Contrary to popular belief, automation does require some level of human interaction. For example, when you step into a self-driving car, it does not automatically know where to go or when to start driving—you have to start the car and tell it your desired destination in order to achieve your goal. Automation, therefore, can only complete a task that has been initiated by a human.
A Rube Goldberg machine is an overcomplicated contraption that performs a very simple task through a chain of events. For example, you push a ball down a small ramp; the ball hits a row of dominoes that knocks down a small pole attached to a string; the string pulls a lever, which releases a piece of candy. Similar to a Rube Goldberg machine, all types of automation needs a little human push to initiate the process and complete the task—in this case, pushing the ball down the ramp.
There are different stages in the automation journey towards The Self-Driving NetworkTM: human-, event-, and machine-driven, each of which must be initiated by a human to complete its assigned task. With human-driven automation, the individual determines what they want to accomplish by providing initial scripts and programming. With event-driven automation, you must gather the necessary information and network events to establish a closed loop and automatically make a decision. While this may seem like a fully automated cycle with no role for a human, the reality is that people actually play an important role by building the rules-based logic that drives this engine. Machine-driven automation is when a machine leverages learning tools to make necessary and informed decisions; humans are only required to take actions when the machine cannot.
There are endless possibilities with Juniper’s automation. I invite you to explore one particular use case by watching our latest webinar to learn how Groupon used Junos OS Automation throughout its data center. For even more examples, visit the Automation Use Cases. All it takes a little push…