Most private cellular networks will connect people rather than things or industrial equipment and so can be fully served by LTE rather than 5G at a far lower cost, according to California-based electronics test and measurement company Keysight Technologies.
Speaking at the recent industrial technology exhibition, Hannover Messe in Germany, Keysight’s general manager of private and dedicated networks, Jagadeesh Dantuluri, argued that about 90% of use cases for private networks were being driven by desire for reliable and secure connectivity around a site or campus for devices such as tablets and IoT wearables, to support a variety of needs that cannot be met with existing WiFi networks. This may be because wider coverage is required, higher levels of resilience, or greater security. LTE is also good enough for the majority of dedicated wireless communication needs, such as connecting security cameras, which tend not to generate higher definition video.
Dantuluri felt moved to deliver this reality check because he believed some enterprises were being lured into premature investment in 5G when they could save a lot of money by deploying 4G instead, which was in any case more mature. He noted that integration with legacy systems, as well as cost, were the main challenges inhibiting widespread adoption of private cellular, and for many these will be better overcome by starting with 4G.
Dantuli also touched on the key question of how private cellular will coexist with WiFi networks already in place, looking at where it will be complementary, where it will be required for new applications, and where it may take over existing ones. In many cases, enterprises will in the immediate future face a three-way choice for given use cases between WiFi as that progresses to generations 6/6E and 7, private LTE and private 5G, with the issue of coexistence between them. Dantuli did admit that even many enterprises for which LTE was right now would require 5G in future, implying that it was wise to ensure that technology or suppliers chosen offered a clear migration path.
As he put it: “Just because 5G has come doesn’t mean people will leave everything and come to 5G. It takes time, effort and money to deploy private 5G, when most companies already have a combination of WiFi and wired connectivity that works. All those systems will not go away, even with 5G. They will coexist.”
He also noted that private 5G remained hampered by a lack of industrial-grade 5G chipsets and devices, as well as cellular system familiarity and expertise in enterprise IT departments. There were also logistical or practical questions, such as which party will guarantee the potentially stringent service level agreements (SLAs) that private networks will require for the more advanced use cases 5G is supposed to meet.
He cited the case of Germany, where some enterprises have spurned operators and established their own private networks in spectrum dedicated to them. In that case the enterprise might consider the provider of equipment such as Nokia or Ericsson, or some third party systems integrator, or even a cloud provider offering networks-as-a-service, as guarantor of the SLA.
These are questions still to be resolved, although are being addressed by these larger enterprises going their own way. For smaller enterprises, private 5G may be a step too far at this stage, and certainly might require the help of an operator. For many verticals, as Dantuluri argued, private 5G is still unproven.
It is worth taking stock here. On the one hand, as Dantuluri admitted, there are those 10% of cases that will require private 5G soon if not now, often because of the need for ultra-low latency as well as high capacity, such as automated mobile robot (AMR) on factory floors, which will continuously transmit and receive large amounts of data in real time or near-real time. But over time this 10% will grow, remembering that even consumer applications increasingly demand the high capacity and bit-rate only 5G can deliver wirelessly, even if they do not need the ultra-low latency. So that 5G migration path is critical.
When it comes to decisions between those three options of WiFi, LTE and 5G, it makes sense to match the requirements with the capabilities and constraints of all three, which include cost, complexity and legacy, as well as technical factors such as bit-rate and latency. We can identify eight broad metrics – cost, complexity, legacy integration, security, mobility/range, latency, QoS/bit-rate, and reliability. On cost and complexity, WiFi and LTE score best, with 5G lagging at present. On legacy integration, WiFi is best, followed by LTE and then 5G, but for security the order is reversed.
In fact, security is one of the strongest cards for private cellular, because it invokes the SIM card approach based on possession of a device, eliminating the threat of credential leakage and ensuring only whitelisted services can access a network. WiFi uses open spectrum that relies on password access and has been hacked many times, with even the forthcoming WiFi 7 only offering limited improvements and reliant on implementation feature such as monitoring and defense in depth, as well as user actions like passwords that are harder to guess, rather than in-built technical capabilities. So even WiFi 7 is unlikely to be deployed for security-sensitive applications.
5G, on the other hand, improves on LTE in several ways, with additional ‘secure by design’ features such as mandates of encryption for inter- and intra-network traffic, as well as in-built scrambling or deletion of data in transit when interception has been detected. It is true that 5G networks generally will be targeted more than previous wireless networks and it remains to be seen whether the additional protections will counteract the expanded attack surface.
Cellular also scores for mobility/range, particularly off-site, or around a large campus. There is also scope for integration with public mobile networks, which will be decisive for some enterprises. While WiFi has traditionally scored for indoor coverage, that advantage is eroded or eliminated by private cellular with small cells. Furthermore, cellular supports transparent handover between cells, with no impact on users, which is not normally the case for WiFi, where devices have traditionally been responsible for their own roaming between access points (APs).
That WiFi roaming deficiency is being addressed by the Wireless Broadband Alliance (WBA) Open Roaming initiative, designed to create a federation of networks between which users can move without having continually to log in. One of the first deployments came in May 2021 when the WBA announced that travelers, employees and retailers at Brazil’s São Paulo/Guarulhos International Airport (GRU) could access the world’s first at-airport WiFi 6 network using Open Roaming, without having to take any action if they had this installed on their smartphones.
Developed by Cisco, Open Roaming was transferred to the WBA in recognition of the need to establish it as a neutral WiFi standard. It goes some way to addressing security deficiencies as well, with protection for user privacy and compliance with the European Union GDPR policies when roaming between WiFi networks. However, it still lacks that token-based protection afforded by SIM cards and also requires interaction with cellular networks as well to enable true global roaming.
Then, on the latency front, there is a clear graduation between WiFi at the bottom and 5G at the top. Currently, WiFi often struggles to achieve latencies below 100ms despite being local, and often slips to 500ms in the event of congestion or interference. This is improving with successive generations, with WiFi 7 targeting the same 1ms latency as 5G. But it will still lag behind because it will lack the deterministic capability of 5G, so that it will not be able to guarantee ultra low latency, sometimes moving out to around 10ms.
Nevertheless, WiFi 7 will close the performance and capacity gap on LTE through several stratagems, notably multilink operation (MLO), which allows single sessions to span multiple channels simultaneously, achieving aggregate data rates up to 30Gbps in theory, over three times WiFi 6. It also supports higher coordination between APs to boost speeds and cut latency.
There is a novel feature to improve performance under interference, which has been an increasing bugbear for WiFi as usage has grown, despite escape into new frequency bands, starting with 5 GHz and then 6 GHz. Qualcomm has christened this interference mitigation ‘preamble puncturing’, describing the new process of isolating the area of spectrum subject to interference and allowing communication both above and below it. Until now, in the event of interference in the desired channel, devices could only use spectrum up to the frequency at which that interference was occurring, and not above it.
The other two of those eight metrics for comparison we identified are QoS/bit-rate, and reliability. WiFi 6 goes some way to closing the QoS gap on LTE, and WiFi 7 will eliminate that altogether, but will leave 5G with a slight edge, especially in the millimeter wave bands. LTE and 5G both score over WiFi for reliability and it remains to be seen whether measures such as Open Roaming reduce or eliminate that.
Currently, private WiFi as deployed in enterprises still dominates general purpose indoor wireless connectivity, and it is in public arenas such as hotels and airports where LTE and 5G have come into contention. In such cases, both WiFi and cellular can suffer from congestion and the outcome of competition may depend on deployment. WiFi in such venues is sometimes run by third party managed service providers, which have woken up to the need to improve capacity to stem an exodus to LTE or 5G on the smartphone side, although laptops often lack cellular connectivity.
Meanwhile, not all enterprises outside manufacturing and engineering share Keysight’s view that the time for private 5G has not yet quite arrived. The US maker of agricultural equipment John Deere has recently been on record arguing that other vertical sectors can benefit from 5G now. The firm has argued that farming can be regarded as outdoor manufacturing and can benefit from high performance, low latency connectivity in a similar way. Certainly, John Deere itself now offers autonomous farm vehicles such as tractors that would benefit from high performance wireless communications, even if it does not quite justify the ultra-low latencies enables by 5G Standalone (SA).