According to the GSA (Global Mobile Suppliers Association), there are now 176 service providers who have deployed 5G technology on a commercial basis. Of those networks, 145 are offering mobile 5G services and 51 are offering 5G FWA (Fixed Wireless Access) services. Here’s the interesting part though; as of March 2021, only seven SA (Standalone) 5G networks have been commercially launched.
This means that the vast majority of 5G deployments today are NSA (Non Standalone), relying on the 4G network, both in the E-UTRAN (Evolved – Universal Terrestrial Radio Access Network) and the EPC (Evolved Packet Core). To give an idea on where 5G is heading, there are 807 commercial LTE networks currently launched, according to GSA statistics. It’s no surprise that from a training perspective, the vast majority of our training delivery at the moment is related to Standalone 5G, since that’s what engineers are dealing with right now in preparation for commercial deployment. As such, I thought it would be worthwhile spending some time to provide a bit of a comparison between the two 5G deployment models; not so much an opinion piece for this blog, but more of a reference guide.
To begin with, Figure 1 outlines the 5G Non Standalone Deployment model.
With the NSA 5G approach, correctly termed EN-DC (E-UTRA New Radio – Dual Connectivity), the service provider can quickly bring a 5G mobile broadband service to market (at least more quickly than Standalone). Non Standalone 5G capable devices will first establish connectivity to the 4G network and if 5G is also available, 5G connectivity will also be put in place. Note that unless 4G is available first, the device will never access 5G in this configuration.
Ultimately, the subscriber has the advantage of using the 4G and 5G radio simultaneously to exchange data with the network, hence increasing data throughput and providing an eMBB (enhanced Mobile Broadband) experience for the customer. The downside of this approach is that because this service is still basically 4G end to end, other 5G features such as URLLC (Ultra Reliable and Low Latency Communication) and MMTC (Massive Machine Type Communication) are not available.
So in this Non Standalone configuration, how does user data flow over the network? Well, the two most common examples of the data that needs to be supported are, of course, Internet traffic but also, for the 228 service providers that have deployed it, VoLTE (Voice over LTE). Figure 2 shows the configuration that is generally used, in which VoLTE traffic remains on 4G, whereas Internet traffic is exchanged with the EPC via 5G gNB. In turn, the gNB is responsible for managing the traffic exchange over the radio; some traffic will be exchanged over 5G directly, whereas some traffic will be sent to the eNB to be exchanged over 4G. This configuration has been historically referred to as Option 3x.
When a Standalone 5G network is deployed, the end to end architecture changes completely. In particular, Standalone 5G incorporates the 5G NG-RAN (Next Generation – Radio Access Network), coupled with the 5GC (5G Core). A multitude of new network functions have been included within the 5GC, each of which performing quite specific roles (where in 4G, a given network node will have numerous roles to fulfil). This is not too much of an issue, since all of the control plane of the core network is virtualized, with network functions operating as cloud native software processes running on a shared virtual infrastructure, managed by Kubernetes. Figure 3 outlines the high level architecture for Standalone 5G, drawn in our NetX style.
The various components outlined in Figure 3 are explained as follows but note that this is by no means a complete list! You can see a more complete picture in Figure 4.
The primary task of the AMF is to manage the mobility of the subscriber. In particular, the AMF will play a key role in the device registration process (including security), as well as track the device’s mobility for reachability and paging purposes.
The SMF manages the PDU Sessions (the user data sessions) associated with an individual subscriber. As such, the SMF will routinely interact with the PCF to determine exactly which Data Networks the device is allowed to connect to, as well as the QoS profile it can expect to be allocated. The SMF will also liaise with the UPF to establish PDU Session connectivity. Finally, the SMF will be responsible for allocating a suitable IP address to the device (IPv4 or IPv6), assuming an IP PDU Session is active.
The UPF is the only network element within the core that is involved with user plane traffic, having data plane connectivity to both the NG-RAN and also the Data Network e.g. the Internet. As such, the UPF is responsible for ensuring data is placed on the correct downlink data flow, as well as ensuring that any dynamic policy rules are suitably enforced (policy rules are provided by the PCF via the SMF). Moreover, as handovers take place in the NG-RAN, the UPF will remain the core network anchor point for the user plane traffic.
The UDM is essentially a central repository of subscriber information which can be used by several different network elements. Information accessible at the UDM includes access restrictions, mobility restrictions, Data Network QoS profiles and security keys associated with subscribers.
The PCF can provide policy decisions to the AMF and SMF on a dynamic basis. These policy decisions or often based on conditions being active within the network, such as congestion, subscriber geolocation or billing. In addition, the Data Network can also potentially provide session level information, such as the subscriber wishing to make a call or view a video. In all of these scenarios, the PCF can dynamically change the way in which the subscriber receives their service, from establishing the correct QoS in the network to completely terminating a PDU Session.
The NEF serves as the entry point into the 5G Core network for CIoT (Cellular IoT) related application servers. Using the NEF, these application servers can interrogate the 5G core network to support device monitoring, provisioning and policy control. The NEF however does not simply serve as a gateway for device control; data can be delivered to end devices via the NEF instead of using the UPF.
From a deployment perspective, with the exception of the UPF all of the highlighted 5GC functions are considered to be within the SBA (Service Based Architecture). This essentially means that these network functions offer services to one another, which are accessible via APIs (Application Programming Interfaces) standardized by the 3GPP. In reality, pretty much all of the control traffic flowing over 5G Core networks will be HTTP/2 based, with JSON payloads being carried when appropriate; the likes of GTPv2-C and Diameter are only used in specific interworking scenarios.
Figure 4 outlines the SBA concept, including the SBI (Service Based Interface) designations for each network function, which is the correct term for the aforementioned APIs.
Where 2018 to 2020 saw a real surge in Non Standalone 5G, 2021 is set to see a significant rise in Standalone 5G. Despite the current low numbers, there’s still plenty of service providers who are imminently about to launch Standalone networks. Yes, NSA will obviously continue to be deployed but hopefully, soon enough we’ll start to see what 5G is really capable of…