The Case for Non-Public Networks for Enterprise Connectivity

As enterprise 5G continues to mature in 2021, private network deployments have received particular attention within the industry. To fulfill implementers’ requirements for the maximum degree of network customizability as well as full control over Quality of Service (QoS) and Service Level Agreements (SLA), the telco industry so far has rested all its hopes on providing fully isolated private networks, with all network infrastructure deployed exclusively on the implementing enterprise’s premise.

As enterprise 5G use cases mature, however, a new notion of Public Network Integrated Non-Public Networks (PNI-NPN) emerges, leveraging public and private network resources. Leading service providers like Deutsche Telekom, for example, have already launched commercial offerings that share network infrastructure between public and private networks.

On the one hand, these deployments will guarantee the maximum possible degree of network integrity by deploying dedicated equipment on the respective enterprise premises. On the other hand, it has the capability to reduce deployment costs (by leveraging already existing public network resources with dedicated on-premises equipment) as well as minimize the enterprise size space that is needed for dedicated enterprise cellular network deployment. Therefore, careful and sensible combination of public and private cloud resources will have the capability to increase the total addressable market for cellular connectivity beyond the large enterprises that are in a financial position to deploy a fully isolated private network on their premises. As this new notion of NPN is shaking up the private networks market, the question is what flavor of non-private networks will prevail, i.e., what network functions will be deployed on-premises and what network functions will utilize already existing resources.

Looking at the basic components of a cellular network (RAN, transport, and core network), there are two different ways to combine public and private network resources. First, a dedicated enterprise network could share the RAN with the public network, while all other functions (including the core network) would be deployed on-premises. The network traffic would be separated within the RAN, with the private enterprise traffic being routed towards the on-premises edge or core deployment, while any consumer networking data (such as employees using their mobile phone for Internet browsing) would be routed towards the public network.

While this compromises on the full isolated of the private network, it also carries several advantages both from a use cases and performance as well as from a cost point of view. First, using public RAN infrastructure will minimize the amount of network infrastructure necessary for private network deployment, which will drive down the deployment costs considerably. Second, an intersection with the public network on the RAN level might also benefit the performance of the enterprise network. On the one hand, it will allow routing certain non-critical network traffic that is generated on enterprise premises (e.g., employees browsing the Internet) into the public network, thereby reducing the processing load on the on-premises private core. On the other hand, it also eases the realization enterprises that require intersection with the public network. In a logistics environment, for example, the intersection with a public operator network will be needed to guarantee full traceability of goods along the entire supply chain. While in the case of a fully isolated private network this would require additional infrastructure to be deployed, in a RAN sharing scenario, data can be handed over to the public network more easily because RAN equipment is shared either way.

Second, the enterprise network could also share the RAN as well as the core with the public network. Particularly, this mostly applies to the control plane because most enterprises would not want to compromise on network integrity as much to move the user plane to an off-premises location. Hence, the on-premises connectivity infrastructure would be sharing RAN as well as control plane with the public network. While this reduces deployment costs for an enterprise network even further, it also allows enterprises to hand over control of network operation and functionality and can therefore be particularly appealing to small- and medium-sized enterprises without the capabilities to manage and operate the network on their own. Integrity of the networking data could be guaranteed by deploying an on-premises edge that could host the data plane.

The emergence of this new notion of PNI-NPNs means that the telco industry needs to rethink its approach to enterprise connectivity.  At the heart of it, connectivity suppliers need to depart from a dichotomous view, in which the only two options for enterprise connectivity are either a fully on-premise deployed private or the provision of a dedicated slice from a public network. Instead, they should focus their efforts on finding solutions to provide the different flavors of NPN, as outlined above.

As the emergence of PNI-NPN increases the value of existing public network resources, it also changes the market dynamics for the provision of enterprise connectivity. While for isolated private network deployments carrier-owned public network resources are less important, they will provide additional value for NPN deployments and hence will increase traditional Communication Service Providers (CSPs)’ value proposition for enterprise cellular connectivity. While this does not automatically put CSPs back into the driving seat for enterprise cellular deployments (especially given the increased importance of cloud-based deployments, which will require a different set of expertise that traditional CSPs might not have), it increases their attractiveness to partnerships and co-creation initiatives between hyperscalers and CSPs. While CSPs bring thorough expertise in network operation as well as extensive public infrastructure through their wide-area macro networks, hyperscalers will add necessary cloud resources as well as expertise in cloud management and critical payload orchestration capabilities between public and private cloud resources.