I heard Brent Ozar recently talked a bit about the SQL Server platform and its future. He also mentioned that Fabric has distracted the data platform team and it isn’t a great product. I tend to agree, and I see too many bugs, holes, and problems. However at the end of this short snippet, he talks about SQL Server with an interesting comment.
Is SQL Server feature complete?
That was Brent’s opinion, which is one that I tend to share. I think that the platform is very feature-complete. There aren’t a lot of things I think I really need in order to choose SQL Server as my database. I wish some functions (FORMAT, MERGE) ran faster, and there are a few items (AGs, replication) that could be easier to work with or were more robust under load. However, overall, SQL Server runs well.
It’s a good choice, there is mature tooling available to help, it’s well understood, easy to administer for the most part, and there are lots of people that have experience on the platform. There are ample reasons to choose SQL Server as a solid relational database platform.
At the same time, if the product is feature complete, then that gives PostgreSQL, MySQL, and assorted other platforms a target to aim for and potentially displace workloads at a lower cost. Even if the features aren’t quite the same, a much lower cost can be enticing.
That ignores the cost, often a very high cost, of switching platforms. However, I do see plenty of people investigating other platforms, not to migrate or move, but for new work. That makes sense, and I suspect that is part of the reason that Microsoft keeps trying to raise the bar with new features. I’d prefer they focus more on stability and performance than new stuff, but I get that doesn’t always sell well.
Are you happy with SQL Server? Looking elsewhere? Ready to learn a few platform? Let us know today.
Steve Jones
Listen to the podcast at Libsyn, Spotify, or iTunes.
Note, podcasts are only available for a limited time online.
Voice of the DBA 
Seems to me like as if Fabric is trying to be a one-tool-fits-all solution which is never a good approach. You have one tool that does it all but nothing very well. It also appears a bit gimmicky to me.
I’m a firm believer in specialized tools that have the ability to work with other specialty tools via some form of passing info/data between the 2 and not an all-in-one. Within the accounting software my industry uses there are 3-5 custom reporting tools or options and no one is ideal/great for every possible type of report. 1 is great at financials but bad with letter/notices while one is good with document style reports like contracts but is absolutely not the right choice for transactional reports. The software vendor has an annual users event and for several years in a row the 1 hour session on the available tools and how you determine which one to use was very popular. This vendor is now taking the complete opposite approach and is moving to a one tool does it all and I am so aghast with how foolish a mistake this is. They are ignoring their own advice they gave customers for over a decade all to make things easier on them; it’s less work to provide support for 1 tool versus several.
Advances in software apps/systems shouldn’t be steps backwards and yet this is what I’ve seen as of late and not just with this one vendor. Microsoft’s push to the new non-customizable OL is a huge leap[ backwards for users. It may make Microsoft’s job easier but that’s not what software advances are supposed to be about, making it easier on the software vendor at the cost of functionality and ease of use for the customer.
LikeLike