The right IT investment, commodity and operations mix for the CIO of a future successful business is naturally constantly evolving, but cloud computing will certainly become a key ingredient for most of the resulting IT solution mixtures.

Carr was always right because from a business perspective, the few business differentiators that could help a business win against its competition from within IT have been tougher to find and develop and disappearing more quickly once established. IT is a mature business, so game changing invention and innovation is harder to find. In addition, as IT commoditizes, a game changing differentiator from IT is harder to find and deliver. Even once you achieve a differentiator with IT, technology is evolving so rapidly that the differentiator can be wiped out quickly. The manifestation of IT services and outsourcing also rapidly destroys the IT differentiators by their access to your differentiator intellectual capital and subsequent market dissemination by the IT services firms as they sell it to make more money.

Companies like IBM saw this a while back and got into the services business to use their client's money to build differentiators in their solutions which they in turn would sell later to their competitors as part of "engagement intellectual capital". That's why in most IBM IGS/EDS/Deloitte, etc. IT services contracts the client doesn't get to own the IC developed at their own expense.

I laugh at IBM's intellectual "jams" because they are actually getting free ideas and other IC from clients and even competitors which they can in turn use and capitalize on to charge those who provided the ideas in the first place. Solution providers are desperately seeking for those differentiators as well but they innovatively use the jam to get unwitting clients and others to provide differentiators for them and even validate the IC for free. It's the biggest con in IT today.

Cloud computing is a further manifestation of the commoditization of IT which will do very much the same as electricity did to that market.

Your analogy of the Ford Model T and A conversion, however, holds the key to successful and valuable IT.

The successful CIO/CTO of the future will be the one who can use internal private IT assets to identify, create and capture an IT differentiator, use the laws of intellectual protection (patents, etc.) to make it last as long as possible and be forward thinking as to how to insert such a differentiator into current commodity IT systems like cloud SAAS with a minimum of financial and business impact(i.e, keeping the Ford Model T to A plant conversion costs to a minimum). The successful CIO/CTO won't also have talent retention issues because his/her IT people will always be in the "fun" part of IT where the work is learning about the new IT technologies and figuring out how to make them into business differentiators, not doing the boring part of IT like running a datacenter, performing help desk work or running a payroll.

I think the analogy between Cloud Computing and Electricity is very powerful, up to a point. The factors that still remain for Cloud Computing to address are:

+ Compliance
+ Security
+ Liability

How will Cloud Computing address the risks associated with events like the FBI raid on Service Providers' Dallas data centers?
How will Cloud Computing address the liability issues of hackers gaining access to personal data stored in their data center?

The vision of a public computing grid is very powerful and attractive and has far reaching implications on business. I believe we will see a multi-phased implementation of Cloud Computing:

1. I see the emergence of select SaaS-like external clouds. However, I believe the emphasis will shift to internal clouds due to the issues mentioned above.
2. I then think we will see external service providers engaged to manage our internal Clouds.
3. As the issues above are addressed, we will begin to see the shift to external clouds, but slowly because the value proposition will not be as high when compared to internal cloud/managed services.
4. We will see an attempt to merge internal clouds into a larger external cloud (joining hands).

The interesting challenge is to decipher how long each of these phases will take. Obviously, the phases are not crisp and well-defined. Also, I doubt if the term "cloud computing" will last through these phases. We will most likely call it something else (maybe "Time-Sharing").

Cloud computing will be AN ingredient, but I doubt that it will be "key". And security et al are factors, but more importantly to the foundation of a competitive marketplace are the customers themselves, what they buy from "Company A", how their purchase contracts are structured with "Company A", etc.

And while software and networking solutions are "propriety" (though not unique - programmers take knowledge with them - you cannot "non-compete" skill set fundamentals and industry standards), the data behind a company's customer base is VERY confidential - to the point that were it government, it would be classified Top Secret. Sorry, but NO company can afford the loss incurred on so many fronts (competitively, monetarily - including privacy suits, and more) that would ensue from "declassifying" that data and putting it in a Cloud for everyone - nor would they be stupid enough, lest they desire company death.

Cloud computing for public domain info (i.e. mapping from satellites, that our tax dollars support a large chunk of anyway, zip codes, area code, shipping, etc) - sure, that has a future in the Clouds, but not company-customer specific data. And therein, at least a while longer, lay the need for company IT.

There is one factor that none of the proponents of Cloud computing seem to address is the cost of Internet Bandwidth. And this is where the analogy with grid electricity fails. At the moment Bandwidth is assumed to be large, costs less and is available on-demand. When cloud computing picks up the demand for large bandwidth is bound to increase and its Economics 101 that the demand side pressure increases the cost as well. ISPs might insist on data metering - either at vendor side like telephones or at client site similar to electricity – so that they can charge on a per MB basis. Cloud computing would then loose its cost advantage and will have to offer something more. CIOs should keep this in mind while opting in for cloud computing service.