Cloud pricing: Are per-second billing models unequivocally needed?

An instance of this is with a per-second pricing indication AWS and Google, for example, have adopted of late.

Cloud tended to start off with per-hour pricing, that has afterwards been lowered to per-minute charging as some-more granular workloads have come by – yet does per-second unequivocally offer anything of value to a customer?

A fake (as will be shown) comparison can be done with a telecommunications industry. Voice calls used to be billed on a basement of a tie price total with a per-minute assign for a call.

With so many calls durability reduction than a minute, though, users found themselves profitable a lot for such brief conversations – and didn’t see a value in a model. It usually took one telecommunications association (Orange) to opt for a per-second charging indication for users to see a value and start switching, and – flattering shortly – a telco rivals altered their pricing skeleton too.

Is a cloud marketplace anything like a telecommunications one? In this instance, no.

Without forcing a block use box into a turn cloud hole, I destroy to see only how per-second pricing can unequivocally supplement anything to a cloud model, and can't consider of a effort that is supportive to per-second loadings: sure, there are copiousness (such as those in financial trading) where response times have to be in a sub-second area – yet this has zero to do with per-second billing for executive estimate unit, storage or network usage.

It simply appears to be a delight of selling hype over reality, with some selling bod carrying a “flash of brilliance” in entrance adult with an offer that nobody else had come adult with.