Revisiting PMN

Private Mobile Network (PMN) adoption is accelerating after a slow start in the last couple of years. A successful PMN implementation that meets your requirements across geographies requires selecting the right architecture, appropriate spectrum, right partners as well as right sourcing model.

Summary

Any successful PMN initiative needs to understand the current and future requirements of applications and services that will be delivered over the PMN. Deciding on the most appropriate architecture will have significant impact on the TCO (Total Cost of Ownership). Similarly, decisions on spectrum to use will have significant impact on both TCO as well as service levels the PMN can deliver. Identifying and selecting the right partners to deliver technology and services for the project will be key to a timely and successful implementation. There are basically two ways of sourcing a PMN; as a managed service with predominantly a monthly recurring cost or self-managed with a high initial investment in equipment and a lower monthly recurring cost to operate the PMN. It’s paramount to understand that different spectrum has different propagation (reach/behavior) and that unlicensed spectrum (e g CBRS) isn’t universally available in all countries when piloting a blueprint PMN to be rolled out in multiple countries.

Market Outlook for PMN

The adoption of PMN has been slow over the past couple of years and failed to meet market expectation. Even though there are many advantages with PMN compared to traditional Wi-Fi or wired networks in a number of use cases. The most obvious use cases being hard to wire scenarios like mines, refineries, oil platforms, ports and airports. In 2025, it seems like more clients are willing to go from single country pilots to multi country deployments.

In the 2025 Gartner? Magic Quadrant? for 4G and 5G Private Mobile Network Services, published on January 6, 2025, Gartner states “the market for network equipment and services for 4G and 5G PMN for the top five industries will reach $45.3 billion by 2033, up from $2.2 billion in 2023.”

A PMN can come in many different architectures

In its most basic, and typically, most cost-effective form, a PMN can consist of buying transport on an existing public mobile network from an established Mobile Network Operator (MNO) with a service level agreement. In a 5G network this would most likely mean buying a slice. In its most advanced form, it would mean having a full stack of all necessary operational systems on site to support fully autonomous operation. The most common architectural options, as defined by 5G ACIA (5G Alliance for Connected Industries and Automation) include (but there may be in between options from the provider):

0.???? The PMN is delivered as a VPN over an existing MNO network using a dedicated APN (Access Point Name) with no site-specific additions.

1.???? The PMN is delivered as a VPN over an MNO network with site specific operator additions (site specific antennas/base stations). This is often also referred to as PNI-NPN (Public Network Integration – Non-Public Network) using E2E Network Slicing.

2.???? The PMN is delivered with site specific operator additions (site specific antennas/base stations), shared core, RAN sharing and local break out of data. This is often also referred to as PNI-NPN using RAN and Control Sharing.

3.???? The PMN is delivered with site specific operator additions (site specific antennas/base stations), local core, RAN sharing and local break out of data. This is often also referred to as PNI-NPN using RAN Sharing.

4.???? The PMN is delivered with a full stack of on-premise servers/systems to be able to operate as an autonomous network. This is often also referred to as SNPN (Standalone NPN).

A rule of thumb is that if option 1 above cost 1X then the cost of option four above is typically 8-10X. Hence, it’s important to understand current and foreseeable requirements on the PMN to make an informed architectural decision.

What are the differences between different spectrum bands on a high level?

When there is a choice of what spectrum to use, it’s basically a compromise between coverage and speed (as well as to some extent also latency and high number of endpoints). The high level and not so scientific explanation is:

• The higher the spectrum band, the shorter the reach is. Lower spectrum bands also permeate walls better than higher spectrum bands. Very high spectrum bands require line of sight to work properly.
• On the other hand, higher spectrum bands are usually available in wider channels which enables higher thruput. As each base station is handling a relatively smaller area, it also supports a denser implementation of endpoints.

This really means that you want different spectrum for different use-cases. E g:

• When planning to cover a big area like a large open-pit (overground) mine, sea or airport cost effectively, you prefer lower spectrum, typically low bands in the range 600-900 MHz to minimize the number of base stations needed. Typical data speed is 30-75 Mbps.
• When planning to cover a medium size building, e g a factory where the use case calls for low latency and high reliability or an arena full with people, you prefer spectrum in high bands, 7 GHz and above, typically bands in the range 24 GHz to 40 GHz. Typical data speed is 1 Gbps or higher.
• Mid bands use spectrum in the range 1 GHz-2.6 GHz and 3.5 GHz - 6 GHz which is optimum for covering campuses consisting of multiple buildings extending to several blocks in size. Typical data speed is 100-250 Mbps. C-Band, that is frequently referenced, resides in 3.7 GHz - 4.2 GHz in mid band and CBRS (available in the US) resides in 3.55 GHz – 3.7 GHz (LTE Band 48 / 5G band n78).

The challenge when dealing with global roll outs is that typically only MNOs (and equipment manufacturers working closely with MNOs) can provide the same (or close to same) spectrum across geographies.

When looking at license free spectrum, which is currently only available in about 23 countries for mid band and 16 countries for high band, whereof 12 of these countries also provide mid band, it’s only about 27 countries (up from 22 countries in late 2022) in total that currently provides some kind of license free spectrum. This number is expected to continue to increase over time as more and more regional spectrum harmonization efforts are under way. The number of countries and the specific countries may not be sufficient for many global projects for quite some time, making license free spectrum a non-starter to use. The license free spectrum allocations vary depending on region:

• In Europe there is a tendency to provide LTE Band 43 / 5G Band n78 (except the UK that uses different bands altogether) in the mid band and the amount of spectrum varies from 40 MHz to 240 MHz and typically in the range 3.4 GHz to 3.8 GHz. For high band, it’s 5G band n258 and the amount of spectrum varies from 400 MHz to 850 MHz and typically in the range 24.25 GHz to 27.5 GHz.
• In APAC there is a tendency to provide 5G Band n79 (Japan also offers spectrum in LTE band B41) in the mid band and the amount of spectrum varies from 100 MHz to 300 ?MHz and typically in the range 4.6 GHz to 4.9 GHz. For high band, it’s 5G band n257 and n261 and the amount of spectrum varies from 400 MHz to 900 MHz and typically in the range 27.95 GHz to 29.5 GHz.
• In the US, the allocated spectrum, CBRS, is in LTE band 48 / 5G band n78 and resides in mid band, 3.55 GHz – 3.7 GHz (150 MHz).

The mid band spectrum in Europe and US is fairly comparable while the mid band spectrum in APAC is in significantly higher spectrum, about 1 GHz higher than Europe and US which means propagation will likely be somewhat different. When doing research for this article, I ran across this Ericsson Whitepaper that I found very helpful, detailed and updated in April 2024..

Who can provide a PMN

The typical providers of a PMN include MNOs, Managed Service Providers (MSPs), Systems Integrators (SIs), Equipment manufacturers like Ericsson and Nokia and in a limited way Hyperscalers like AWS on a regional basis

Don’t forget about LTE

Even if the primary initial use case for investing in a PMN works fine on 5G alone and supports the target spectrum band, consider secondary use cases that may benefit from LTE as well at a low additional investment level.

PMN as a Managed Service or Self Managed

Most organizations don’t really want to become a mobile operator and having to master all the associated skills including radio planning and operating a Network Operations Center (NOC). In fact, many MNOs outsource the operation of their own network to managed service providers like Ericsson and Nokia. Contracting a PMN as a managed service makes a lot of sense in most situations.

Recommendations

Start by trying to identify all the use cases that will drive the PMN requirements. Ensure the prospective vendors can show a roadmap for how their technology can be extended in phases to support the full scope and needed architecture of the PMN. Keep in mind that PMN technology is evolving quickly and that some requirements may be dependent on a future 5G release. Architect for modularity to insulate as much as possible against lacking vendor long term viability and ensure the selected vendor(s) can support across the entire target footprint and provide access to required spectrum.

It may look attractive for organizations with a CapEx heavy tradition to buy and operate their own PMN infrastructure. For most organizations, it’s usually a better and lower risk option to buy PMN as a managed service from an MNO or MSP with strong operator relationships.

GARTNER is a registered trademark and service mark of Gartner, Inc. and/or its affiliates in the U.S. and internationally and is used herein with permission. All rights reserved.

My first post on this topic was published on November 7, 2022: https://www.dhirubhai.net/pulse/5g-pmn-private-mobile-network-its-all-spectrum-leif-olof-wallin/