cloud infrastructure can be considered an extremely powerful tool for organizations of any size. Allowing things such as easily scalable storage all the way to easily accessible resources such as applications through the internet. Taking advantage of the ease of use that is cloud-based storage is a must for organizations that are looking for another form of backup and recovery for their current form of data storage. This should not be replacing any other form of backup and recovery, as it is never bad to have multiple means of backing up important data. Another advantage of Cloud Infrastructures, if you opt for a program that allows you to use say the processing power of said program’s machines in any shape or form, this could end up saving extreme amounts of money by dodging an increased labor cost and even increase in specific bills. An example could be to say you require processing power for video rendering, but just don’t have any machines currently that can get the job done, and installing these new machines could prove to be a hindrance or even too costly at the moment. There are quite a few services that will let you pay a specific amount of money (Per GPU & Render Time for example) to do just that. One of the biggest attractions of taking advantage of cloud-based services would have to be the extremely high rate at which you would be saving money. Taking an example from D.G. Chandra and M.D. Borah’s “Cost-benefit analysis of cloud computing in education,” we see they compare the pricing of setting up a full network of computers with proper staffing and hardware costs in comparison to monthly fees from signing up for and maintaining 30 virtual desktops, over a course of three years total. They estimate a total cost of $1,018,841 over a 3 year period for an on-premise setup in comparison to a  total of $1,667.15 for the 30 virtual desktops. This estimates that these cloud services can cost mere fractions of that of on-premise setups when accounting for staffing requirements and other maintenance fees, which also doesn’t even include the possibility of requiring upgrading hardware within these machines later down the line.