Spotlight on Tech

Planes, Networks and Automobiles – Getting to Autonomy

Anshul Bhatt
Chief Product Officer, BU Intelligent Operations
Rakuten Symphony
February 20, 2024
minute read

Network automation is critical for mobile operators, but it can feel like an overwhelming task. But looking at the future evolution of your network, motivation to begin the journey towards autonomous networks isn’t too hard to find.

For example, machine-to-machine communications will need to support billions of attached devices as sensors, home appliances, cars and other machines seek network access. Growth in 5G networks and beyond will perpetuate virtualization, massive MIMO, edge computing, network slicing technologies, programmable networks and more. All of these contribute to better services and a more manageable network - at the cost of significant complexity. Adding more engineers to look after these networks is not scalable or cost-effective. Neither is the old telco model where you could offload the increasing costs to consumers.

Understanding that AI and the journey towards autonomy can be mapped out over time can make the tasks feel doable because a giant end destination is achieved by taking reasonable steps.

In this series of blog posts, we want to take a deep dive into AI by providing an understanding of the steps in the automation to AI journey and then exploring use cases that can make the most impact on Rakuten Symphony services and contribute to reaching the full autonomous network goal.

In this blog, we will be describing the major phases of the autonomous networks journey and what the benefits are of having an autonomous network. To that end, the mobile network industry has borrowed from the automotive industry and the airplane industry to explain the steps to automation success and the benefits of that success.

The Journey from Manual to Autonomous

Autonomous cars have captured the driving public’s attention for several years now and in considering how to get from no automation to fully autonomous, the automotive industry has developed a six-stage taxonomy that describes the journey to autonomous driving and what to expect at each step of the way.

The TM Forum has adopted a similar model for the mobile industry to describe the network automation process. These steps include:

  • Level 0 no automation – this is self-explanatory and describes a fully manual network that is slow to react to issues or changes and requires extensive technical staff to stay operational or to update or change the network.
  • Level 1 assisted operations – Some task-focused automation is in place, but overall, the network still requires significant manual effort and logistics to make changes.
  • Level 2 partial automation – the network has islands of automated functionality that come from upgrading specific use cases. However, staff is still in charge and making decisions.

It might be helpful to know that most of the industry is somewhere in these pre-automation or early automation phases.

Levels 3-5 are the stages of automation where networks are starting to gain applied autonomy benefits.

  • Level 3 conditional automation – This is the first stage that features some amount of end-to-end automation. There’s still some human oversight needed but it’s for limited situations that are very complex.
  • Level 4 AI-assisted automation – Expanded end-to-end automation that comes with autonomy within certain boundaries or limitations. The result is a network that needs very minimal human engagement or intervention.
  • Level 5 autonomous networks – This stage is a complete intent-aware autonomous network with no human intervention at all for deployment and operations.

As we’ll see in a later post, there are smaller steps within each level to identify specific automation actions that solve identified network challenges while also helping to advance toward the goal of an autonomous network.

Autonomous Networks Handling Complexity in Real-Time

The benefit of reaching level 5 fully autonomous networks is explained by the plane metaphor which compares the automation for a jet airplane versus that required for a fighter jet. The jet is designed for stability and for a human pilot to be able to control. Jet airplane automation makes the pilot’s job easier and improves safety, but it doesn’t remove the pilot from any of the plane’s operations.

Fighter jets, on the other hand, operate at such great speeds that a human pilot can’t react fast enough in many situations. System wide automation is needed to keep these planes in the air. New automation software brings new capabilities to the plane. Like a fighter jet, 5G and 6G networks of the future will have a level of complexity and need a fast response time such that having autonomy provided by system wide automation will be mandatory.

Now that we know the path to autonomous networks, we can start mapping out the journey. The next post in this series will feature a number of use cases at various network levels to ensure the network maximizes the AI advantage at each level on its path to becoming a fully autonomous network.

If you are attending MWC, be sure to visit the Rakuten booth (Hall 2 Stand 2D46) where we will be demonstrating how we are making these new levels of automation possible.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Network Automation
Future of Telecom
How can Symphony help?
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
By clicking “Accept All Cookies”, you agree to the storing of cookies on your device to enhance site navigation, analyze site usage, and assist in our marketing efforts. View our Privacy Notice for more information.