A lot of the attention in 5G development has been focused, rightly, on 5G New Radio (NR). But to offer 5G services a new system architecture and core network is also needed.

Defining the next-generation architecture is the responsibility of the 3GPP's System Architecture (SA) Technical Specification Group on Service and System Aspects. The study phase, completed last year, outlines what this new core network -- known as NG Core -- will look like and provides a guide to specification in the normative phase, now underway in 3GPP Release 15.

For more detailed information, I've written a white paper on this topic, Designing Cloud-Native 5G Core Networks. It's available for download HERE.

The key study document is "Architecture for Next Generation System" (TR 23.799), which sets out the requirements for NG Core. In its conclusions TR 23.799 identifies two representations of the next-gen architecture: a point-to-point architecture and a service-based architecture.

These architectures are now being developed further in the "System Architecture for the 5G System" specification group known as TS 23.501. Both contain the same functional elements (more or less) and support the same N1, N2, N3, N4 and N6 interfaces.

The point-to-point architecture is shown below. As with prior generations, it specifies discrete interfaces between control-plane elements. This is probably the model operators with aggressive deployment schedules will use -- at least initially, perhaps before migrating to the service-based model at later date.

Figure 1: Point-to-Point Next-Gen Reference Architecture Source: 3GPP TR 23.501, January 2017, Figure 4.2.3-2

The main difference in the service-based architecture is in the control plane where, instead of predefined interfaces between elements, a services model is used in which components query an NF Repository Function (NRF) to discover and communicate with each other over APIs. This is closer to the cloud native networking concept and attractive to operators that want flexibility and adaptability. The challenge is that it is harder to implement using today's cloud platforms and, at this stage, appears likely to be part of Phase 2 deployments.

Figure 2: Next-Gen Service-Based Architecture for NG Core Source: 3GPP TR 23.501, January 2017, Figure 4.2.3-1

Both the options above are known as 5G "standalone" (SA) mode -- i.e. they allow 5G NR to be deployed without dependencies on the LTE network (although in practice integration will be required for mobility, etc.). There is also an option for operators that want to move very quickly to deploy 5G NR to use the existing 4G Evolved Packet Core. This is known as non-standalone (NSA) mode. In this model 5G NR provides extra user plane capacity, but the LTE radio is used as the primary cell and for control-plane messaging (i.e. to set up and manage sessions and mobility). You can read more about this in the "RAN Architecture and Interfaces" report (TR 38.801).

There are quite a few moving parts and the route operators select will be dependent on a range of factors related to timing, infrastructure cycles, market positioning, and so on. One way that can help transition been these options is to deploy "cloud native" core networks for both 4G and 5G.

There's a lot more to say on this topic and so, together with Nokia, on the March 23, 2017 Light Reading will host a specialist 5G Core webinar: Cloud-Native Architectures Will Be Key to 5G NG Core. Free registration is HERE. It would be terrific if you could join us.

— Gabriel Brown, Principal Analyst, Heavy Reading

This blog is sponsored by Nokia.