We already have Cloud computing systems available, however when will we be able to shopping list order our computing power or infrastructure – consider two pricing models, shared and dedicated:
With shared pricing, you’ll be able to pay a fee for Kb/RAM/s which is a flat fee for every Kilobyte of RAM used per second, a fee per CPU clock cycle and a fee per Kb of space used. Alternatively you could purchase 2 hours access to 32Gb RAM and 3.2GHz for a flat fee.
Cloud computing is being driven forward by service providers such as Amazon, Google, and Yahoo! however I wonder if there will ever be a need or solution for truly instant on demand infrastructure or resources. GoGrid and Amazon EC2 seem close, however they’re not as straight forward or as powerful as you’d expect, and I’d personally be worried about runaway costs.
Pricing of both services seems good – I’d expect prices to be slightly cheaper than their hosted server equivalent. I can only think of a few uses for this kind of technology, primarily short time distributed computing projects, brute forcing, rainbow table generation, large data processing, etc.
I’d really like to see this kind of technology used in place of regular hosting environments though. Possibly in a scenario where your hardware and resource requirements were persistant but evolving on a regular basis. I’m struggling to think of such a scenario though..