Would Open-Sourcing Microsoft SQL Server Actually Fix Anything?
At the PASS Summit last week in Seattle, I was having a discussion with a few people who’d moved on from managing SQL Server over to working with PostgreSQL. They were talking about the advantages of open source – and there are certainly many – but it prompted me to ask, “If Microsoft open-sourced SQL Server, would that actually fix any of SQL Server’s problems?”
First off, open source means that when you’re experiencing a problem, and you believe it’s a bug in the product itself, you have the ability to read the source code to get to the bottom of the issue. That wouldn’t have made much difference to me in the past because I probably wouldn’t be able to make heads or tails of the source code. However, today in the age of AI, it’d be much easier for me to point Claude or ChatGPT at the SQL Server source code repo, explain the problem, and have it go churn through the code looking to find the root cause.
However, just because you can prove that there’s a bug, doesn’t mean Microsoft will fix it.
Six months ago, I identified a bug in sys.dm_exec_query_stats, reported it through the official channels along with a simple query to reproduce the issue, and Microsoft … simply ignored it. It took repeated public shaming over social media before they agreed to ship the fix in some future cumulative update. Open source doesn’t fix that.
Even if you coded a fix, and submitted a pull request, that doesn’t mean Microsoft would accept it. The documentation team has done a fantastic job of managing pull requests in their Github repo, staying on top of them and merging contributions in quickly – but they’re a rare shining example at Microsoft. For example, the VS Code team has over three thousand open pull requests, dating back months, and just clicking around through random ones shows that a lot of people have put a ton of work into VS Code bug fixes and improvements that are dying on the vine. Open source doesn’t fix the problem of the software vendor not dedicating enough resources to handle public contributions.
If anything, Microsoft has been backing away from its open source work around SQL Server. The documentation team I was just touting? Their open source work ends next month, and they’ll no longer accept pull requests. The open-source Azure Data Studio was shut down earlier this year, but even when it was active, there were usually thousands of open issues at any given time, and the team wasn’t given enough resources to make progress. The still-active VS Code has thousands of open issues, too. Open source work makes life much, much harder for software vendors: they get a deluge of unqualified code contributions, inaccurate bug reports, and whines about feature requests. It would take more people to sift through all the noise to get to the good signals.
Open sourcing SQL Server would fail.
Not because open source is bad, or that closed source is good.
It’s just that Microsoft’s awkward, expensive support process has served to keep the bugs quiet, shield the developers from having to fix those bugs, and generate more revenue through support.
If they open sourced it, people wouldn’t need to pay as much for support or licensing – so Microsoft’s revenue would go down – while at the exact same time, their expenses would go up because they would need to deal with the avalanche of pull requests and the exposure of now-buried bugs.
You could argue that in the long term, the product’s quality would eventually go up, and that the user’s experience would be better – and I’d agree. But in the short term, revenue would go down, expenses would go up, and there’s no business manager that’s going to advocate for that.
Instead, Microsoft will continue to invest in Postgres because somebody else deals with the bug reports, code reviews, and support requests. It’s an easier way for them to make money by selling a hosted Postgres service without having to bother supporting an on-premise boxed product version.
That’s not to say they’re not investing in SQL Server! They are, especially since a similar engine powers stuff in Azure SQL DB and Fabric – but I don’t believe open source would ever be in that product’s roadmap because it wouldn’t fix anything, but rather make things worse. I love open source – but it just doesn’t make sense for Microsoft’s data platform staffing levels.
Free, 3× a week
Get my new posts by email
Three posts a week, plus a Monday roundup of the best database news from around the web.