The probability of each node/machine/server failing is very low. But in the setting of a large data center, everything scales, and so do the chances of some node failing at any given time. That's why we make the tradeoff between redundancy and recoverability.
The probability of each node/machine/server failing is very low. But in the setting of a large data center, everything scales, and so do the chances of some node failing at any given time. That's why we make the tradeoff between redundancy and recoverability.