What Can IT Operations Learn from Self-Driving Cars?
Self-driving cars are quickly becoming a reality. Tesla owners can now download a new software update that enables the car to steer itself, although not without controversy. Google has logged over a million miles in self-driving cars, and other companies including Nissan, Volkswagen, and BMW are also experimenting with the technology.
And while there are still issues to sort out, some lessons can already be learned from the self-driving car. As IT teams are turning to automate their software-defined datacenters, they can apply some of these takeaways to ensure automation doesn’t get in the way of safety and stability.
Automating the Right Behavior
The self-driving car is driven by some pretty sophisticated software. But before even one line of code was written, engineers had to build a comprehensive model of the perfect driver.
For IT teams, the temptation to just start coding automation scripts is clear. However, a great deal of trouble can be avoided down the road if you take the time to ensure the configuration you are modeling is thoroughly vetted before any scripts are written.
Even if you think your datacenter is free from misconfigurations, better safe than sorry is the approach to take here. Every new company we work with starts with a quick health check of their environment, and each and every one of these health checks has revealed some hidden configuration risks that were unknown to the IT team. While your company may very well be the one exception to the rule, is this something you want to bet on?
Accounting for Interference
Modeling the perfect driver may be enough for running a car in a closed-loop racetrack with no one else in sight. Safely navigating on a busy road requires modeling of the dynamics of other drivers, pedestrians, traffic lights and signs, obstacles, and other objects around you. Obviously, this adds compounding layers of complexity to the processing requirements of the software running the car.
That’s true for the software-defined datacenter as well. Creating a perfect script for automating storage provisioning is practically useless if it creates a discrepancy with your database layer, to give just one example. This can get tricky, as the database configuration itself may change after the script was written. Having the ability to quickly test these interdependencies across the IT infrastructure is a key requirement for the safe deployment of automated provisioning in today’s dynamic datacenter environment.
The self-driving car is equipped with an array of cameras, radars and sensors that constantly monitor the car operations as well as its surrounding. If the car is changing lanes and another vehicle is moving too close, the car will automatically change course and return to its original lane.
While your datacenter’s environment doesn’t change at the pace of a car driving 60 miles per hour on a highway, changes in the datacenter are frequent. What was deemed to be perfect yesterday may no longer be at the same state today. Daily changes to configurations across all layers of your infrastructure, minor as they might be, may inadvertently introduce critical risks that could result in a system failure and unexpected downtime. In an environment that keeps changing on a daily basis, ongoing validation of the entire environment across all domains is required with every change.
The Bottom Line
For all the power of automation, the right checks and balances must be put in place for safe operations of a car on the road, and the same holds true for your datacenter. The speed and efficiency afforded by automation must be met with the agility and effectiveness of software-driven automated monitoring and validation.
To learn more, check out our eBook on Agile IT OperationsGET FREE EBOOK