The paper outlines four possible deployment models for bringing 5G into enterprise networks, as well as the key considerations for choosing each one, such as the nature of the application, latency in the core and RAN interfaces, and the location and manageability of services. The four models are as follows:
- On-premises core network and application services – Data sovereignty, site resiliency, and application latency requirements are ensured by keeping all traffic on-prem. Access to conventional enterprise cloud-based applications is enabled, subject to normal limitations around resiliency and latency.
- On-Premises user plane and application services – The paper outlines several reasons to move the control plane to the cloud, such as the need for control plane aggregation in a multi-site 5G core network deployment. All other 5G elements and the application services are on-premises, except the 5G control plane elements.
- Cloud-based core network and application services – User plane traffic from 5G devices will always have to enter the cloud. In such deployment models, it may be possible to move the 5G core network and user plane elements to the cloud where the applications services are located.
- The hybrid model – There are some application services in the cloud, and some are on-prem. To support such a model, there can be two different Data Network Names (DNN’s), one for supporting applications that are on-premises and another for supporting applications in the cloud.
