The 5G core will be essential to an effective 5G business model that serves multiple enterprise sectors, rather than being constrained by the MNOs’ more usual consumer-centric approach. Only with 5G NR Standalone, and the full cloud-native, sliceable core, will operators be able to exploit the full potential of their new platforms and address high value, high growth segments in a profitable way, rather than being confined to low margin consumer markets or just one or two chosen verticals.
This means that deployment of 5G Standalone (SA) – which has a full 5G core rather than relying on an enhanced LTE core like 5G NSA – is being accelerated, by many MNOs, rather than delayed by the Covid-19 pandemic. A recent survey of 45 5G operators by Wireless Watch found that over half of them expected to roll out a 5G core at least six months earlier than they had planned before the pandemic, because there is an intensified need to differentiate their services, and support enterprise programs to recover from the crisis.
There are both aggressive and defensive drivers at work. Some MNOs actively believe they can steal a march on competitors, old and new, by deploying a 5G core early, and being able to support differentiated experiences for a wide range of industries, especially by harnessing slicing. Others are more defensive, feeling they must adopt the core because, otherwise, enterprises and their private network partners will steal the cream from the top of the 5G milk.
The defensive operators – which have not fully worked out their own 5G case for industrial services, but are scared of new competition – will be even more spooked by Ericsson’s latest bid for 5G leadership. The vendor has announced 5G SA trial kits specifically designed for enterprises to try out 5G in a low risk way.
Although these are available through telcos – the launch partner is Deutsche Telekom – it is a very small step to providing them to alternative service providers. Even with an MNO involved, such kits take some of the power and mystique away from the operator. And they would make it easier for a less experienced mobile provider to brush up its skills, should Ericsson row back on its current commitment to target enterprise and private 5G only through MNO channels.
So an acceleration of 5G core deployment from late 2020 onwards is on the cards, although commercial 5G SA will still be the preserve of a relatively small group of operators until 2022 at least, because of the risks and complexities of the migration, not just to a new core with advanced capabilities like slicing, but to a cloud-native platform, and in many cases, to integration with other new telco technologies like AI/ML or edge computing.
But that early 5G SA elite group will be larger than we previously predicted, and there will be more intense activity among other MNOs at least to trial the new core, to lay foundations in terms of cloud-based infrastructure and skills, or to deploy the new technology in secondary or private network cores in order to gain expertise and confidence.
The past week alone has seen announcements from SK Telecom, China Telecom, Vodafone, Deutsche Telekom and others about their 5G SA roadmaps, though few will have been under the pressure of a regulator’s mandate to roll out a 5G core. In Singapore, the Infocomm Media Development Authority (IMDA) made 5G SA deployment a condition of winning one of the main 5G licences, which it auctioned in April. Another regulator which went beyond the usual ‘technology neutral’ stance was France’s ARCEP, which has made support for network slicing a condition of operators gaining the maximum possible amount of spectrum in the forthcoming auction – though it is not clear yet whether it will mandate full 5G slicing enabled by SA.
The winners in the Singapore auction were Singtel and a joint venture (JVco) between the second and third MNOs, Starhub and M1. By only offering two nationwide licences, IMDA excluded the fourth player, TPG, but the SA requirement meant it was not all plain sailing for the established operators. Singtel and JVCo have each been assigned 100 MHz of 3.5 GHz spectrum, while Singtel, StarHub and M1 each gained their hands on 800 MHz of millimeter wave spectrum (band unspecified).
The operators have protested against the requirement to deploy 5G SA from day one, saying the technology is not yet sufficiently mature, but they will introduce the capability from January 2021 nonetheless. Singtel is working with Ericsson and JVco with Nokia.
In South Korea, SK Telecom has been trialling Intel Xeon Scalable processors and Ethernet 800 Series network adaptors as potential enablers of its Samsung 5G core, which it aims to roll out from late 2020 or early 2021. Intel said the project began as an R&D exercise but has been successful enough that it will be commercially ready this year.
The partners have developed a 5G SA user plane function (UPF) based on the 3GPP’s Release 15.
“Aside from the specification, the R&D focused on the software and hardware architecture, and optimizing for reducing the packet processing latency,” said Park Jong-kwan, VP and head of the 5GX technology group at SK Telecom. He said the trials had
demonstrated up to 78% lower latency by using intelligent packet classification and steering.
Release 16 creates a migration path to containerizing all network functions, and that could require more R&D with Intel on control plane optimization, Park added. “Our initial challenge was to make the low latency 5G core, the UPF with high efficiency,” he said. “But with the 5G core becoming more cloud-native … more collaborations are to be engaged with Intel.”
Alex Quach, general manager of Intel’s wireline and core network division, told SDxCentral: “In this case we were able to look at taking and prioritizing, and directing enhanced mobile broadband traffic, separating that prioritization from the IoT traffic, or ultra-reliable low-latency traffic.” That traffic is directed to the network interface card (NIC) instead of a CPU so that it can achieve, at a constant CPU utilization rate, lower latency in the network.
And in China, China Telecom’s CEO Ruiwen Ke told the recent GSMA Thrive virtual conference that his company “is now working rapidly to develop new innovations in 5G SA for cloud and network integration so that we can build secure, reliable products on the cloud and network”, as part of a roadmap for slicing, edge computing and AI.
Ruiwen also said China Telecom is developing guidelines and standards that other operators can follow in their 5G SA deployments. “China Telecom will be the coordinator for the global industry to draft the 5G deployment standard. This is the first global guideline for 5G SA deployment,” he said. 5G SA networks are expected to start rolling out in China late this year. China Mobile had originally planned to be the only non-greenfield MNO to go to SA from day one, but in 2019, decided to accelerate deployment by using NSA for its first phase.
In Europe, Deutsche Telekom has become a major supporter of open networks for 5G, conducting trials with multiple partners with technologies such as O-RAN. Its latest is with Mavenir, which has provided its containerized 5G core for tests on the German operator’s cloud network infrastructure.
Mavenir said the test validated “combo nodes” which enable voice call handover from 5G to 4G, and also focused on automation, on-boarding, cloud-native, CI/CD, and on performance requirements for resiliency, fault tolerance and performance.