As a consultant and architect for complex IT infrastructures, I repeatedly encounter the “infrastructure bias”. In short, people take infrastructures of all kinds for granted and lose sight of the art in complexity over time. Nobody calls their energy provider in the evening and thanks them for having had reliable electricity all day for years. However, anyone who has looked at the European interconnected grid knows that this would be more than justified on some days of the year. French nuclear power plants are taken off the grid because rivers are too warm, and engineers in Switzerland open the weirs of hydroelectric power plants to keep the grid frequency at 50 hertz – all invisible to the consumer!
We all know the problems of the German railway from painful experiences. I don’t want to defend Deutsche Bahn as an organization in any way here; rather, my point is that we have the densest rail network in Europe in Germany, with a length of over 33,000 kilometers – despite various extensive closures. It is first an enormous achievement that this complex system functions at all. If problems exist on the Rhine, the impact is naturally felt as far north as the south. It is simply a critical bottleneck, and trains run statically one behind the other without the possibility of dynamic evasive maneuvers. Of course, I am greatly simplifying the subject of rail here, and there are overtaking points. But the complex rail system is overloaded or lacks capacity, which is why there are increasingly obvious problems. Since the shortage, and perhaps mismanagement, has been fermenting for some time, every rail rider feels the need for more capacity and investment in infrastructure. Politicians and rail managers also argue with this, and demands for more money for the railroads usually reap approval across a broad front.
In most IT infrastructures, “the suffering” is less obvious. The systems are not as old as the rail network, and many admins sacrifice “something” out of “nothing”. Most IT infrastructures are about 20-30 years old at their core, and physical components such as servers and firewalls have been replaced regularly. Against this background, there is no “Deutsche Bahn effect” noticeable to most users; applications run, mail arrives, and all is well. This is where the “infrastructure bias” comes into play: IT infrastructure is maintained chiefly by stout-hearted people and suffers rather silently. Decision-makers notice in day-to-day business that the applications are running and everything seems in order, the infrastructure components are “invisibly” tucked away in cabinets, and everything goes about business as usual. Intangible components such as Active Directory directories, operational applications or authorization structures are often much older. Such literally “intangible” components are, by their very nature, even more invisible than the physical components of the infrastructure.
It is precisely this “invisibility” that threatens disaster; in most cases, long overdue renewals or reviews of the architecture are postponed. At some point, the day will come when significant damage is caused by a failure or, even worse, the entire operational process is interrupted by ransomware. Here, good advice is usually expensive, and the money is inevitably loose. For this reason, I advocate that companies invest more time, money and attention in precautions regarding their IT infrastructure availability.
One should not fall for the “nothing will happen at our company” thinking; the question is not if, but only when. Commercial decision-makers should develop a basic understanding of the complexity of IT infrastructures, but IT managers should also give them opportunities to build up an understanding of this complexity. Both parties should approach this topic with respect for “the art” of the other party.