Thursday, May 26, 2005

Consolidate. Standardize. Prune.

How is it possible that someone with whom I almost totally disagree has, nonetheless, arrived at the same exact conclusion as me?

The person I'm referring to is Nicholas G. Carr,

author of 'Does IT Matter?'
Best known for his article 'IT Doesn't Matter' which appeared in the Harvard Business Review two years ago, Carr's most recent article, 'The Next IT Revolution', published in the May 23, 2005 edition of CIO Insight quotes:
  • a Hewlett-Packard study that found most corporate servers use just 10 to 35 percent of their processing power
  • an IBM estimate that average desktop-capacity-utilization rates are just 5 percent
  • a Gartner estimate that 50 to 60 percent of a typical company's network storage capacity is wasted
Carr compares utility computing with electrical power generation. He states:
Between 1905 and 1930, when it became possible to centralize power generation in big utilities and distribute current over a network, a revolution took place. Many manufacturers resisted, but lower costs, freed-up capital, enhanced flexibility and fewer headaches made even the largest companies switch. Similarly, advances in networking, hardware virtualization, grid computing and Web services are making IT utilities possible, in which IT will go from being a complex corporate asset to being a simple variable expense. Utility computing today merely hints at coming changes.
Carr's analysis reminds me of the following totally erroneous quotes from the past:
  • "I think there is a world market for maybe five computers."
                   -- Thomas Watson, chairman of IBM, 1943
  • "There is no reason anyone would want a computer in their home."
                   -- Ken Olson, chairman of DEC, 1977
  • "640K ought to be enough memory for anybody."
                   -- Bill Gates, chairman of Microsoft, 1981
In my opinion, Carr just doesn't get it. He can't quite grasp the true power of the exponentially exploding drivers that fuel IT's growth:
  1. transistors
  2. storewidth
  3. bandwidth
I suggest computing more closely resembles telephones and televisions than electrical power generation. How many phones are in your house? How many TV's? My guess is there's unbelievable under-utilization, but who cares? Locality is what's important.

With so much cheap storage, virtually free bandwidth, and limitless processing power, IT innovation will stem from moving the mountain to Mohammed rather than moving Mohammed to the mountain. Imagine the asynchronous behind-the-scenes movement of data where copies of most frequently accessed databases literally reside on local computers. The analogy here is like virtual memory where most commonly referenced pages reside in physical memory for rapid access. Think how Lotus Notes revolutionized email by synchronizing desktops and servers.

I wholeheartedly agree with Carr that significant savings can be achieved by
  • consolidating data centers
  • standardizing platforms, products, and applications
  • pruning redundancies, reducing inefficiencies, and eliminating unnecessary replication
On the other hand, I categorically disagree that IT doesn't matter. Unlike electricity, which enables the operation of equipment according to immutable physical laws, IT powers business through a highly complex and ever-changing dynamic involving technologies, processes, and people. Often, business problems get defined in terms of available solutions. Moreover, the technological solution to a business problem frequently changes the problem. There is a complex and intricate relationship that connects emerging technologies to paradigm shifts. With technologies continuing to advance at an exponential pace, rest assured that IT DOES MATTER!


Blogger ITscout said...

Vinnie Mirchandani, founder of Deal Architect, offers a couple of very interesting postings on his blog about Nicholas Carr: Nicholas Carr: The Francis Fukuyama of IT? and The end of corporate computing? The beginning of chaos...

2:40 PM  

Post a Comment

<< Home