Private 5G Fundamentals

Introduction to private networks

Enterprises or industry verticals are undergoing a rapid digital transformation driven by a need to achieve better operational efficiencies and overcome industry-specific challenges. In conjunction with high bandwidth, low latencies, and faster data speeds, highly reliable and secure connectivity private 5G paves the way forward for digital transformation.

What are private networks?

A mobile private network is a local area network deployed for the use of authorised devices within a specific geographic area such as manufacturing concern or any campus setup. Due to security concerns, authorised devices that can use public networks may not have access to private networks.

Figure 1: Private 5G enabled Bosch campus in Germany

In terms of the network technology stack, public and private 5G does not have much difference. The accessibility rights and the security measures are the major differentiators. These networks are commonly owned and controlled by a single enterprise for reliable connectivity and secure communications essential for the nature of services an enterprise offers.

Enterprises can deploy private networks both physically and virtually.

Why private 5G?

Private networks have been in existence for a couple of years now. Private networks have witnessed their presence in various network technologies such as 3G, 4G, and 5G. Today, 4G private networks have a dominant share as they are self-sufficient for a majority of the use cases and applications.

Private 5G networks help address the connectivity needs of a diverse group of devices in a wireless mode.
These networks play a significant role, especially for applications and use cases that are mission and business-critical and demand a real-time output such as AGVs, asset tracking, predictive and preventive maintenance.
Besides industry 4.0 and public safety, utilities represent the other areas where there is traction for private 5G networks. The following attributes of private 5G make the implementation of these use cases successful:

High availability – High availability ensures the high availability of connectivity services.
High reliability– Reliability refers to the capability of transmitting a given amount of traffic within a predetermined duration with a high success probability. It requires sufficient network coverage and capacity, as well as robust handover functionality.
High Security– Private networks enable storing data locally/on-site, which is not possible in the case of public cellular networks. Therefore, it is applicable for businesses where data security is paramount. This feature ensures protecting people, processes, and technology are protected from any outside threats.
Interconnection – The flexibility of private networks to use public networks as and when required. This model is suitable for moving objects like connected or smart ambulances or vehicles in logistics. Such use cases often have to move between both public and private networks. Hence, better integration between both makes the implementation of these use cases a success.
Customisation of services– Enterprises can customise the network needs through service-based architectures.

Private 5G vs legacy networks

Figure 2 below shows the comparison of 5G with other legacy networks used for enterprise use cases such as Wi-Fi and 4G LTE:

Comparison of private 5G with 4G LTE and Wi-Fi

Source: twimbit analysis

Private 5G network technology stack

The private 5G components are fully deployed on-site and owned by an enterprise, or some components are in integration with an MNO’s network. As the report progresses, we will examine various deployment models. Below are the essential components required to build a successful private network:

Spectrum
Radio Access Network (RAN)
5G Core network and control plane
UDM
Public edge and center cloud
On-site edge cloud (MEC)
UPF

Types of spectrum for private 5G

There are two broad categories of spectrum viz., licensed and unlicensed

Industrial spectrum – Government or regulators allocate spectrum in different bands for private dedicated enterprise usage in some countries. For example, In 2021, the government of South Korea announced to auction 600 MHz spectrum in the 28 GHz band and 100 MHz in the 4.7 GHz band for enterprise use.
Public spectrum – Operators acquire this spectrum through a competitive bidding process or in a few countries via beauty contests. The enterprises can further avail this spectrum on a lease basis
Shared spectrum – Multiple stakeholders can share a spectrum. A shared spectrum, as the name suggests, can be shared amongst multiple stakeholders. The spectrum is either administered by the regulator or a certified third-party administrator.
Unlicensed spectrum – Unlicensed spectrum bands are naturally inclusive as any enterprise can freely use them, but there is a high risk of a data breach. Figure 3 shows some of the examples of unlicensed spectrum band allocation:

Source: Qualcomm

How Mobile Edge Computing (MEC) plays an important role?

Mobile Edge computing enables the computing process in proximity to the end-user (Edge Cloud), thus reducing the amount of data travel to the centralised cloud or data center. Private 5G, when combined with edge computing, leads to low latencies and high performance. However, the choice of network deployment models determines whether edge computing occurs on-site or on the public network. Further network slicing capabilities allow the enterprise to prioritise data that need faster processing at the edge and is more critical than the other.

MEC combined with 5G is more than just low latencies

Data sovereignty, security and privacy, is ensured
Customisation of end-user services generates a diversified revenue stream
Reduced backhaul in case of local edge

What are the different private 5G deployment models?

According to 3GPP, there are broadly two models of private 5G deployment

Standalone non-public networks are dedicated on-prem networks with the complete set of network components (5G core, gNB, UPF, CP, UDM, MEC) deployed on-site. They are deployed irrespective of the availability of public networks in a particular geographic area. Private networks are also suitable for remote locations such as mines, offshore oil rigs, and underground sites where the public network cannot supply its capabilities.

Model 1 – Standalone: local frequencies

The layout uses local 5G frequencies. To deploy a standalone NPN without using MNO’s frequencies, the enterprise should be able to avail it directly from the government if the country has such a spectrum policy. In other words, the government has set aside a spectrum quota for private use.

Figure 4 below shows a fully private network using local frequencies, no sharing with the public network:

Model 2: Standalone: licensed frequency

In another case, MNO licensed frequencies can build standalone private networks the standalone networks can also be built using MNO’s licensed frequencies with the architecture remaining in complete control of the enterprise.

Figure 5 below shows a fully private network using the operator’s licensed frequencies, no sharing with the public network:

Hybrid network models – In the case of hybrid models, private networks share some of the components of the network architecture with the public network. Thus, all other components except the shared component are deployed on-site.

UPF, 5GC CP, UDM, and MEC are deployed in the enterprise and physically separated from the public network. Only 5G base stations (gNBs) located within the enterprise are shared between private and public network (RAN Sharing).

Data traffic of the devices belonging to the:

Private slice (a private network), which diverges to the private UPF in the enterprise,
Public slice (public network), which diverges to the UPF in the mobile operator’s edge cloud.

In other words, private network traffic such as in-house device control data, in-house video data, etc., stays in the enterprise only, while public network service traffic such as voice and Internet is transferred to the mobile operator’s network.

Figure 6 below shows a private network built using RAN sharing architecture:

In this model, the dedicated UPF and MEC are along with gnB (base station) in the enterprise while UDMs and CP are in the mobile operator’s edge cloud. Both public and private networks share UDMs and CP.

Figure 7 below shows a private network built using RAN and control plane sharing architecture:

In this model, MEC and UPF exist only in the operator’s edge cloud, while gNB is deployed on-site (in the enterprise). Private and public networks share 5G RAN and Core, which are logically separated. Since the MEC will take place in the edge cloud of the mobile operator, this architecture may experience latency as a major issue

Figure 8 below shows a private network built using RAN and core sharing architecture:

Which model is suitable for your enterprise?

Each model of private network deployment has its pros and cons. Varying needs of the enterprise, its scale and location (single premise or multiple) of operations, its budget, and which attributes of the network are the most important are some of the questions an enterprise must address before making a decision on which model is best suited. Figure 9 shows the various pros and cons of each type:

Key drivers of private 5G

Some of the opportunities and enterprise challenges that are driving this migration are as follows:

Limitations of legacy networks– Existing narrowband mission-critical networks reaching the end of their life cycle. Also, the private networks offer improved coverage, bandwidth, reliable and secure communications.
Concerns for improved security– A need for improved security and control over the enterprise data is amongst CIO’s top priorities.
Data-heavy use cases such as augmented reality (AR), virtual reality (VR), and HD video-based applications are used for remote operations, real-time imaging, monitoring and surveillance.
Continuous investments and innovations in 3GPP technologies

Challenges in building a private 5G network

High Cost- Private networks (LTG or 5G) require huge CAPEX as compared to the Wi-Fi networks. Because of the high cost, only financially capable large enterprises or government authorities for different vertical industries deploy private networks. Comparatively, hybrid models are less cost heavy than fully private networks.
Integration with legacy networks – There are challenges with integrating existing legacy networks with the private 5G network technology
Complexity and Lack of IT expertise – Majority of enterprises lacks the IT expertise required to build and manage private networks, a complex process overall. As a result, enterprises have to outsource managed services. Now, here comes the role of buying flexible (Opex-based) network ‘as-a-service’ subscription, deemed an all-in-one network solution package inclusive of hardware, software, management tools, licenses, and lifecycle services.
Non-availability of the industrial spectrum – To build a standalone network fully independent from the MNOs network and enterprises, those who want complete control over the spectrum need to acquire such a spectrum directly from the regulator. As discussed previously, some countries reserve their spectrum exclusively for enterprise use, non-availability of such a spectrum is one of the challenges in building a private network.
Lack of bespoke network solutions – Due to the fragmented nature of the market,varying enterprise needs amongst different verticals require different network solutions, providing bespoke network architectures is currently not feasible for the suppliers