Rarely does a technology I work with it make it to the pages of the Wall Street Journal, but Microsoft’s recent announcement that SQL Server will be available for Linux is one of those times.
I’m sure I’ll get questions about this. So, I thought I’d prepare my answers here. (If you want analysis of this announcement from a technical guru, check out Brent Ozar’s post.)
SQL Server on Linux is Irrelevant to Some
For my mid-market BI clients, the announcement is irrelevant. Yes, Linux has tremendous market share in many areas of web development and start up software. But in mid-market ERP and BI, it’s almost non-existent.
So if a customer asks about running SQL Server on Linux, I’ll make the point that, for them and other mid-market companies, SQL Server on Linux isn’t a mainstream solution.
And I really like mainstream solutions, for two reasons:
1. Continuity
If a bus takes out my team and I, the customer can move on (preferably after a brief period of mourning).
2. Lower costs
Mainstream solutions are more cost effective, which I’ll explain.
System Costs vs. People Costs
In a mid-sized business, the cost of an operating system is a small percentage of the cost of the software. It can be as little as five percent.
And the software cost is usually a small percentage of the lifetime cost of the people who implement, extend and maintain that software. In a simple environment, the software cost may equal the people cost. In a more complex environment, the software cost may be only 10 percent of the people cost.
So, if you want to keep costs down, you need to focus on the people cost. And to do that, you need a mainstream solution. Because with a mainstream solution, the required knowledge and skill sets are more widely available.
In addition, when you come across an issue (and all software has issues), you might even be able to find what you need through an Internet search. (This is absolutely NOT to say that the Internet is a substitute for training and experience.)
But if you’re on a platform used by a small number of similarly sized companies, getting help isn’t going to be easy. And the couple of thousand dollars you saved on operating licenses will be quickly burnt in labor and system downtime.
The same argument applies to SQL Server for Linux. If it takes off, the first people who know anything about it will be able to charge a premium—if you can find them.
And good luck finding any hints or tricks on the Internet.
Again, I want to repeat that this is true in my mid-market ERP/BI universe. If you’re developing the next great web app, the economics can be totally different. Which is why Linux is a dominant force in certain areas of the market.
We’re surrounded by vast amounts of different technologies. The key to success is learning what’s not relevant.