I´m a big fan of Autopoiesis, Cybernetics, Relational Biology, Dissipative Structures and other modern concepts related to systems organization. Recently I learned about the IBM proposal of Autonomic Computing. But I live restricted to use a very bad computer, a very slow Pentium II runnig Linux, and I tend to see things from this perspective.

I always complain about how my machine gets stuck sometimes. But every time I´m forced to use a Windows machine (at least a bad-managed one), I see things can be worse, and make friends again with my box.

Right now I sat in this Windows machine, worst of my lab, started the boot, and went for a coffee. Got back, still waited some more… Logged in, and the show begins. It´s always the same: I have to wait a long time between double-clicking the mozilla icon, and the actual program get open. I often double click twice or thrice, and every time I end up with more then one copy of Mozilla, of course. Yes, I know it´s my fault. But I refuse to be a slave of my machine. It´s a diplomacy thing, you must do stupid things sometimes to make the rules clear.

One of the reasons it takes so long is the anti-virus program. It always try to make an update, and it seems it´s a terribly hard proccess!... Then I cancel it… The whole proccess takes a lot of time to do nothing.

Today I decided to open Micosoft Word too. It took its time, opened eventually… Then I stared to look my mail box, and was surprised by an error window saying that Office couldn´t start the “help agent” or whatever, that little stupid paper clip that bugs you all the time. I was delighted. I hate the thing, and it bugs me even when it doesn´t work!... Changing for the Einstein avatar doesn´t make things better.

So, what I started to think isÇ how a system like this would evolve in an autonomic computing world?...

I always say half-jokingly that an autonomic Windows system would boot up, then remove itself and put Linux in the box instead.

But seriously, it would be interesting to see if a computer managed by a bad autonomic-computing system would get stuck with things like installing and updating anti-viruses…

If an autonomic manager only think about functionalities, it will start to fill the machine with things like anti-viruses, and “help wizards” and all, and the machine could end up wasting all its resources in those stupid features that doesn´t help at all. It could eventually use up all its resources in pointless tasks, as it happens today with the slow and old Windows machines in the lab.

So, an autonomic computing system must watch out so its autonomicity doens´t become a virus (or cancer) eating up all resources in the machine.

There could be perhaps a scenario where the autonomous part of the operating system gets into a positive feedback loop, trying to do MORE things when it sees that there is LESS resources. But doing more, it will eat up more resources!!...

A good autonomous computing manager must watch out for its own size, then. It must sometimes make decisions about when to give up on watching out for viruses, so to give the user a nice Quality of Service…

At least in a poor-man´s machine like mine.