I didn't want to write a top-level comment and I'm sure few people will see this, but I scrolled down very far in this thread and didn't see this point made anywhere:
The article focuses almost entirely on technical questions, but the technical considerations are secondary; the reason so many organizations prefer cloud services, VMs, and containers is to manage the challenges of scaling organizationally, not technically.
Giving every team the tools necessary to spin up small or experimental services greases the skids of a large or quickly growing organization. It's possible to set this up on rented servers, but it's an up front cost in time.
The article makes perfect sense for a mature public facing service with a lot of predictable usage, but the sweet spot for cloud services is sprawling organizations with lots of different teams doing lots of different mostly-internally facing things.
I agree with almost everything you said; except that the article offers extremely valuable advice for small startups going the cloud / rented VM route: Yearly payments, or approaching a salesperson, can lead to much lower costs.
(I should point out that yesterday, in Azure, I added a VM in a matter of seconds and it took all of 15 minutes to boot up and start running our code. My employer is far too small to have dedicated ops; the cost of cloud VMs is much cheaper than hiring another ops / devops / whatever.)
Yep. To be clear, I thought it was a great article with lots of great advice, just too focused on the technical aspects of cloud benefits, whereas I think the real value is organizational.
The article focuses almost entirely on technical questions, but the technical considerations are secondary; the reason so many organizations prefer cloud services, VMs, and containers is to manage the challenges of scaling organizationally, not technically.
Giving every team the tools necessary to spin up small or experimental services greases the skids of a large or quickly growing organization. It's possible to set this up on rented servers, but it's an up front cost in time.
The article makes perfect sense for a mature public facing service with a lot of predictable usage, but the sweet spot for cloud services is sprawling organizations with lots of different teams doing lots of different mostly-internally facing things.