i worked for a hybrid hosting and cloud provider that was partnered with Electronic Arts for the SimCity reboot.
well half way through they decided our cloud wasn’t worth it, and moved providers. but no one bothered to tell all the outsourced foreign developers that they were on a new provider architecture.
all the shit storm fail launch of SimCity was because of extremely shitty code that was meant to work on one cloud and didn’t really work on another. but they assumed hurr hurr all server same.
so you guys got that shit launch and i knew exactly why and couldn’t say a damn thing for YEARS
I wonder if that's related to "the wrong cloud". Imagine if someone wrote some super slick code that worked really really well in the original cloud, and just couldn't figure out how to make it work in the new cloud, so everything is just an awful workaround.
Unless you're really deep into a particular provider's unique-esque products (Lambda, Azure AD, Fargate, etc), this is exactly why things like Terraform exist.
Oh for sure, but the games industry is one of the few that still does some weird stuff because a lot of the software is only expected to last 5 years or so at most, and needs to get every drop of performance.
I could definitely see some hyper optimized cloud API looking really great and then not having an equivalent in another ecosystem (or at least not one that could be quickly swapped out just before release).