow, opaque refresh chain into a fast, governed data asset:
✓ Report refresh time reduced by approximately 50%.
✓ A single, unified platform serving the entire user base.
✓ Governance centralised in one place, improving access-level management through row-level security and data masking.
✓ Improved scalability, more tables can be mapped into the reports and larger data volumes held in the Warehouse, while remaining easy to maintain.
✓ Eliminated the load on the on-prem systems, resolving throughput concerns as data volumes grow in the future.
✓ Enabled CI/CD pipeline integration with the current architecture, giving greater flexibility to make and deploy any future changes.
Cut report refresh time by ~50%, enabling more timely operational and clinical decisions.
Country-scoped access controls closed a real data-exposure gap across SYNLAB's 20+ country operations.
Replaced a fragmented legacy chain with a single place to manage data access, quality, and change.
Removed on-prem load, allowing data volumes and use cases to grow without re-architecting.
CI/CD integration replaced manual, high-risk pipeline updates with faster, safer deployments.
Before undertaking a migration like this, it helps to know exactly where your current environment stands. Our Microsoft Fabric Transition Sprint assesses your data platform and gives you a clear, low-risk path to migrate with confidence, the same kind of foundation that shaped SYNLAB’s results.
Navigation
