5G and edge computing: a match made in heaven

5G’s innovative network architecture is often combined with other new technologies, such as edge computing. We talked to Wouter Van Tilborg, Innovation Manager at Orange Belgium, and asked him what edge computing is, why it’s relevant to 5G, and what the major applications and challenges are.

What is edge computing and how is it related to 5G?

Wouter Van Tilborg: Edge computing processes data close to the source rather than in central data centres. That general principle can be implemented in many ways. One example is edge AI, where a camera analyses images locally rather than sending them to a data centre.
With Securitas and Robovision, for instance, we are working on a 5G use case, where cameras are placed near glass bulbs and recycling containers. Through edge processing, analysis of the environment can be done quickly and locally.

In the context of 5G, edge computing is mainly about bringing 5G network equipment closer to the end user. It reduces the distance between the antennas and the core of the 5G network on the one hand, and the end user’s data processing, on the other. That is how 5G achieves ultra-low latency.

So what exactly is ultra-low latency?

Wouter Van Tilborg: Ultra-low latency is a response time under 10 milliseconds. This is necessary for industrial machinery that needs to be instantly and automatically adjusted based on measurements from wireless sensors, or for a power grid that needs to instantly respond to spikes in energy consumption or production, for example.

The antennas and core of the 5G network are kept in the same location to keep latency as low as possible. Telecom operators implement this principle by placing their 5G core equipment in the customer’s data centre.

Doesn’t that present certain security challenges?

Wouter Van Tilborg: It does, because of course we still have to comply with regulations, such as NIS2, an EU directive on the cybersecurity of network and information systems provided by telecoms operators. 5G does have the advantage that every network function is virtualised and runs as an autonomous service. All those services communicate with each other via a message bus. It is important that those communications between services are sufficiently secured and that there is always a trade-off between security and functionality.

Are there other scenarios where it’s advantageous to combine 5G and edge computing?

Wouter Van Tilborg: Edge AI, the technology I mentioned before, is now emerging. Cameras are given computer power so that recognition of number plates, faces or objects can be done by the camera locally without having to send the images over the network. Performing the processing locally avoids additional network delay and saves bandwidth.

However, we often need to refine such image recognition with data that is not local to that particular computer. This does require the images to be sent over the network to a server in the cloud where those images are used to retrain the AI model. 5G’s high bandwidth is perfect for uploading large amounts of camera footage.

Will 5G enable more new edge computing applications to be developed in the future?

Wouter Van Tilborg: Definitely. Operators are already placing 5G network functions close to the end user. The next step is to roll out customer applications on the same edge infrastructure, such as Push-to-X applications. For example, we are working on a crucial communication network for the Belgian Ministry of Defence, linking edge computing to 5G. With AI models and workflow engines for augmented reality glasses, 5G will enable even more applications with edge computing. If you have invested in edge infrastructure to ensure the scalability of the 5G network, you can also use it to offer applications in a managed service model. We are now looking at those options.