During some meetings with one of our primary vendors and our IT leadership we were discussing managing the ‘cloud’. During the dialogue an interesting analogy was used in referring to the transition we are making. The example was given of the ‘stealth’ bomber that requires about 2,000 adjustments to be made every second in order to fly. No pilot can make that many decisions individually. The pilot’s role is different and requires them to ‘trust’ their automation and instrumentation more than ever.
We then talked about how managing the ‘cloud’ is similar as we have lost specific control on where an application is running. With virtualization, that app can shift to anywhere within the environment. At any given time, you may not know exactly which physical server it’s located. We have many people that are afraid to let go of that kind of control. But we need to get to a point where we worry less about the underlying infrastructure as our applications become agnostic to the hardware platform. As we cross over that physical boundary, it opens us up to more agility and flexibility. However, the transition is sometimes difficult and requires a concerted effort and focus.
Think back about that pilot. How do pilots make the transition from one type of plane to another? They don’t just jump into the cockpit and adapt while heading down the runway. Rather, they incorporate a very critical training device called a simulator! They spend hours and hours in that ‘safe’ environment that ‘simulates’ the ‘production’ environment. They can practice and crash without burning! Their reliance on the instrumentation evolves over time. They move from ‘blind trust’ to more of ‘experience’ as they continue the training and the simulation. The more hours in the simulator, the more confident they are and they better understand the capabilities of the new environment.
So how many of us provide a safe environment for our admins to really become familiar with the new environment? Too often, we cram the learning into 2-5 day training and then force them to go live. That’s a scary situation and can cause some real anxiety. That’s why so many resist the change. They just don’t feel comfortable with the ‘new environment’. We must provide ‘simulators’ and plenty of hours for folks to ramp up.
Another point that was brought up was the failures of so many organizations that attempt to run the new environment like they do the old. Same processes, same tools and same org structure. Which result in not maximizing the capabilities of the new environment and technology. Yet, if you tried to take the legacy environment and change your processes/tools you would break it. So there needs to have an approach that keeps legacy running down its’ existing path-while consistently looking for opportunities to improve.
The new environment, however, is going to require us to jump into ‘simulators’ and start testing out the new instrumentation and get to the point that we understand the new capabilities. We then will need to build new processes—most likely very streamlined in comparison to what we are doing now. These new processes should yield quicker turn-around times for the business and more standardized infrastructure. And to be successful, we may need to tweek our traditional silo’d organizational structures. Possibly create a separate 'Cloud Ops' group?
What are your thoughts?