Microsoft MDS end of life: your options before 2028

Microsoft MDS end of life is no longer a rumour. Microsoft shipped SQL Server 2025 without Master Data Services, and SQL Server 2022 is the last version that has it. If your data warehouse reads its customer groups, cost centres or product hierarchy from MDS, you have a deadline. It is probably earlier than you think.

The MDS end-of-life dates

Microsoft removed Data Quality Services (DQS) from SQL Server 2025 too. If you run DQS next to MDS, plan for both.

The real deadline is your next upgrade

The patch dates are not your deadline. Security fixes for SQL Server 2022 keep coming until 2033, so MDS keeps working until then.

Your real deadline is the day you must upgrade that SQL Server instance. MDS does not exist in SQL Server 2025. So the server that hosts MDS cannot go to 2025 while MDS still runs on it. If your company upgrades every few years, or an auditor asks for supported versions, that day comes sooner than 2033. Plan the migration before the upgrade forces your hand.

Check your servers: is MDS still in use?

An MDS database always has the table mdm.tblSystem. This query lists every MDS database the login can open. It skips databases it cannot enter, so it does not fail on error 916:

1
SELECT d.name
2
FROM sys.databases AS d
3
WHERE d.database_id > 4
4
  AND d.state = 0
5
  AND CASE WHEN HAS_DBACCESS(d.name) = 1
6
           THEN OBJECT_ID(QUOTENAME(d.name) + N'.mdm.tblSystem', N'U')
7
      END IS NOT NULL;

Do not search for the mdm schema alone. Other products use that schema name too. On my own test server, a check on mdm.tblModel found a database that was not MDS. The check on mdm.tblSystem did not.

Then see which models still live. Run this in each MDS database:

1
SELECT Name, LastChgDTM
2
FROM mdm.tblModel
3
ORDER BY LastChgDTM DESC;

A model that nobody changed in five years may not need a migration at all. A model your ETL reads every night does. Also check your SQL Server version with SELECT @@VERSION. On SQL Server 2019 or older, mainstream support has already ended.

First decide what your MDS really does

Most of the choice follows from one question. Does your MDS hold reference data for a data warehouse, edited by a few people and read by ETL? Or is it the start of company-wide master data, with customers and products from many systems that must be matched and merged?

For the first, you want something close to MDS that runs where MDS runs. For the second, you are shopping for an enterprise MDM suite, and MDS was never really that.

Write down what your models use before you compare tools: subscription views, staging tables, business rules, hierarchies, versions, the Excel add-in. Those are the parts that must have a place in the new tool.

Option 1: stay on SQL Server 2022 until 2033

This costs nothing today. It also freezes that instance: you cannot upgrade it, and the MDS web app does not get better. Microsoft lists Internet Explorer 9 or later as the only supported browser for Master Data Manager. Choose this only if you plan to retire the data in MDS anyway, or to buy time for a migration you have already planned.

Option 2: build it yourself on SQL tables

Plain tables, a few stored procedures and an edit screen. This works for one or two simple lists. For more, you rebuild what MDS gave you: validation, approvals, an audit trail, security per entity and the views your warehouse reads. You also maintain it yourself, for as long as it runs.

Two tools can take the edit screen off your hands.

SQL Spreads puts an Excel front end on SQL Server tables. It sets up in under a day and checks data types. It has no approval workflows and no hierarchy management. Prices start at USD 39 a month per Designer seat (details).

Microsoft Dataverse with Power Apps keeps you in the Microsoft stack and connects to Power BI. It is cloud only and needs Power Platform licences. It is not built for master data management, so approval workflows, hierarchy management and subscription views are missing.

Option 3: a replacement that works like MDS

These tools aim at teams that leave MDS and want the same way of working.

Primentra

Primentra runs on your own SQL Server, from SQL Server 2016 and the Express edition up, on Windows 10, 11 or Windows Server 2016 or later. It uses the MDS building blocks: models, entities with a Code and a Name, free-form and domain-based attributes, one staging table per entity, and permissions from model to entity to attribute. Active Directory groups map to roles.

Its MDS Direct Migration reads the MDS database with a SQL login and runs a dry run first. It moves models, entities, attributes, members, explicit hierarchies (as a Parent attribute), subscription views, and MDS users and groups. A run can be undone, and a second run brings over only what changed. MDS business rules do not migrate yet; the run lists them so you can rebuild them. The vendor says the install and a typical migration each take under an hour (migration details).

It does not have collections, file attributes, per-row permissions, a rollback of saved changes or a cloud edition. The price is EUR 7,500 a year per production server, with unlimited users. Development, test and acceptance servers need no licence. A trial runs 60 days. Feature requests go through a public tracker, where you can see what is planned (pricing, tracker).

CluedIn

CluedIn is Azure-native, graph-based MDM, and it has its own page about moving from MDS. It is cloud-first, the price starts at about USD 28,500 a year and depends on volume, and Azure costs come on top.

Grunddata

Grunddata hosts reference data that people edit in Excel or Google Sheets, from the EU. It aims at the same teams that leave MDS. It starts at EUR 1,500 a month for one environment and 10 users, so EUR 18,000 a year or more. Your data then lives outside your own network.

Option 4: an enterprise MDM suite

Profisee, Semarchy, Informatica and SAP MDG do far more than MDS ever did, from match and merge at scale to data quality tooling. You pay for that, and the projects take months. If your master data spans many source systems, or SAP sits at the centre, this is the category to look at.

Profisee fits the Microsoft stack well, with Azure and SQL Server, and has stewardship workflows and match and merge. Semarchy has a modern interface and a flexible data model, but it is not native on SQL Server. Informatica brings deep data quality and many connectors. SAP MDG is mainly practical when SAP is your system of record.

The alternatives side by side

Prices and setup times come from the vendors and from the comparison pages at primentra.com/compare and primentra.com/mds-alternatives. They change. Ask each vendor for a quote before you decide.

How to test an alternative

Whatever you pick, test it on a copy of your real MDS model, not on a demo. Restore a copy of the MDS database on a test server. Move one model that your ETL reads every night. Then point a copy of that ETL at the new views, and compare the rows with what MDS returns. If the counts match and your editors can do their daily work, you have your answer.

Still working in MDS until then? This saves a lot of clicks: reorder MDS attributes with SQL instead of the web app.