This is kind of an odd article. I'm not sure if my understanding is at fault or the author's. I don't quite see the overlap / conflict between Docker and PaaS that the author seems to be talking about, but I may just not quite have got what they mean.
The point of PaaS to me is that I have someone else providing the services that I need and these people (Amazon, Azure et al.) are going to know a lot more about managing it all than I do. Or quite frankly even if they don't, I want to focus on my applications, not on administration of all these services.
That's why PaaS is my GoTo. IaaS is my fallback if PaaS can't do what I need because with IaaS I can do a lot more, but at a very marked increase in work outside my core business. As PaaS gets better and better, I see the need for IaaS reducing for my use cases (and many people's). I think there will be a big shift away from IaaS towards PaaS because if the latter gives you what you need and lets you focus on what you really care about, why wouldn't you?
But where Containerisation fits into the above conversation, I'm not quite sure. I think the author is saying that Containerisation obviates the need for PaaS to some extent? Or that people think it does and he's saying that actually people are wrong and it doesn't... But I'm not quite clear. Containerization is, I guess you could call it, a refinement / supplement / tool for IaaS and thus keeps parity between the two approaches, offsetting the complexity of IaaS to some extent? I'm not sure if that's the argument.
Paris, because maybe I'm just being like her today.