Are things becoming too automated? That is a question I cannot believe I am asking. I have spent most of my virtual career creating automated processes from my virtual environments. I spent a great deal of time creating a “toolbox” of scripts to perform almost all of the tasks that I found myself needing to do in an automated fashion. A lot of my peers and I were all creating automated build and automated configuration of our environments from the very beginning and now are seeing a great deal of the automation that we created in one way or another being added to the product suite. Is that a bad thing? No, on the contrary, this shows the active deployment of the technology and gives us insight to the direction the technology may follow.
Here is where things become different. In the “old” days, we admins were building and creating almost all of the automation processes ourselves. We pretty much knew it all like the back of our hand, and we had a complete and solid understanding of how things work. Once the creation process for the automation is moved away from the administrators of the environment, some of the knowledge and understanding of what is really happening under the covers disappears. Of course, this makes things easier for most levels of administrators when setting up and getting the environment deployed, but can it easily take away from the knowledge or understanding of how things really work.
There are two sides to any coin, and the positive for built-in automation is the ease of use for almost all environments, which makes maintaining the environment easier. But, virtualization and cloud computing are also continuing to mature and branch out from the idea that a single hypervisor will handle all our needs, and tools are being built to “connect” to different hypervisors and different public clouds. Currently the automation from one cloud to another is still in its infancy and still has a way to go, but all this automation needs to be really looked at and not just used.
Consider what this automation is really doing. We are moving virtual machines and/or applications from one environment or platform to another. When crossing different platforms, the way security is handled and presented can be quite different from each private platform, and even more so when you add the public cloud into the mix. Consistency throughout the environment, as well as security, can go right out the window as soon as the migration completes.
Trusting technology is a good thing, and this trust helps a lot of us administrators sleep at night, but we must not get too complacent about the technology and must really get back under the hood to know exactly what changes inside and out of the virtual machine or application. This understanding needs to be fully realized and documented for migrations going in either direction, to or from our different platforms. We must not trust without digging in and examining the automation in its entirety so we do not create or allow unattended access and/or unattended changes to happen. Maybe things are not getting too automated, but really our understanding of the automation is what has truly fallen behind.
Great post! We absolutely not trust automation we can’t create/understand by ourselves. Otherwise how will you know how to prevent the automation from becoming self-aware?
Only YOU can prevent SkyNet!