Abstract , money bleeding attack
Well ,well
This might be not a news for you , i was looking around about this topic but honestly i was not able to gather informations, so i thinks that or it's new (nothing is new on internet) or it's quite obfuscated or not so much considered.
We are habit to heard some kind of attacks ... ddos , deface , data exfiltration etc etc
but some new formula are having now the opportunity to be relevant for the company budget
Please note that this is NOT a pro and cons about cloud or onprem ...
What i'm writing it's just a consideration about some challenge that are obfuscated in a onprem solution , that are big deal in the cloud
Let' start with a small example in a onprem solution, you have a data center where we suppose you are hosting website ... now the website can be an e-commerce or something else but we should agree that is exposing services.
This is not always true ... but a data center has the following fees
- monthly for racks
- monthly for power
- monthly for connection
- monthly for a lot of other stuff ... but let's simplify
Still talking about a data center , the fees often related to predefined resources
(yes... you can have a solution like hpe green lake or something similar)
Summarising the topic above in some cases you have a fixed fee for fixed services
It's also true that a data center is often oversized ... maybe is not double for availability zones ... maybe is not 50% more , however is always more than the traffic you are managing , otherwise you cannot survive during the spike, seasonality , hardware fault

Again this is an oversimplification about the resources usage ... it's rare that the real capacity is covered by machine with 80% of load and it's alo rare have spare racks ... better to use it and decrease the general usage overall.
Now ... what you see is what you are paying
Consider now the mindset about the cloud , the same infrastructure can be considered more cost and needs oriented
- Everything can be automatised with autoscaling
- The usage can be different for seasonality ... also by the daily trend
- Pricing is more close to the number of user served
Last point it's really important
On a datacenter this kind of think is not always present... you have fixed cost ... so the goal is more close to be able to handle more customers with the same resources
In the Cloud it's still part of the discussion but will share some part of the topic with the goal to be able to manage customers with less resources