Readers Still Don't Like Carr's Outlook on IT

By CIOinsight  |  Posted 05-27-2005

Readers respond to Nicholas Carr's May 23 column, "The Next IT Revolution":

Consolidation's a Good Idea, But Not a New One

I couldn't agree more with Nicholas Carr's assessment of today's IT world. He suggests a consolidation of corporate computing power … what a novel idea!!

It was the devolvement from mainframes and dedicated (intelligent) terminals to (dumb) PCs and server farms that caused all this inefficiency in the first place.

So, everything old becomes new again.

Bob M.
Harrisburg, Pa.
iamnomad@att.net

Consolidation Comes in Small Steps, Not Just Major Movements

In Nicholas Carr's recent article "The Next IT Revolution," the comparison of IT infrastructure to electricity may be a convenient analogy at first blush, but as soon as you delve deeper, it is not so clear.

Electricity is the same for every company—information is not. Granted, very large corporations can benefit significantly by consolidating their data centers.

For the other not-so-giant companies, the real message (not articulated so well in the article) is this: Consolidation is really happening at the micro level (e.g. blade servers, Virtual OS utilities, Web services, etc.) as opposed to the grandiose data center level.

As far as his comments on the "overinvestment in IT labor," the point is made that businesses can dramatically reduce IT infrastructure and labor costs by moving all corporate information assets and processing outside the company.

This seems a bit of an overreach for me (if you think your IS department's help-desk support is bad now, wait until all your applications and core data reside in Bangalore or Manila!).

A related unanswered question is, "How do you support the infrastructure necessary to access your information and applications if you have neatly removed all or most of your IT infrastructure and support by moving these components to the outside?" Wouldn't you still need switches, routers, network technicians, etc.? That means firewalls, antivirus, and help-desk support … it kind of goes full circle.

I applaud Mr. Carr's forward thinking, but personally I think the next IT revolution will arrive in much more powerful individual computers, with much improved security, networked in a geographically unconstrained peer-to-peer mesh. That kind of revolutionary change would be a more realistic way of reducing IT infrastructure inefficiencies.

Perry Cherpes
plcsoccer@aol.com
production management executive
for a global application-development services company

Click here to read an article by Tom Pisello of Alinean highlighting the strengths of IT that Carr forgot. and Click here to read the second round of reader responses.

Underutilization? Or Headroom?

While I do believe that Mr. Carr's thoughts about consolidating, standardizing and pruninghave merit, he fails to mention the security aspects of all the consolidating.

One of the main reasons a lot of systems are underutilized is that they are isolated from each other for security reasons and have special purposes.

Another reason is that folks buy systems that are usually much bigger than initially needed so they can grow into them instead of replacing them every year or so. So, I think the article is somewhat misleading.

Thanks,
Rod Lindley
rjtyree@sbcglobal.net
Programmer II for
a West Coast financial institution

Obsolescence Is the Real Problem

The opinion piece you're carrying from Nicholas Carr on the refrain that "IT Doesn't Matter" has the great virtue of brevity. I will try to respond in the same spirit.

Lots of money is spent on computing infrastructure, or in Carr's words, "keeping the lights on." Fair enough. In fact, lots of money is spent on literally keeping the lights on.

We're willing to build a lot of lighting infrastructure even though we rarely expect to use it more than a fraction of the day, or in some cases not for days at a time.

We don't expect to build things out of this kind of light, nor does it offer a "strategic advantage" that would differentiate us from the competition. But it does supply a necessary condition for certain kinds of creativity that might indeed be distinctive.

Of course, an information infrastructure will be more complex than a lighting infrastructure, and rightly so, given the value and uniqueness of what is being transported. That observation doesn't provide us with a license to waste resources on excessive infrastructure, any more than we would install more lighting than conditions require. But what exactly are those conditions?

In the case of lighting, and indeed of most engineering works, we already know the answer in considerable detail. But can we make the same claim about information infrastructure?

Clearly, if all you do with a computer is read e-mail, then we can budget very modestly and precisely for your needs, that is, as long as the content is forever restricted to plain text or some other fixed technology. In that case, no computing infrastructure will need to be replaced until it physically wears out.

In reality, a lot of infrastructure is working perfectly well at the time it is replaced, and if Nicholas Carr is concerned about the extent of that waste, then he is by no means alone.

The problem seems to be purely a result of obsolescence, and there is no end in sight. You could try to anticipate future developments and engineer your information infrastructure accordingly, and we all try to do that, but we are all constrained by the materials at hand, if not by our imaginations.

When all you've got is coaxial Ethernet, it's pretty hard to anticipate an infrastructure that will scale to 10Gb/s, not concretely at any rate.

Everybody wants to talk about grids and utility computing, as another example, but those who promote them the most seem to have the least experience in what their deployment actually entails.

Grid infrastructure fundamentally depends on network infrastructure, which is also subject to obsolescence. Grid also depends on an identity infrastructure which is still in the early stages of being worked out conceptually, to say nothing of security.

So, what was the point again? Originally, I think it was that IT "didn't matter" because it "wasn't strategic." That claim now seems to have been diluted into a form related to resource utilization, which is progress of a sort, but onto ground which has been extensively tilled already.

It might be more interesting to hear what Carr has to say about the nature of obsolescence, and what, in economic terms, we can do to address it.

Best wishes,

Dan Razzell
Starfish Systems
www.starfishsystems.ca