I'm about 1/2 done with Big Switch, and its good.
Carr's thesis is that information services (hardware and software) are now and increasingly operating on the scale of the electrical power utilities. Just as companies don't generally create their own power, they now no longer need their own IT depts. As network access speeds and reliability approach that available on one's own computer, the network itself becomes one big machine. He's look at and past things like Amazon's EC2 "elastic computing cloud" that allows companies to use Amazon's computers (H&S) as if they were their own. Another example is 3Tera's AppLogic, a cloud computing platform.The customer pays for the computing power consumed when they consume it--just like we pay for electricity. Wow!
Carr briefly highlight large-scale consequences of the electrical gridon society and suggests that similarly large-scale effects will follow from the utilitization of computing. I think he's right. Consider the consequences of large-scale, utility-style cloud computing on digital preservation. If say higher education institutions outsource their computing utility-style to global third-party providers, then preservation of the digital content (an oxymoron; what we mean is curation of digital content over time via migrations) also moves to the third party. There the scale is much larger, part of the ongoing access to content, and costs to individual institutions is amortized across the aggregate of all the institutions using the third party computing utility. In short, indiviual institutions (here colleges and universities) need not themselves directly work to preserve their digital content. They have out-sourced it to their computing utility. It (digital curation--my prefered phrase for digital preservation) still has to be done but not repeatedly at a local (say we call it--retail) level. There are many trust issues here, but then most of us trust our utilities now for water and power.