Self-Driving Cars to Self-Driving Data Centers

Machine Learning and AI delivers on the self-driven promise

Self-driving cars are making headlines every day with the vision of a car that runs itself, maintains itself, prevents accidents, and gives an alert when help is needed. Such a concept gives rise to further thoughts, such as, what would a self-driving data center look like and how would it change IT as we know it? A self-driving data center is not just a vision, as we are already progressing toward full self-driving capability with different levels of automation and machine learning-based solutions to improve agility, reduce costs, and focus on higher-value business outcomes.

Ali Siddiqui, General Manager, Agile Operations at CA, and I first made this analogy in our  previous blog, and the idea sparked great interest, especially regarding how an existing data center with mainframe, which is facing a skills shortage, can be gradually turned into a self-driving, autonomous data center.

A Washington Post article by Aaron Cole of The Car Connection details five levels of self-driving cars – all of which have a direct correlation to the growth and development of the autonomous data center, automated leveraging machine learning and artificial intelligence:

Level 0:

This is our starting point. It’s what most of the human race drives today – cars with no automation, where the person at the wheel is in complete control. Historically, this has described the corporate data center perfectly. IT operational tasks are done by humans be it troubleshooting or root cause analysis across a complex technical stack. For example, with a Level 0 data center, if a server threshold runs out of capacity, you have to figure out if it’s an application problem, a server problem, a data flow problem, etc. – then set about fixing it. Not infrequently, multiple subject matter experts have to be called in. The mean time to repair is high, because the entire process is reactive in nature.

Level 1:

At Level 1, cars provide limited driver assistance, such as automatic emergency braking or cruise control. In the data center, this translates to basic automation in a single domain – say monitoring networks.  There is some automated data collection which increases productivity and offers insight to the network expert but the feedback loop is still delayed and reactive in nature.

Level 2:

More convenience enters the automobile world at Level 2, with features such as automation of both steering and speed. However, driver engagement is still a “must-have.” In similar fashion, data centers at Level 2 can recognize “normal” and “abnormal,” and can apply solutions to current problems that have been encountered and addressed in the past. While IT administrators still keep a watchful eye on the data center, routine matters can be taken out of their hands.

Level 3:

Here is where things start getting really interesting. At Level 3, conditional autonomy enters the picture, and cars can handle the road unless the automation system itself fails. Through machine learning, clustering, and correlation, a data center at this level can bring various knowledge feeds to bear on system activity, understand relationships and causality, compare situations to past events, remediate problems in real-time, and even predict issues before they occur while simultaneously recommending appropriate action. Unless the artificial intelligence itself is in jeopardy or an especially acute and unique issue arises, IT administrators are feel free to focus on other value-added IT projects and leave the system to run itself.

Level 4:

Cars are nearly autonomous at Level 4, able to stop themselves safely even in the event of a systems failure. For the data center, this means high automation that encompasses applications, workload inputs, and the whole business environment.

Level 5:

This is it: The Holy Grail where the driver becomes the passenger and the car is in total control. In the IT realm, this is the data center that is completely self-sufficient: better at healing itself, optimizing itself, and running itself than would be possible even with the most elite IT operations team. This will enable agility with control, which is the most essential building block of the future, so that business and IT leaders can focus on a rapidly-changing business environment to truly become a Digital Enterprise.


According to most predictions, fully self-driving cars are just a few years away from hitting the mainstream. Self-driving data centers are on a similar fast track, surging forward with all the power of today’s machine learning and artificial intelligence. These are exciting times – in the coming months, you will see how machine learning-based intelligent automation becomes a critical component of modern data centers, including mainframes powered by CA’s modern software factory and intelligent automation as a service. Stay tuned!

Ashok Reddy
Ashok is responsible for CA’s DevOps Line of Business including the Developer Products, Continuous Delivery,…


Modern Software Factory Hub

Your source for the tips, tools and insights to power your digital transformation.
Read more >
Low-Code Development: The Latest Killer Tool in the Agile Toolkit?What Are “Irresistible” APIs and Why Does Akamai's Kirsten Hunter Love Them?Persado's Assaf Baciu Is Engineering AI to Understand How You Feel