"Why is it stupid, dangerous and short sighted? People say those things about cloud vendor locking like it's a obvious truth. I really annoyed by this arrogance among my systems minded peer group."
It's not arrogance. It's decades of experience in this industry, and simple business reality. If you let yourself get "locked in the trunk", you've made yourself vulnerable to whoever owns the trunk. If an entire industry gets locked in the same trunk, the only sensible decision for the owner of that trunk is to raise the rent. What are the people in it going to do?
Most recently the industry was here with Microsoft Windows. Our whole industry was so beholden to Microsoft that we basically danced to their tune. Computer manufacturers didn't sell "computers", they sold "Windows machines". Microsoft could move in and take over entire sub-industries with impunity, like the spreadsheet industry and office suites in general, networking software industry, the browser industry, pretty much anything that was software and they decided they wanted.
Yes, there are entities that are too small to be worried about this. If you're a five-person startup, you've got bigger problems than whether or not your cloud vendor is going to become a monopoly in five years and jack the prices through the roof for your services. But larger entities need to be more careful.
There's other cases too; you can end up over-exposed to Oracle databases, for instance. Some small-medium companies end up designing custom hardware for some purpose and end up overexposed to the only people willing to make that hardware at a price you can profit from it. It's not a good idea to be stuck to one entity like that, and even if you judge the benefits to outweigh the costs, that doesn't mean the costs don't exist and you can just forget about them. It's something you need to constantly reevaluate.
I agree that those are all valid problem. My point is more that it's not such hard lock as people assume who never built a "cloud native" infrastructure. I wish we would discuss more specifics.
So what are we talking about? Snapchat running on app engine is probably one of the more extrem examples. Let say they aren't happy with that anymore. So far they deployed easily by using the google apps api and since want to keep a similar workflow. They could setup kubernetes for that. Same for managed data stores. Maybe there is no 1:1 open source equivalent, but most likely there is one similar in design which won't require much change in your code.
Of course all this takes time and live migrating your site isn't trivial.
But so does building on bare metal from the start. Sure, you don't all the fancy self-service features right away, but you want them eventually.
So even if you realize that moving to the cloud, or a specific cloud was a huge mistake, it's not like being locked in a trunk.
I know its cool to blame every Microsoft success on the windows monopoly, but Excel won on its merits. Compare Excel 3 with Lotus 1-2-3 of the same era. One had a GUI the other was stuck in 80x40 character land.
Excel 4 is about when the product really started to resemble its modern form. But even V1 on the Macintosh was a significant improvement over character based spreadsheets.
Simply blaming it on Windows ignores the very real innovations made by early Excel versions. Lotus lost simply because they were late to the GUI, and even then only did it half-heatedly.
It's not arrogance. It's decades of experience in this industry, and simple business reality. If you let yourself get "locked in the trunk", you've made yourself vulnerable to whoever owns the trunk. If an entire industry gets locked in the same trunk, the only sensible decision for the owner of that trunk is to raise the rent. What are the people in it going to do?
Most recently the industry was here with Microsoft Windows. Our whole industry was so beholden to Microsoft that we basically danced to their tune. Computer manufacturers didn't sell "computers", they sold "Windows machines". Microsoft could move in and take over entire sub-industries with impunity, like the spreadsheet industry and office suites in general, networking software industry, the browser industry, pretty much anything that was software and they decided they wanted.
Yes, there are entities that are too small to be worried about this. If you're a five-person startup, you've got bigger problems than whether or not your cloud vendor is going to become a monopoly in five years and jack the prices through the roof for your services. But larger entities need to be more careful.
There's other cases too; you can end up over-exposed to Oracle databases, for instance. Some small-medium companies end up designing custom hardware for some purpose and end up overexposed to the only people willing to make that hardware at a price you can profit from it. It's not a good idea to be stuck to one entity like that, and even if you judge the benefits to outweigh the costs, that doesn't mean the costs don't exist and you can just forget about them. It's something you need to constantly reevaluate.