resiliency mostly pays off when there's a crisis. You can reduce latency in systems by sourcing things closer to use, but having a fallback farther away.
It's a tradeoff though. If I have two suppliers, the odds that I have 'a supplier issue' are doubled. But the odds that I have no supply at all is lower.
I learned much later that the only way to get IBM and some other big vendors to do anything at all for you is to move your checkbook. That is, if the checkbook might be coming out of your pocket, they're interested. If you threaten to put it away, they are very interested. The rest of the time? Fuck you, we have other checkbooks to worry about.
Edit: Now I got the cluster thing. I meant cluster in the sense of risk, meaning you have a single point of failure and high degree of risk concentration. But your point is really interesting, I never thought about these aspects of cloud computing.
And if you have to spend most of your income on stuff, you too proportionately pay more sales tax than someone with excess to invest.
Well, this is the problem, right. Resiliency only pays off when there's a crisis, otherwise it's a net cost. So if you're outcompeted by the other non-resilient companies before that crisis happens, it doesn't help.
I feel like societal resiliency needs to be rewarded (in many contexts, utilities are obvious but also food supply, etc) but I don't know what a good way of doing this would be, except making it government run and making resiliency a primary objective. I don't see how a for-profit system that rewards short-term gains could ever be made to value resiliency.