As industry digitally transforms, there is an exponential increase in the demand for IT infrastructure, and coupled with that an exponential increase in the overall IT costs. In cloud computing platforms, consumers only pay for the compute resources they use and only when they use it. If a large enterprise IT operation can power off compute resources when not in use, this can save millions of dollars or more. The challenge here is developing the ability accurately predict, in a repeatable fashion, when a compute resource is not being used, turning it off and Eliminating the need to pay for it.