One reason cloud computing can be cheaper than conventional IT infrastructures is because it spreads the cost of operations across several customers.
According to a recent research paper from Microsoft, cloud computing can save money three ways: through supply side savings: operating in large datacentres with a lower cost per server; demand-side aggregation, smoothing demand peaks between customers, and multi-tenancy efficiency, or sharing the cost of the IT resources.
Microsoft argues further that the cost of cloud computing falls as providers build larger systems, because the cost of each unit computing power measured in MIPS falls as organisations buy more. So by pooling the demand from a number of businesses, cloud computing drives down the costs to a lower level than the companies could achieve on their own.
But the flexibility offered by cloud computing does not automatically mean that companies save money over the lifetime of an application or service. Some will, and some will not.
Taking one albeit simplistic example an Amazon S3 storage bucket costs $0.093 per GB per month. That makes the monthly cost of 1TB of storage roughly twice the cost of a basic desktop hard drive, or close to the cost as an "enterprise" drive. However, those raw costs do not include the host server, software, management, power and cooling, or bandwidth, all of which Amazon supplies. This illustrates the importance for IT buyers of making like-for-like comparisons.
However, even among those who do not save money, the ability to preserve capital budgets, or to have more flexible IT, might outweigh any additional cost.
This is the view of James Staten, a principal analyst at Forrester Research. He points out that companies often see the greatest benefits from the cloud when it is used to handle peak workloads, or for specific projects. Long-term rental of cloud services will not suit all businesses, especially enterprises that have the capabilities and scale to build and run large systems in house.
