Last week I had the pleasure of traveling to Redgate headquarters in Cambridge, UK. It was my first time traveling to England, being at the main office, and meeting so many coworkers IRL. I had (Gluten free!) fish & chips, a proper English breakfast, and came home with a sampling of traditional English sweets to try with the family. It was a delightful time.
Obviously, my English adventure isn’t really what this brief article is about, but I thought I’d set the stage for why licensing is on my mind recently. (although can we all just agree that they’re cookies… not biscuits? I digress…)
Do Open-Source Users Purchase Tools?
I joined Redgate almost six months ago because I have used and love their tools from my SQL Server days and because I appreciate how supportive they are of the database community. In the few months that I’ve been here, I continue to see that genuine spirit of support in every conversation across teams. 🎉
That doesn’t change the fact that Redgate is still a business that creates great tools to support top-notch Database DevOps organizations. It’s because of that success that they can continue to invest in the communities that they serve.
Which is why it’s not surprising that I’ve had some form of the following question asked of me nearly every week since I’ve joined the company:
“People/companies seem to be drawn to Postgres because it’s free and open-source. Do they purchase tools to support their work when the database is free?”
Typically, my response articulates my personal experience managing development teams and organizations, including our need to find the right tools to help us scale. In our case, that usually meant that the most important parts of our business were supported by applications that were 💯 in areas like monitoring, alerting, schema change management, and more.
I believe that’s true for most organizations, and I’m guessing many of you have a similar experience. As developers, we’ll often work to DIY a solution up to a point, but at some threshold level of effort, we need help keeping all of the plates spinning efficiently. Once we reach that point, it makes sense to find a company and solution we can trust to manage those burdens. This frees our team to focus on other, more important tasks that a solution can’t help with. You know, the stuff that makes your application or solution yours!
So yes, companies will spend money on tools that help them solve problems, regardless of how cheap or expensive the database licensing cost is. A free piece of software that doesn’t perform well is still costing everyone a lot of money and time.
But I realized last week that I was missing an important piece of the licensing discussion, even though it was a part of nearly every decision we made in my teams.
It’s the Cognitive Load, Silly!
As part of my visit to Cambridge I lead a discussion on what Postgres is, how it’s different from SQL Server, and why organizations are increasingly turning to it for their relational database needs.
As I was preparing the slides for my presentation, I came to the “Postgres is free, so why would anyone purchase tools” slide again and a totally new thought came to mind. Typically, I spend a few minutes talking around the permissive license in context of how it saves money and how people will evaluate the need for other products that aren’t open-source. I usually say something like, “While most organizations will probably reduce their capital spend by moving to PostgreSQL over the long-haul, there’s no magical rule somewhere that says it will always work out that way. That’s 100% true for everyone.”
Instead, I realized I hadn’t been discussing one of the biggest reasons companies are drawn to the freedom that open-source licensing provides with a product like PostgreSQL.
It reduces the cognitive load around licensing.
To some extent, regardless of how many ROI calculations you do, there is a (somewhat) unmeasurable impact of choosing PostgreSQL. During development planning sessions or yearly budgeting conversations with the finance team, organizations don’t have to cross-reference a grid of features, versions, server core comparisons, and more.
I can’t tell you how often I used to sit in planning meetings with questions like these…
- Can we do online (concurrent) index rebuilds?
- Why isn’t the server using all the memory we’ve allocated?
- Can we have a read-only warm standby?
- Will that Indexed/Materialized View be used without extra work?
…to which the answer was often, “Our current license doesn’t support that.”
I’m not dunking on commercial database products at all. (I mean that!) This was just a new way of looking at a problem I had experienced many times but didn’t fully appreciate the way our processes and discussions changed because of it.
It’s a total game changer when the team isn’t focused on managing available product features to reduce (and manage) licensing. Thinking back and considering the hundreds (or thousands??) of hours that my teams had to invest understanding licensing, the cognitive load aspect takes on new meaning.
Is That Really Surprising?
I know. Most of you are probably thinking, “Come on Ryan! That’s literally the most obvious thing you could say about licensing with open-source or Postgres. Sorry you’re so slow on the uptake, dude.” (Truth be told, I’m slow on the uptake for many things, and maybe this is just another example.)
Regardless, I think for many people this is a helpful (new) perspective on one of the main drivers that leads organizations to choose open-source, and specifically PostgreSQL. All too often the focus goes straight to a conversation about reducing operating costs. But I think that’s only part of the conversation and I’m glad I was able to articulate it to the team.
PostgreSQL has often been ahead of its time on many things. This is just another example of how that plays out in the trenches. 😊
3 thoughts on “The Cognitive Load of Licensing”
What will also matter is administrative overhead. Procurement can be quite the daunting task, getting worse with the size of the organisation. What budget can the DB be paid from? Is it a permanent budget or project-bound? Whose signatures do I need for that and how many pretty ppt slides do I have to draw to get them? How many meetings will that require?
It’s just way easier to simply grab PostgreSQL from the shelf next door and worry about such stuff later, when a project is already running and it’s too late to decide for another DB engine anyway. Especially given the comfortably large ecosystem of consulting firms who know how to grow clusters way beyond reasonable limits nowadays 😉
Absolutely! Great point. It’s these kinds of little things that add up over time into major savings!
Great thoughts Ryan – and no, I don’t think you were slow on the uptake 🙂
I don’t think I had thought about the PostgreSQL license in this way before you articulated it. Thank you!