Every private 5G network has a core: the part that lets devices join, keeps track of them as they move and routes their data. Where the private 5G core runs, on site or in the cloud, decides whether your network keeps working without the internet. This page explains the three common designs in plain language and the questions to ask before you buy.
Does private 5G work without the internet?
Yes, but only if its core runs on site. A private 5G network keeps working without the internet when the core is on site (on-premises). If the core, or the part of it that lets devices join and move, runs in the cloud, a lost internet connection can stop devices from connecting, even though the radios on site are still powered on.
Key facts
- The core handles three jobs every device depends on: letting it join, keeping it connected as it moves between radios, and routing its data.
- Research on private 5G networks that move between sites reaches the same conclusion: for local services such as device authentication to keep working, core functions have to run locally.
- Keeping data traffic on site (“local breakout”) is not the same as running the core on site. The site can still depend on a controller elsewhere to keep working.
- With the core on site, an internet outage only affects things that need the internet, such as email, cloud apps and remote management.
What the core does
The radios on the ceiling or the mast carry the signal. The core is the network’s brain. It checks each device’s SIM before letting it on the network, keeps the connection alive when a robot or scanner moves from one radio to the next, and decides where the data goes. Without a working core, a device can see the signal but can’t use the network.
Three ways to run a private 5G core
1. Core in the cloud
Only the radios are on site. A cloud 5G core runs in a data center or public cloud and is reached over the site’s internet connection. This keeps the on-site equipment small and simple to manage centrally. The trade-off is dependence: if the connection to the cloud is lost, devices may not be able to join or stay connected, so the site’s operations depend on its internet line.
2. Split design with local breakout
Data traffic stays on site, but control of the network, such as letting devices join and managing them as they move, stays with a central controller elsewhere. Your data doesn’t travel to the cloud, which helps speed and privacy. But if the controller can’t be reached, the site can still be affected, because the part that runs the network isn’t there.
3. Core on site (on-premises)
The whole core runs on site, together with the radios. A remote controller can still be used for monitoring, updates and expanding the network, but the site doesn’t wait on it. If the internet goes down, devices keep joining, moving and sending data on site. Only services that need the internet wait until it’s back.
Side-by-side comparison
| Core in the cloud | Split with local breakout | Core on site | |
|---|---|---|---|
| Where data travels | Through the cloud | Stays on site | Stays on site |
| Who lets devices join | The cloud | A controller elsewhere | The site itself |
| If the internet goes down | Devices may lose the network | The site can still be affected | On-site operations keep running |
| Equipment on site | Radios only | Radios and a local gateway | Radios and a gateway or server |
| Best fit | Sites where a short outage is acceptable | Sites that need data kept local | Sites where operations must never stop |
Designs vary between vendors, and some cloud and split designs keep existing connections alive for a while after the link drops. The only way to know is to ask, and to test it.
Why it matters more as sites automate
A person with a scanner can wait a minute for the network to come back. A fleet of robots can’t. When a robot loses its connection it stops, and every robot behind it waits. On a site that runs on robots, cameras and connected machines, an internet outage that stops the network stops the operation. That is why the location of the core has become a business question, not only a technical one.
Questions to ask a private 5G vendor
- Where does the core run: on site, in the cloud, or split between the two?
- If the internet goes down, can a device that was switched off join the network?
- Can a device move from one radio to another while the internet is down?
- How long can the site run without the internet: minutes, hours or indefinitely?
- Does any license or security check need to reach the cloud for the network to keep running?
- What exactly stops working during an outage, and what happens automatically when the connection returns?
Ask the vendor to show it: unplug the internet during a pilot and watch what the devices do. For a side-by-side view of how vendors describe their own designs, see the private 5G vendor comparison.
How EdgeNectar runs the core
EdgeNectar builds autonomous private 5G with the whole core on site, together with AIDEN, EdgeNectar’s AI engine that monitors and fixes the network. If the site’s internet connection goes down, robots, scanners, cameras and other on-site operations keep working, and AIDEN keeps looking after the network. Only apps that need the internet, such as email, wait until it’s back.
EdgeNectar’s remote controller is used to monitor the network and to add radios or SIM cards when the site grows. When the internet connection returns, it reconnects on its own. See how the architecture works.
FAQ
Does private 5G work without the internet?
It can, if the core runs on site. Then devices keep joining, moving and sending data during an internet outage, and only services that need the internet are affected. If the core runs in the cloud, devices may lose the network when the internet goes down.
What is a private 5G core?
The core is the part of a 5G network that lets devices join, keeps them connected as they move between radios, and routes their data. Every private 5G network has one, either on site, in the cloud or split between the two.
Is local breakout the same as an on-site core?
No. Local breakout keeps data traffic on site, but the part that controls the network can still run elsewhere. An on-site core keeps both on site, so the network doesn’t depend on a remote controller to keep working.
Is an on-site core harder to manage?
It doesn’t have to be. With autonomous private 5G, the on-site network monitors and fixes itself, and a remote controller handles monitoring and expansion, so the site doesn’t need a telecom team.
Sources
- Kochems and Schotten, “AMMCOA – Nomadic 5G Private Networks” (arXiv, 2018): “for certain network services, like authentication of UEs, to be available, some of the core network functionalities have to be instantiated locally.”