by Bryon D Beilman
I was reading the information week article "Automating the Private Cloud" by Jake McTigue and it caused me to pause and think about how we automate IT Management for our own customers and what we see inside small, medium and enterprise IT environments. Jake's article talks about what companies are doing to build private clouds which is essentially commoditizing hardware to dynamically handle business work loads. The one area that is perhaps overlooked is to automating the environment to provide resource pooling, self-healing and improved application availability.
One of the best practices that we use is to build a customer Run Book, that has all the details about how the environment is built, managed and how exceptions and issues are handled. Our customers appreciate this and we are essentially defining the processes and steps so that a monkey can handle the core IT tasks. We try to make the common tasks so well defined that it can be done perfectly every time by a lower Tier 1 person or someone who is not familiar with the environment. I don't think any of us want to be compared to monkeys, but by documenting, and scripting the environment so that a monkey can do it allows us to focus on interesting technical and business challenges and do the uninteresting and repetitive work in a quick, repeatable way.
The next evolution is to really turn monkeys into robots. In Jake's article he talks about different cloud automation software that allows the management task to be handled automatically by a given set of rules. This concept has been around for a long time. In 1999, I remember a product called BMC Patrol which has now been integrated into their full management suite. Now 14 years later BMC and others all are focusing their efforts on the cloud (private and public). We too are focusing our efforts on using a new product suite that moves us from RunBooks , monitoring and alerting to doing some automation and self healing. We don't use the BMC product, but another one that is designed for IT consulting and distributed environments. In most operating systems things can be abstracted to a file or a process and those can be monitored, restarted, checked and proactively managed.
We again are not trying to offend anyone by saying your monkey job is now being replaced with a robot, but we don't have that issue at our company because our staff are all focused on helping our customers solve interesting business and technical problems and to be able to program a robot to do the job is perhaps more interesting than doing the job itself. Not all jobs can be done by robots , or monkeys for that matter, but as companies like i-robot have shown, letting robots do some of your daily tasks is not such a bad thing.