NETWORWORLD has a review about Niccholas Carrs’s new book where he argues that the IT department is dead. Carr is known for his provocative views. However, you often hear similar assertions about the impact of cloud computing these days. Usually, they are about the showdown between Google and Microsoft. Many think that Google can replace applications on the Windows desktop with their web-based applications. We are living in a fast moving world. So it can’t be wrong to listen to these assertions every once in a while.

Maybe today’s web applications are not yet powerful enough to rival with desktop apps. But that might change in the future when Rich Internet Applications (RIAs) become more widespread. The question is if most of the computing is done somewhere in the cloud, i.e. in large computer centers, who then needs a Windows admin? You might argue that these computer centers are operated by admins, too. The point is, however, that you obviously need less system operators if everything is centralized.

This line of argumentation is in a way plausible. Cloud computing is on the rise, there is no doubt about it. However, I think there are also some arguments that this development won’t have negative effects on the future need of IT administrators.

First of all, the fact that more computing is done in large computer centers in the future doesn’t necessarily imply that there will be less computing on the desktop or on corporate servers. We see these giant computer centers already rising for quite some time. But the computing that is done on desktops has been growing at the same speed. One major reason for this is that the user interfaces get more and more sophisticated. I believe that we will see some major innovations in this field in the near future. I am talking about 3D, touch screens, speech control, virtual reality and such stuff. We will need any number cruncher available to do the necessary computing of these new user interfaces. Most of them are not yet available for ordinary desktops because they don’t have enough computing power. Note that you’ll also need more computer power in the data centers for these applications. I think Second Life is a good example. Without a modern PC, Second Life is not really fun. And their graphics are fairly simple compared with modern computer games.

Second, I doubt that all companies will outsource their computing. The IT infrastructure is essential for most businesses nowadays. Hence, many won’t just rely on external computer power. Another point is that that most companies have confidential data that they don’t want to be seen on the computer screens of their competitors. It is a matter of fact that the probability of your data getting into the wrong hands rises if you store it somewhere in the cloud. And last but not least, companies who are able to control all their IT themselves are more competitive because they can react much faster to innovations. This was always a major disadvantage of IT outsourcing and I don’t see how computer clouds could change this. Any kind of centralization takes its toll on flexibility.

Third, even if a bigger share of the computing is done in external computer centers, it will not have a negative effect in the need of in-house system operators. We already outsourced some of our backend apps to larger computer centers. But this didn’t reduce the number of people who manage these systems in my department. On the contrary, we need more and more IT staff. The applications are running on servers that are operated by others now, but we still have to manage the applications ourselves. The complexity of software increases steadily. So you need more people to manage them who know about all their settings. These settings are different in every environment and they have to be changed more frequently than in former times. That’s why admins who are familiar with the local environment and who are always available have to do this work.

I am sure there are many other arguments for and against Carr’s claims. What do you think?