The Evolution of IT
By John Parkinson
If you look at the history of evolution in natural species (assuming you believe it happens, that is), you see two sets of forces at work: both long- and short-term incremental improvements through somewhat random mixing of characteristics (survival of the fittest--or more probably, the best adaptive) interspersed with rare but significant "wipe the slate clean" extinction events (I think there are four or five in the paleontological record) that change the options available in the incremental landscape in fundamental ways.
Even so, some species persist across major evolutionary discontinuities. Sharks are virtually unchanged for 300 million years, cockroaches for at least 100 million. Paleo-biology calls these "keystone" species because they anchor large parts of the ecosystem across discontinuities and form a nucleus for rebuilding.
If you are willing to draw an analogy with the IT ecosystem (after all, we evolve a lot faster than biology and burn a lot more unsuccessful "species" along the way) you can see that (a) we are probably due for an extinction-event-like discontinuity sometime soon and (b) we have some candidates for both shark and cockroach as keystone species.
The looming discontinuity is the energy cycle and the limits to how much energy we can invest in our IT infrastructure. We will see this manifested in the shift to cloud computing, but that's going to make the problem worse, not better, because of the concentration of energy required for large-scale data centers. A lot of today's top technology species, all evolved amidst abundant energy, will be wiped away by this.
Mainframe = shark (not dinosaur, as it is so often characterized). Not only is today's mainframe a linear descendant of the original design from 40 years ago, it's successfully absorbed the innovations (DNA) of many other "species" along the way, growing stronger and more dominant in the process. It's also remarkable energy efficient.
As for the cockroach? The microprocessor fits here. Microprocessors are everywhere, have mutated into many kinds, seemingly have minds of their own and are very hard to get rid of. Of course, they are also a critical part of the technology ecosystem (even mainframes use them) and in the right circumstances, they create a great deal of value.
John Parkinson is CIO of TransUnion. To read his columns for CIO Insight, click here.