We have several interesting tendencies as an industry. My favorite examples include coding everything on our own and reinventing the wheel 100 times over. Another one is coming up with industry standards: “Hey, we just got used to convention A, but let’s definitely introduce convention B - it yields the exact same results but makes life just a little more miserable."
Why do we do that? My best guess is that we all believe we are a little smarter than average. A dangerous assumption. Another reason is vendor lock-in. That’s right. A nasty topic that no one admits, but it's still true. I do think that this is the number one reason. This is why I’m dedicating this week's newsletter to just this topic: lock-in!
3. Lock-in has many faces
I’m a big fan of Gregor Hohpe, his books, and really most of what he says. He wrote an amazing post on “Don’t get locked up into avoiding lock-in”. His basic argument: your application life-time is probably so short anyway, you shouldn’t spend too much time worrying about lock-in.
A pretty strong statement and I like it, though you can’t fully trust the guy’s intentions - Gregor is on the payroll of our friends at GCP and AWS and lock-in is what makes the world go around. I’d suggest you want to look at the principle of Pareto efficiency. If you can get low lock-in without compromising on developer experience or creating extra work: take that route.
In my opinion, the single biggest example of Lock-in is probably AWS Lambda. I know first hand that this was one of the big drivers inside of AWS engineering management and the reason why they pushed out Kubernetes support for so long. This article talks about it in depth (early days from 2017 but still really relevant).
But at Humanitec we like to be impartial or try to at least. So I’ll also share the counter-view on this by Thoughtworks, check it out!
2. What is Humanitec doing to avoid lock-in?
This week at Humanitec we made sure we don’t fall into this trap ourselves. We released a feature called “export manifests”. As the title suggests this enables users to generate a fully functional representation of their configuration that you can run out in the wild. So we follow the industry standards of Kubernetes manifests. And we save you a lot of time and money even if you eventually want to leave or even, if your dev infrastructure differs from production and you don’t want to manage everything with us or if you’re into Gitops. All of this is possible now. Here’s our release announcement if you want to take a look.
I’m also excited to share one of our nicest content pieces in a while: our interview with Jan Löffler on how and why they built an Internal Developer Platform (IDP) at Zalando. Jan is Plesk’s CTO now and just really dives into the details. Really worth checking out!
1.
To wrap up, we are organizing a really cool webinar with our partners at CircleCI that you might want to check out. We will guide you through how to build your Internal Developer Platform with CircleCI and Humanitec.
Anyhow, stay lock-in free or free while avoiding lock-in, just as you please.
