Running IT is hard enough when all the CIO has to contend with are changing business demands. But these are far from the only “condition fluxes” IT has to deal with.
Take an example: Imagine some aspects of the business you support depend on a reliable, predictable, and immutable definition of the second.
No, not the second what, as in what comes immediately after the first.
The second. The unit of time representing 1/60th of a minute. The unit of time defined by the frequency of cesium atoms oscillating between their two stable quantum states.
The unit of measure on which, the meter, lumen, gram, ampere, degree (Kelvin), and mole depend.
The unit of time that’s in the process of being redefined by the world’s metrologists, a job title I just know most KJR community members aspire to. And before I mosey on to the next point, a thank you to my now-retired business partner, Steve Nazian, for pointing out the coming temporal redefinition to me.
Okay, maybe I’m overstating its potential impact, given that the change in a second’s length will be, at most, about one part in nine billion. Or maybe not, depending on the future of quantum computing.
Anyway, the question to be answered this week isn’t about the second itself. It’s about the coming challenge of tracking down all of the change’s ripple effects. Physicists and engineers around the world might, that is, have a whole new Y2K-like challenge in front of them.
Which (at last!) almost brings us to the point of this week’s rhetorical random walk: How well-documented is your technical architecture?
One more step and we’re there. The point isn’t about your architecture. It’s about you technical architecture management practices, not about how complete and up to date the documentation of your technical infrastructure, platforms, information repositories, applications, interfaces and integration, is not to mention the many-to-many mappings connecting your applications and business capabilities.
Here’s the point
Technical architecture management includes the concept of what’s usually but erroneously called technical debt, to correct which error I’ve suggested the term “chronodebt.”
I’ve defined chronodebt as the accumulated cost of remediating all IT assets that aren’t what engineering standards say they should be. chronodebt is what your IT organization owes the god of time.
Which leads to the question of what we should call the accumulated cost of remediating incomplete and inaccurate documentation of your technical architecture.
It’s chronodebt applied to IT management. It’s a debt that doesn’t come due until it suddenly approaches knee-capping-level urgency.
It matters because there will come a time … there always comes a time … when you need this information Right Now, because you need to track down the ripple effects of a redefined second, a modified business process, the integration requirements of new or updated software, integration of a newly acquired business, the need to provide documentation for a divestiture … really any change imposed on IT by changing circumstances beyond its control or ability to influence.
It’s management chronodebt – the entirely predictable result of striking the wrong balance between those management practices that are most urgent and which are most important.
Bob’s last word: The first law of preventive maintenance is that while you always have a choice between paying now And paying later, paying later almost always costs more.
But in order to implement any program of preventive maintenance, whether it’s managing IT architecture’s chronodebt or a factory’s mechanical and electrical systems, company leadership first has to recognize the existence of what we might call “management chronodebt” – the cost of having failed to invest in the management practices needed to manage technical chronodebt.
New on CIO.com’s CIO Survival Guide: The Edison Ratio: What business and IT leaders get wrong about innovation. Because when it comes to innovation, great ideas matter, But in the end, understanding the need to act on fewer great ideas matters more.
Completely inappropriate for KJR, but who can resist? I hereby nominate the distinguished judge J. Michael Luttig for membership in Bob and Ray’s “Slow Talkers of America.”
There are times when it’s perfectly appropriate to purchase technology, make minimal investments, and run it until it’s dead but too dumb to fall over. Think, for example, about capabilities like your smart ID card system that prints the ID on a smart card that supports access control. You may get everything you need and spend only a little by keeping the old clunker running for a long, long, time. Then someday, you just start over with a new system. If the complexity is low, as it may be here, you don’t miss out on much along the way, and the effort to replace isn’t that big.
Once you get beyond a little complexity and a little business criticality, though, it’s a better play to think evergreen, i.e., updating, enhancing, and moving forward throughout the life of the system. You spend as you go, but you put off that massive rip-and-replace for a much longer time and when you do get there, you may reduce the risk and magnitude of the replacement endeavor. All that, and you get the benefits of new capabilities along the way as you improve the system.
While we IT leaders are not entirely off the hook, we have a big advantage in today’s offerings of IaaS, PaaS, and SaaS, where vendors will do most of the work to keep the system up to date.
My default is to go evergreen to avoid that management chronodebt, but be open to the edge case where the time, attention, and money required to keep something fresh isn’t worth the pain avoidance it provides.