Private networks are expected to be one of the primary ways in which 5G services are monetized. A recent IDC forecast estimated that on a worldwide basis, private LTE/5G wireless infrastructure revenues will reach $8.3 billion by 2026, compared to $1.7 billion in 2021 and reflecting a compound annual growth rate of nearly 36%. In a session at the virtual Test and Measurement Forum conference, Adnan Khan, director of technology and market development with Anritsu, outlined the basics of private networks, as recapped here.
What is a private network?
Private networks have been established within 3GPP standards as “non-public networks” or NPNs, which is designed to be deployed by enterprises, school districts or other entities to provide their own wireless service for their own purposes, explained Khan. Private networks can only be accessed by those who have the right device and permissions/configurations.
Who is interested in deploying private networks?
Khan cited research from the Global mobile Suppliers Association (GSA) that the highest level of private network interest and adoption currently is by manufacturing companies, followed by the mining sector, educational institutions and power or water utilities. Khan noted that the current mix of verticals that are most interested in private networks is likely to change over time—Anritsu has already seen the Covid-19 pandemic drive higher interest in private networks from healthcare providers, for instance. The key challenge for private network adoption currently, he says, is the kind of devices that are currently available.
Why are private networks needed?
The primary reason is for an enterprise or organization to have control over its network, Khan said, including being able to control and improve its coverage and performance, as well as to have control over its security measures.
What are different ways to deploy 5G private networks?
Khan described four deployment models: A “standalone” private network deployment, which is a fully isolated network that meets an enterprise’s secuirty and performance requirements but is interconnected with a public network, such that devices primarily operate on the private network but can operate on a public cellular network as well. Another model would be deploying a private network with a shared Radio Access Network with a public network operator or spectrum provider, while the network core and compute resources are those of the private network customer. A third option would be to deploy such that user-plane data is confined to the private network, but the spectrum and the control-plane data is shared with a public network; and a fourth option is a network slice of a public network.
How is private 5G different from the typical carrier-deployed 5G networks?
Private 5G networks will naturally be a much smaller scale than public carrier networks. While the overall framework is similar, Khan pointed out that the types of devices that need to be supported on private networks may be far more diverse and with very industry-specific demands, than the consumer device base dominated by smartphones: Robots, sensors, actuators, machinery, in sizes and RF characteristics. Private 5G networks may operate in licensed, shared or unlicensed spectrum, but will also require spectrum planning and the integration of various elements. Khan said that private networks, particularly those where an enterprise may have dozens or hundreds of sensors on, say, a piece of machinery, may have a significantly higher density of connected devices than a typical public network.
What are some of the testing needs of private 5G networks?
Khan outlined several types of testing that come into play in private 5G networks: spectrum mapping, in order to have a clear understanding of the RF environment that the network will be operating in; testing of end-to-end operations in order to validate and understand network performance, including latency; service assurance; and device testing. Prior to deployment, Khan said, devices might be tested for functionality, but they should also be load/stress tested. The potentially high density of devices may affect their operation, and in a smart factory situation, he pointed out, devices may be running constantly for hours—functionality-focused testing may not capture how well a device will handle that level of stress.
Useful testing KPIs for private networks, Khan said, include RF quality and coverage; throughput; latency; service availability; service reliability; security testing; mobility testing; and location accuracy testing. Depending on the device types and the services running on the network, he pointed out, the needs of sensors versus virtual-reality goggles will help determine the specific parameters.
What role might Open RAN play in private networks?
Khan said that while Open RAN is still evolving, he believes it will be a “very big play” in private networks because one of its focus on driving network infrastructure costs down as well as cutting out vendor lock-in and offering the flexibility of multi-vendor deployments. He thinks that Open RAN may come into play in “cost-conscious” private networks, as well as those that are “security conscious” where, say, a company wants a highly secure and isolated network deployed in a particular location. Still, he added, there is much more testing of Open RAN that needs to be done, particularly on the interoperability aspect—but even that may prove to be more doable in a private network setting, he said, where devices only stay on that network and interoperability with legacy network elements or with other, public cellular networks isn’t needed.