Simulation
Move the demand slider and switch hosting models to see how fixed on-premises capacity compares with autoscaling on Google Cloud.
Change the inputs and watch how the outcome changes.

Move the demand slider and switch hosting models to see how fixed on-premises capacity compares with autoscaling on Google Cloud.
In an on-premises data center, hardware, power, cooling, maintenance, and support were all bought up front as capital expenditure. At off-peak demand almost every server is idle, but the full cost of the fixed capacity is still yours, so most of the investment is wasted.
Starting scenarioEvery input at its default
With the hosting model set to on-premises and a steady baseline, most servers carry regular load. The remaining headroom bought for peaks still sits idle and unrecoverable, so part of the capital expenditure continues to be wasted.
What changed
With the hosting model set to Google Cloud and off-peak demand, autoscaling scales compute down to the policy minimum, so there is no idle capacity to pay for. Physical infrastructure is offloaded to Google, and consumption-based pricing bills only what is actually used.
What changed
With the hosting model set to Google Cloud and a steady baseline, the autoscaling group runs a right-sized set of instances matched to actual demand. You are billed only for the resources actively consumed, with no idle headroom cost.
What changed
With the hosting model set to Google Cloud and a traffic surge, autoscaling automatically adds instances within the policy maximum to absorb the spike, and you are billed only for the burst duration. When traffic falls again, Cloud Run or a managed instance group scales instances back down to reduce costs.
What changed
With the hosting model set to on-premises and a traffic surge, demand exceeds the fixed hardware capacity. New servers cannot be provisioned instantly, so the workload bottlenecks and performance and user experience degrade.
What changed
Consumption-Based Pricing Replaces Upfront Capital Costs
Consumption-based pricing means an organization pays for the Google Cloud resources it actually consumes rather than paying for fixed capacity in advance. In an on-premises data center, routers, switches, cabling, power, cooling, maintenance, and support require large upfront and ongoing costs. In Google Cloud, many costs become usage-based, so organizations can avoid paying for idle capacity and adjust spending when demand changes.
Domain 1 of 9
27 practice questions
Practice this DomainLast updated