Platform as a Service (PaaS) allows us to create and launch applications in no time at a fraction of what it would have historically cost.
Not so fast.
The provider of PaaS has a goal to make money with their service. They will realize that they didn’t do somethings right and they need to “tweak” what they provide to keep their costs under control, open a new market, or deprecate features that are not used and it no longer makes sense to maintain. Or worse, deprecate the entire platform because it didn’t prove financially viable. All understandable.
However, the platform providers goals don’t necessarily align with the goals of the developer who is consuming the platform. The developer used certain features of the platform so that they could focus on their value adds to their customers. Now that the platform is taking care of that grunt work they are solely focussed on their clients. Except the platform just changed an API, deprecated a feature, etc. etc. Now the developer is working to the agenda of the platform. Their goals for their clients have been side tracked because the platform changed. It may be worth it, it may be a change that helps the developer, but from my experience all too often it isn’t. The developer bought a Platform as a Service to enable them to do something they didn’t want to do. If the platform has done that job they just want to count on it being there so that they can build the truly specialized stuff.
Yes, PaaS allows us to come to market fast, but I am not confident that it enables sustained velocity for the developer who is using it towards value creation for their customers. Choose PaaS with eyes wide open and be prepared for the agenda changes, or probably better, plan to migrate from the platform to what you control once market fit has been proven for your product. Or, bypass PaaS from the start, pick an open source solution for the work you need done, host it on a scalable service, and control all functions of the product that you provide.