I am currently trying to establish a cross charging model in my company where VM consumers (e.g. projects) are charged for the creations and usage of VMs. Creation is charged as a one-off Capex charge, usage as an ongoing Opex charge. We are not yet switching to a utilization based model, so at the moment the charges are based on requested VM size.
Would anyone be able to share their cost model, or give an high level overview how theirs works, I would be interested how other companies approaching this problem. My main issues are:
* How does VM lifetime get charged? Currently I assume one year lifetime, and a hardware/software depreciation of three years.
* How does VM size influence the charging? With larger VMs the over subsciption ratio (e.g. for vCPUs vs. cores) reduces, therefore large VMs should cost substantially more than small VMs.