I’ve been watching the rise of OpenStack with an ambivalent mixture of curiosity and dread over the past few years. With the recent announcement of OpenStack support baked into Red Hat Enterprise Linux 7, it seems OpenStack is emerging as the de facto standard for virtualization and abstraction in the private could. Another new requirement for the working systems administrator. It could be that I’m just getting old, but when I look at the diagrams of OpenStack, I see a lot of moving parts.
My philosophy in dealing with computers is to always keep it simple. In my experience, the more interconnected components there are to an application, the more likely it is to break. Likewise, the more complex, the harder it becomes to diagnose the failure and fix what’s broken. When I look at this diagram by Ken Pepple, I see moving parts, connected like gears, and I have to ask “how fragile is this?”.