Temporarily renting the capacity to handle spikes in load is called

Cloud deployment strategies differ in how they handle demand that exceeds an organization's private, on-premises capacity. One well-defined pattern lets an application temporarily extend into a public cloud provider's infrastructure only for the duration of a demand spike, then scale back down once normal load resumes — avoiding the cost of permanently over-provisioning private hardware for rare peaks.

The stem describes exactly this behaviour: capacity is rented on a temporary, on-demand basis solely to absorb spikes in load. This is called cloud bursting — the private environment “bursts” into the public cloud when load exceeds its own capacity and retracts once the spike passes.

• “Cloud sharing” describes multiple tenants concurrently using shared infrastructure, not a temporary capacity-rental response to a spike.

• “Cloud computing” is the broad umbrella term for delivering computing resources over the internet in general, not the specific bursting-on-spike strategy.

• “Cloud loading” is not an established term in this context and does not describe renting temporary capacity.

Because only cloud bursting names this temporary, spike-triggered capacity rental, it is the correct term.