Computing services are different from electrical power services in one very important way. The provision of electricity is a pure push operation–the quantity in question comes pouring through a hole in your perimeter, and the service provider has no idea what you are up to. In total contrast to that, using commercial cloud computing is more like sending your rings, bracelets, and brooches out to be repaired–the service provider has your family jewels in hand.
A power utility does not have possession of your means of production. They do not have access to your trade secrets, your intellectual property, or private data related to your customers. The provision of electrical power is essentially a one-way transaction–there’s virtually nothing for the power company to steal, lose, or break.
Commercial cloud service providers have physical, virtual, and administrative control over your data, applications, and level of service. The risk implications of that are significant.
To repeat what I said in my previous blog entry, I actually do believe that the power metaphor is a useful one that provides some important insights into the economic drivers, and network effect implications of the provision of cloud services. The two situations are analogous, but not totally so. The electrical utility model provides virtually no guidance about the confidentiality implications of public clouds, which is unfortunate and sometimes counterproductive.
You only have to trust your power company to a relatively small degree, and as I discussed yesterday, you still might want to maintain your own redundant capability in house. The utility model of cloud computing totally glosses over the trust implications. If you want to use a CSP, you have to trust them–a lot. Utility computing providers have far more control over your destiny than do power utilities.