Knowledge Popularization | 7 Deployment Solutions for 5G Private Networks

[[315546]]

What is 5G private network?

5G private network, also known as local 5G network or private 5G network, refers to a mobile communication network dedicated to a specific industry or enterprise.

Different from private networks, the 2/3/4/5G mobile communication networks that our smartphones are connected to today are called public networks. The public network is a "public place" where hundreds of millions of mobile users share the same network, frequency resources and network equipment.

The concept of private network is not unfamiliar. It has existed as early as the 2/3/4G era, such as railway private network, public security private network, military private network, etc.

As we enter the 5G era, private networks will become more and more popular.

This is because the greatest value of 5G lies in the 2B vertical industry market and in entering all walks of life. In the future, factories, parks, transportation, hospitals, oil fields, power grids, ports and other industries may have 5G private networks to use 5G's large bandwidth, low latency, high reliability, multi-connection network capabilities and edge computing to achieve digital transformation and improve production efficiency.

Since 5G private networks are so important, how can they be deployed?

This article introduces seven deployment solutions for 5G private networks.

1. Enterprises independently build their own private network based on the frequency band, which is completely isolated from the public network

Before explaining this picture, let us first understand the composition of the 5G network.

To put it simply, the 5G network is mainly composed of UDM, 5GC CP, UPF, MEC and gNB.

• •UDM (Unified Data Management), manages user data, including user identification, user contract data, authentication data, etc.
• •5GC CP (5G core network control plane), the 5G core network control plane includes multiple functions such as AMF and SMF, which are simplified in this article and collectively referred to as the control plane. It is mainly responsible for handling connection management, session management, mobility management, and carrying signaling or control messages.
• •UPF (User Plane Function) refers to the user plane of the 5G core network, which carries data traffic and is responsible for forwarding traffic between the wireless access network and the Internet, reporting traffic usage, and implementing QoS policies.
• •MEC (Multi-access Edge Computing), the 5G core network user plane UPF sinking and MEC can form an edge node, sinking cloud computing capabilities thousands of miles away to the edge close to the data source, allowing data to be stored and processed locally, thereby reducing network latency, better protecting the security and privacy of local data, and enabling various 5G private network applications.
• •gNB is a 5G base station.

As shown in the above figure, although the scale of 5G private network is much smaller than that of 5G public network, it has all the necessary functions, and is also composed of UDM, 5GC CP, UPF, MEC and gNB.

The 5G private network deployment method shown in the above figure is that vertical industries build their own 5G mobile private network including UDM, 5GC CP, UPF, MEC and gNB based on the 5G private network frequency band, and completely isolate it from the 5G public network.

The so-called 5G private network frequency band refers to the 5G frequency band specially allocated by decision makers for vertical industries. With the 5G private network frequency band, vertical industries are no longer restricted by operators in terms of spectrum resources, but can build their own private networks.

At present, many countries around the world have allocated 5G dedicated frequency bands, or so-called 5G local frequency bands, for vertical industries. For example, Germany has allocated 3.7-3.8GHz, the United Kingdom has allocated 3.8-4.2GHz, and Japan has allocated 2575-2595MHz, 28.2-28.3GHz, 4.6-4.8GHz and 28.3-29.1GHz.

The advantages of this deployment method are:

•Can ensure the absolute security of corporate data (because it is completely isolated from the public network)

• Low network latency (from the core network to the wireless access network, they are all deployed locally, and data transmission is only a few kilometers, so of course the latency is low)

• The network is autonomous and controllable (the end-to-end network is built by ourselves and is completely isolated from the public network. Even if the public network is cut by an excavator or congested, it will not affect the private network at all)

However, the disadvantages are also prominent. The main one is the high deployment cost, which is generally unaffordable for small businesses. The subsequent operation and maintenance costs are also high, and a group of professional operation and maintenance personnel must be trained.

2. Operators help enterprises build independent private networks based on public network frequency bands, which are completely isolated from the public network.

Similar to the first deployment method, the main difference is that the frequency resources used are not the 5G private network frequency band, but the operator's 5G frequency band.

For example, China Mobile currently owns 5G frequency bands of 2515-2675MHz and 4800-4900MHz. Based on this deployment method, China Mobile can use the 4800-4900MHz frequency band to deploy a 5G private network for a certain oil field that is completely isolated from China Mobile's 5G public network.

3. RAN sharing between private and public networks

Enterprise private networks still deploy UDM, 5GC CP, UPF and MEC, but 5G base stations (gNB) are shared with the public network.

Under this deployment mode, wireless data traffic is diverted on the 5G base station. Data traffic belonging to the public network will be transmitted to the public network UPF, while data traffic belonging to the private network will be transmitted to the private network UPF.

In other words, data generated by sensors and ultra-high-definition cameras in factories will be retained within the company, while data traffic from employees' smartphones such as surfing the Internet and watching videos will be transmitted through the public network.

The advantages of this deployment method are similar to those of the first and second methods, both of which can ensure the absolute security of enterprise data and low network latency. The difference is that due to shared RAN, enterprises can save some deployment and maintenance costs.

4. Sharing of RAN and control plane between private and public networks

Not only the 5G base stations (gNB) are shared with the public network, but also the control plane is shared with the 5G public network.

The so-called control plane sharing means that the control plane functions (authentication authentication, mobility management, etc.) of the private network and the public network are all performed by the 5GC CP and UDM in the public network.

In this deployment mode, the gNB and UPF of the enterprise private network are connected to the 5GC CP of the 5G public network through the N2 and N4 interfaces respectively, and the user information of the devices in the enterprise private network is also stored in the operator's 5G public network instead of being stored within the enterprise. Therefore, data security and privacy protection may be a little worse.

However, since MEC and UPF are still deployed within the enterprise, low network latency can still be guaranteed.

5. End-to-end sharing of public and private networks

From UDM, 5GC CP, UPF, MEC to 5G base stations, including the control plane and user plane, 5G private network and public network are shared end-to-end.

It is actually end-to-end network slicing. Based on the 5G public network, a "slice sub-network" is cut out for the 5G private network end-to-end.

In this deployment mode, the security of user information and data traffic depends on the network slicing capability. The low latency guarantee depends on the deployment location of the operator's edge cloud (UPF and MEC). If the operator's edge cloud is located close to the enterprise, the network latency will be low.

6. N3 LBO

LBO, Local Break-out, local break-out.

Let’s first explain the new device in the network, MEC DP, which is the data plane of MEC. It is responsible for providing a data forwarding path between the wireless access network and the core network to achieve local offloading of data traffic.

MEC DP has the ability to parse and process N3 GTP data flows. For example, MEC DP parses the target IP address of the GTP packet from the base station. If the IP packet is local traffic, it will be routed to the internal private network. Therefore, through MEC DP, public network traffic and private network traffic can be separated, thereby ensuring the security of private network data.

But why not deploy UPF directly in the enterprise private network like the fourth deployment method?

Because UPF is too expensive. The cost of UPF is much higher than MEC DP.

Therefore, the N3 LBO deployment method is conducive to reducing the cost of enterprises deploying private networks.

7. F1 LBO

Here we need to explain the architecture of 5G base station (gNB).

The 5G base station is divided into three parts: CU, DU and RU:

CU, the central unit, handles high-level protocols such as RRC and PDCP and is responsible for non-real-time configuration and control decisions.

DU, distributed unit, handles layer 2 functions with high real-time requirements and some physical layer functions.

RU, antenna unit, refers to part of the physical layer and RF, antenna part.

Among them, CU can be deployed in a cloud-based manner and can be integrated with the core network UPF and MEC. DU/RU connects to CU through the F1 interface.

It is not difficult to see that the difference between F1 LBO and N3 LBO is that the 5G base station adopts an architecture that separates CU from DU and RU, in which CU, UPF and MEC are deployed in the operator's edge cloud, while DU/RU are deployed within the enterprise. In this way, private network traffic no longer goes through the N3 interface but through the F1 interface to the edge cloud.