Managing imperfection

MOST OF THE time, this column focuses on how the CTO excels in the corporate environment due to his understanding of the complex intersection between business and technology. The CTO's value proposition is in his deep understanding of technology issues and how they apply to driving business goals. While what you are about to read is a little less positive than usual, those earlier statements remain true.

Free IT resource Hear how top CIOs turn change into a competitive advantage. Sponsored by HP Free IT resource Try Sun servers, workstations and storage products free for 60-days. Sponsored by Sun Microsystems

However, what I have left out in many of my earlier columns is a simple fact that many CTOs know all too well: Technology can often be a messy, frustrating, hair-pulling conundrum where solutions to unforeseen problems are elusive and the time to deal with surprises is difficult to come by. Managing technology problems in the age of the Internet can be quite complicated. The means to fix a particular problem can be just out of your reach and control at any given moment. If you're a control freak, the CTO position is probably not a good fit.

All IT executives are trying to identify an operating philosophy to make operations less prone to chaos and more efficient. Recently, a reader asked what I thought about a centralized IT philosophy, where all IT issues are managed within a centrally managed IT department, versus a decentralized philosophy, where individual departments are responsible for managing IT issues within their budgets. My answer is simple: The Internet makes the truly centralized IT department an impossibility and decentralization is simply a foregone conclusion.

This decentralization sometimes makes it difficult for CTOs to effectively manage all the problems that come their way. Since your company is connected to the Internet, your extended IT department now includes every IT staffer at your hosting provider -- I'm sure many CTOs reading this know their provider's network engineers by name. Go ahead and count the folks who work at VeriSign because they are probably managing your root domain name records, which determine to some degree whether or not your Web site works and your e-mail gets delivered. If you have ever had a tricky DNS problem, then you know exactly what I'm talking about.

The reporting and resolution of technology issues has also changed as the Internet has become core to most IT operations. If you are running a publicly accessible Web site, you probably get several problem reports on a daily basis from your customers.

Operating under the old "the customer is always right" adage means you could be in for some hours of head-pounding frustration. Performance over the Internet can be somewhat subjective, and the poor performance reported by one of your users on their end can mean 1. the user on the other end has a defective cable modem, 2. a backhoe cut a cable somewhere in Kansas, 3. your staff misconfigured your Apache servers, 4. UUNet has a poor peering arrangement with Cable & Wireless in a particular location, or 5. the staff running one of the networks along the way just installed a buggy new firewall. CTOs, it's your duty to work with your staff to figure out which of these five things caused the problem.

Or maybe it's something else entirely. If you find that the problem is not something over which you have hands-on technical control, you have to apply political pressure in the right places, and finding the person or entity that needs pressuring is often difficult -- you might not even have a business relationship to give you the leverage you need.

The Internet is an amazing piece of infrastructure for delivering IT solutions on a global scale. Too bad you can't always control the darn thing.