Google sure is nowhere near as capable or even willing as Microsoft or Amazon; heck even IBM through Red Hat could prove a much better steward for its enterprise customers.
But I think the morality is to never put all your eggs in the same basket, because any of them can fail for external reasons — e.g. horror stories about people locked out of any service because payments failed due to some stupid action by their bank; months to fix issues... don't be that guy/business with a single point of failure.
So, contingency plans, and a healthy distribution of your flows.
Have several banks/cards, several clouds, front load balance with a cross-platform HA solution, then for everything behind peg providers against each other; optimize for cost or convenience or whatever.
This is the resilient way; and possibly even antifragile as you develop new synergies in this meta-space. It also sends a message that interop is value to you, to the market.
When you operate at that level, there may be room for GC in your system — typically where you can afford it to fail / disappear but it's better/cheaper as long as it lives. You're already ready for the day it dies, because you make that risk assessment a parameter in your arch.
Indeed! Love the spirit, work your way bottom-up as you encounter problems, remain practical yet keep that 10,000 ft view. All compounded by splendid execution, to the point that you are elevated 'upstream' so to speak by the master[1].
(There might also incentive for Google to move away from anything even remotely touched by Oracle; case in point being Java, despite the open sdk for that).
But I think the morality is to never put all your eggs in the same basket, because any of them can fail for external reasons — e.g. horror stories about people locked out of any service because payments failed due to some stupid action by their bank; months to fix issues... don't be that guy/business with a single point of failure.
So, contingency plans, and a healthy distribution of your flows.
Have several banks/cards, several clouds, front load balance with a cross-platform HA solution, then for everything behind peg providers against each other; optimize for cost or convenience or whatever.
This is the resilient way; and possibly even antifragile as you develop new synergies in this meta-space. It also sends a message that interop is value to you, to the market.
When you operate at that level, there may be room for GC in your system — typically where you can afford it to fail / disappear but it's better/cheaper as long as it lives. You're already ready for the day it dies, because you make that risk assessment a parameter in your arch.