We are clearly entering the age of utility computing, in which computing will be delivered as a set of packaged services out of a plug in the wall (or increasingly often wirelessly via WiFi and 4G cellular) to end-users, just as electricity and TV are delivered. Those services will be paid for according to metered use or standard monthly fee, again like electricity and cable TV. Users often will consume those services using computing appliances, such as the iPad. And, very significantly, those end-users won’t care where those computing services come from, any more than they think about where their electricity is generated.

This is, of course, precisely the model of Cloud computing services. Software-as-a-Service (SaaS) providers like Salesforce.com don’t talk to the CIO. They go straight to the CFO and appropriate department heads such as the Sr. VP of Sales. Their basic message is: “We can provide the functionality you want at an order-of-magnitude less cost, charge you only for what you use so you always know exactly what you are paying for, with no CapEx, minimum commitment, no guesswork about future needs, and no distracting techie issues or unintelligible techie talk.

Of course internal IT has advantages as well — in particular the ability to integrate multiple applications tightly and to provide direct control and oversight over sensitive data. But those will only count for so much. The bad news is that IT shops, which have been shrinking for over a decade as services have migrated to domestic and increasingly overseas outsourcers, will inevitably see more erosion.

But if they want to hold onto even the most core applications in the face of this onslaught, they have to maximize internal efficiency. While their costs do not necessarily have to match Cloud service prices, they have to get costs down to the same ballpark while at least matching the competition’s service levels. And virtualization is a key to doing that.

Cloud services drive their costs low through the efficiencies of large volume computing that only the largest corporate IT shops can directly match. They do this by focusing on one or a very small set of related services and combining the compute loads of multiple clients. A SaaS vendor, for instance, runs multiple clients on a single software, with a single staff, often on the largest available hardware platform. This also allows the vendor to hire the best experts in its particular service; and of course because its business is providing that service, its entire focus is on delivering that service in the most efficient manner.

Obviously internal IT cannot operate in that manner. However, virtualization can help mid-sized IT shops achieve similar levels of efficiency and flexibility in hardware and staffing. Virtualization allows IT to combine the compute loads of multiple applications on a single large server, increasing utilization from an average that often can be as low as 15% up to 85% or higher, with the ability to dynamically reassign computing resources as demand changes.

This does not completely even the field since internal IT must maintain multiple applications and vendor relationships and is sometimes forced by internal politics to make decisions that do not support efficiency. And by itself virtualization does not answer all the management advantages that Cloud computing vendors offer. Some of those will be the subject of future blogs. But without maximal virtualization IT isn’t in the ballgame, and the first you know that the game has changed may be when the CFO calls to tell you that the SaaS contract is a done deal.