When we think of autonomous driving, the first thing that comes to mind is self-driving cars—vehicles that operate without us having to do anything more than pressing the "start" button or simply giving the command to begin their journey.

But what does autonomous driving of a system really mean? 

Although the concept of autonomous driving isn’t fully standardized, there are some references that help us lay the groundwork to understand it better:

Automobiles

In the automotive field, the Society of Automotive Engineers (SAE) has established a standard definition for vehicle autonomy, describing its ability to make decisions independently, without human intervention. This autonomy is classified into different levels, from level 0 (no automation) to level 5 (fully autonomous). The SAE J3016 standard defines these levels of autonomy for vehicles, providing clear guidance on what constitutes an autonomous vehicle.

Industry 4.0 and Robotics

In the context of Industry 4.0 and robotics, the concept of autonomy is highly developed. It refers to the ability of a machine or system to operate without direct human intervention, following principles of optimization, automatic adjustments, and data-driven decision-making. Standards such as ISO/IEC 18629 define automation and its applications in the industry, considering autonomy as the ability to execute processes without constant supervision.

Information Technology (IT) and Data Centers

Focusing on our own environment—autonomous systems in IT—there is no formal definition of “autonomous driving” as a specific term. However, we can find related concepts in publications from organizations such as:

• NIST (National Institute of Standards and Technology): which addresses autonomous systems and cybersecurity within its framework for smart networks.
• IEEE (Institute of Electrical and Electronics Engineers): which publishes frameworks and definitions on system automation, including how certain processes within technological infrastructures can become autonomous, along with ethical considerations—an interesting topic to explore in another article.

Based on all this, a possible definition of autonomous driving of a system would be: “The ability of a system to operate independently, making decisions and performing tasks without direct human intervention, but with controlled supervision based on data.”

How do they do it?

Looking at the automotive industry, one of the most significant advancements has been the automation of manufacturing processes. Assembly lines have been automated for decades. However, what will make cars truly “autonomous” is their ability to make real-time decisions using sensors, algorithms, and interconnected networks The same concept applies to manufacturing processes: factories not only produce cars more efficiently but also more accurately, thanks to every part of the process being connected to a network that ensures everything runs smoothly.

A similar example can be found in precision agriculture, where farmers use drones and sensors to monitor and optimize crops, making data-driven decisions based on terrain conditions, humidity, and sunlight. Even though farmers aren’t physically in the field, they can manage processes more efficiently thanks to systems that process this data and provide accurate insights.

In energy management, smart grids allow companies to autonomously manage electricity consumption. These systems regulate energy usage based on demand, optimizing resources and preventing excess, with decisions made by algorithms that keep everything under control.

How does this translate to the Data Center?

This is where our deep reflection began. We asked ourselves: What are other sectors doing that we aren’t? And we went back to the basics, reflecting on how Data Centers operate—where the data is, how it’s collected, whether it’s reliable, where processes and people are involved. Unfortunately, we reached the same conclusion as 15 years ago: we are still working in silos, integrating technologies only partially, and—most importantly—we don’t fully trust the information our systems provide.

Looking at sectors like logistics helped us simplify the idea. A Data Center is a critical environment, but at the end of the day, what is a Data Center if not just a “box”?? Upon closer analysis, it’s a box where everything flows: equipment coming in and out, maintenance, supervision rounds, audits… Everything can be simplified if we understand that every action happening in a Data Center should be recorded, traced, and understood—just like in a logistics chain.

So, the first key concept is The Box: the place where things enter, leave, and move, all within a process that can indeed be automated.

What else do we need? Information to train the system so it can make decisions. This information comes from the entire Data Center ecosystem—processes, equipment, people, and existing technologies. That’s why it’s crucial that every process follows clear steps—from equipment entry to exit—whether as an action (like repairing a server) or as information (like a usage report). Everything must have traceability to register these inputs.

 What do we get from this? Those outputs will mark the beginning of automation, enabling the system to make decisions and execute them.

When? Only when we have trust in the system and, above all, confidence that the information from those inputs is accurate. This will only happen if all processes inside the box are fully traced and no step is skipped. (Again, processes are always key).

In summary:

   The Box: The place where things enter, leave, and move, all within an automatable process.
   Input: The information needed to perform any task within a Data Center.
   Output: The result of that process, fueled by information from the inputs.

This inevitably leads us to the connection between processes (through standardization and automation) as the path to achieving autonomous driving of the Data Center.

Is that all? Not quite! For this to become reality, we need crust and security, which can only develop over time. Think about autonomous cars—we’re still fearful and don’t fully trust the technology. Yet, 25 years ago, we never imagined a car could park itself without human intervention, and today it does. This phenomenon is called desensitization and is part of the Technology Acceptance Model (TAM) proposed by Fred Davis in 1989. This model suggests that technology adoption depends on two key factors: perceived usefulness and perceived ease of use. As users interact with a technology and experience its benefits, their attitude toward it improves, making adoption easier.

So… hopefully that moment will come! Hopefully, we’ll have technologies that truly overcome these barriers, being useful and easy to use. At least on our side, at Bjumper, we’re going to try!