%402x.svg)
Models, hierarchies, business rules, users and every downstream feed, migrated to Profisee on Azure or wherever you run it, with no disruption to daily operations.
How long MDS has left, by SQL Server version
Ended July 2026
13 October 2027
9 January 2030
12 January 2033
MDS removed
End of extended support for each version, as published by Microsoft.
From our two MDS migrations: a top-tier Dutch commercial bank and a global financial trading institution
Microsoft removed MDS from SQL Server 2025 (version 17.x) and supports it only in SQL Server 2022 and earlier. The instance that hosts MDS cannot move to 2025 with it.
DQS is removed in 2025 as well, and an upgrade fails on an instance that still has DQS installed.
Duplicate detection, survivorship and quality checks sit in scripts and side databases that only one or two people understand.
An unsupported component sits in the customer or legal-entity master. Auditors and supervisors ask about it, and the answer gets harder every year.
Assessment: one to two weeks
An inventory of models, rules and consuming systems, the list of what the utility will not move, the target architecture, and a fixed price for the migration.
Build: most of the programme
Model, rules, matching, integrations and the deployment pipeline, built and tested in Profisee.
Parallel run: about three weeks at the bank
Both systems in production, with outputs compared every day by the operations teams until every discrepancy is closed.
Cutover and handover: the final week
The switch, MDS decommissioned, and workflows, roles and runbooks handed over to your team.
Every MDS model, entity, attribute, hierarchy, business rule, subscription view and the systems that read them.
Mapped in the assessment, before anything moves
Entities, attributes, hierarchies, relationships, users and groups, and the data itself.
Profisee's MDS Migration Utility, included in every Profisee subscription
Data quality rules, business rules, workflows, and the attribute and hierarchy types the utility does not support.
Rebuilt by hand and tested in Profisee
Matching and survivorship, so every customer, product or legal entity has one authoritative record.
Profisee's matching engine
ERP, CRM, the data platform and Power BI read from Profisee, and each consuming system is validated inside its own workflow.
Orchestrated pipelines (Azure Data Factory in both migrations)
Stewardship workflows, access roles, and one deployment pipeline for Profisee models from development to production.
Handed over with the knowledge to change them
One golden record per commercial client, for KYC, sanctions and lending, from up to 500,000 master data records in six systems. Leaving unsupported software removed the audit exposure, and five matching strategies and more than ten data quality rules replaced custom-built tools. Read the case study →
The client team runs the platform on its own: within weeks of go-live it was adding new data models and configuring workflows without us. Trading data, cost structures, legal entity hierarchies and reference data moved across ten entity types, on time and on budget, with ten automated data quality rules. Read the case study →
What Profisee does after the move: integrated with Dynamics 365 in 15 weeks, with product launches 40% faster, 16% fewer SKUs and 23% procurement savings. Read the case study →
Profisee as SaaS or PaaS on Azure, bought through the Azure Marketplace, or on AWS, Google Cloud or on-premises.
Your ERP and CRM systems, for example Microsoft Dynamics 365 and SAP, plus the spreadsheets and side databases that feed MDS today.
Mastered data delivered to Databricks, Snowflake, BigQuery or Microsoft Fabric, where Profisee also runs as a native Fabric workload. Power BI reads it through Profisee's own connector, and Microsoft Purview keeps the catalogue and lineage.
The same team builds what reads the golden record: the data platform and the Power BI reports it feeds.
Yes. Microsoft states that MDS is removed in SQL Server 2025 and remains supported in SQL Server 2022 and earlier. Extended support for SQL Server 2022 ends on 12 January 2033.
You can, until its support ends. The instance that hosts MDS then stays on 2022, and Data Quality Services blocks the upgrade too. Planning the move alongside your next SQL Server upgrade keeps it to one project.
Profisee is a subscription, and on Azure you can buy it through the Azure Marketplace. Book a 30-minute call and we will go through what the licence and the migration would cost for your MDS estate.
It moves entities, attributes, hierarchies, relationships, users and groups and the data. Data quality rules, business rules, workflows and some attribute and hierarchy types are rebuilt by hand. We list these in the assessment, so the price covers them.
It took 10 weeks at the bank and 12 at the trading institution, from assessment to cutover, including a three-week parallel run at the bank. The timeline scales with the number of models, rules and downstream systems.
Microsoft's own Purview documentation calls Profisee “the premier upgrade and replacement solution” for MDS. It runs as SaaS or PaaS on Azure, as a native Microsoft Fabric workload, and connects to Purview, Power BI, Databricks and Snowflake. Astral Forest is a Profisee partner.
The ones you already run: Databricks, Snowflake, BigQuery, Microsoft Fabric and Power BI on the analytics side, and ERPs such as Dynamics 365 and SAP on the operational side.
A list of your models and the systems that read them is enough. You speak with Michał Dębski, co-founder and the solution architect who would run your migration, and you leave knowing what it takes and what the utility will not move. The migration starts with an assessment of one to two weeks, which is part of the migration and fixes its price. The bank and the trading institution have since referred us to other banks.
