In some areas, often particularly important areas, IT seems to evolve by making mistakes and then expending effort to correct them. Think network security, think email filtering, think viruses….

For this reason, one guide to the future of IT is to identify the engineering mistakes that currently need correcting and wait for the products that correct them to appear. It may take time, because you have to wait until the engineering errors start causing problems. As the industry grows and the size of systems and networks grow, the problems will emerge because of the demand for scale. Also there’s no guarantee that the “solutions” that appear will solve the problem once and for all.

Note for example the virus problem. It was and still is an authentication problem. Software that had not been properly authenticated was being allowed to run on computers. The AV technology that emerged and made quite a few people wealthy, didn’t actually solve the problem. It didn’t provide a solid authentication process. The emergence of whitelisting products that provided a true authentication process was inevitable. but it never happened until scalability issues demonstrated that AV technology was broken.

Consider identity management. This is a market that appeared and grew from nothing to $billions very quickly. When you think about it, it all stems back to the fact that mainframes and computer networks never got round to establishing viable repositories to deal with a specific and fundamental issue. There needed to be a common point of definition for people (users), processes, data sources and hardware. At best, only half a job was ever done on any of these items.

The above diagram, which I refer to as “the diagram of everything” represents, in a very simple way, the components of a computer system; people, software, data and hardware. In order to actually know everything within a system a complete map of each of these elements and their relationships to other elements needs to be known.

The importance of having such a map increased as soon as computing environments became interconnected in a major way and the components needed to be flexible. As things stand, there is very little that is static and dependable in a computer network (over time) – and the situation is getting worse. So the whole environment needs to be tracked – at least it does if you want to do sensible things like enforce governance rules, monitor performance, optimize resource usage and all those good things.

The Temptation of the Cloud

Right now we have very little technology that can track the behavior of a network and manage it well. Management technology doesn’t scale. The identity management and user provisioning market sorted out some of the issues in some areas – security at least now has a basis to work from. Another market that is on the rise, because the capability is desperately needed, is the market for asset management products. Real-time discovery of hardware and software assets is becoming increasingly urgent.

However, I get the sense that, from a technology management perspective, many things are out of control – although they were never properly in control – and they are getting more out of control as time passes.

This is one of the huge attractions of cloud computing. If you put as much as your computing as possible into the cloud then you will be spared the task of managing it. It’s a way of “scaling your IT down.” You lose complexity without losing capability.

You can think of it as core-competency computing. You only keep control of the applications that really matter to your organization and you consign the rest to the cloud. The management problem then sits with those who, like Google, can manage very large networks. It’s the direction that the IT industry will inevitably take.