The increasing cost of information technology

I was reading a blog post recently by a CIO from a large retail bank, bemoaning the ever increasing cost of information technology.  Why, he asked, is the cost of information technology ever increasing, when the cost of hardware, processing power and memory constantly falling?  As Pogo said, we have met the enemy, and he is us.

While Moore's law stipulates that the cost of processing power will continue to fall at a spectacular rate, and the cost of memory will fall perhaps even faster (my MP-3 player has almost as much memory as my laptop from 2 years ago), these laws denote the fact that we can make the infrastructure incredibly inexpensive.  So inexpensive, in fact, that we've ignored the ever increasing cost of capturing and holding data.  I'd be willing to bet that while the cost of hardware and processing power has fallen at a steady, regular basis for years, the increase in the amount of data we capture, manage and hold has grown exponentially.  Even in an IT downturn, such as the one we are in right now, firms that make data storage systems are doing relatively well, simply because we can't keep ourselves from capturing and storing ever more data.

And, once we've captured and stored that data, we've got to protect it.  So, even with inexpensive servers and software, we add another layer of security, which means more people have to administer the applications and more layers of integration and complexity.  And, of course to justify all the data we capture we have to analyze and report it.  Which means more software to slice and dice and report the data, and more people who can manipulate and manage the data.

Ultimately, we have in any IT environments several "honey pots" of data that must be protected and analyzed.  And these do nothing but grow.  Soon we are seeking "off site" locations for the data to be housed and stored.

Perhaps we need a new Moore's law, which would be the rate of increase to capture, store, protect, analyze and manage one gigabyte of data for one year.  Unlike Moore's law, which suggests the price of processing power should fall over time, I suspect the Phillips law (immodestly named) will indicate a significantly rising cost over time, and demonstrate the true cost of information technology.  We need to begin to ask the question - do we really need all of this data, and all of the infrastructure costs associated with capturing, managing, storing, recalling, analyzing, archiving and reporting it?

Comments

[smarks]

(Coming late to this.)

I don't think the cost of storing data increases over time.

You're correct to note that there are costs associated with keeping data online and securely backed up. For example, a 1TB drive costs less than $100 today. However, to keep 1TB of data online 24x7, and backed up securely (including offsite copies, disaster recovery) etc is considerably more than $100. I'm not sure how much it costs. Maybe $1000. Call this the "fully burdened" cost of storing the data. IT departments have always been concerned and aware of these issues.

Now consider, say five years ago. I doubt that the fully burdened cost of storing 1TB was any less then than it is today. I'm certain it was more, and indeed going back a few years it would have been unthinkable for a business to consider keeping 1TB of data online.

I think the reason for IT's ever-increasing cost is the *demand* for keeping and manipulating more data. The cost per GB or per TB has fallen over time, I'm sure, but the demand for storing and processing data has quite possibly outstripped this reduction in cost.

I think it's a fair question as to whether aggregate IT expenditures have really improved productivity. Certainly there have been some improvements; the question is whether the improvements have justified the costs.