What applications/services really need 5G?

This blog is guest authored by Dean Bubley, Founder & Director, Disruptive Analysis.

5G represents an important upgrade for the mobile communications industry – or more accurately, an important series of upgrades arriving in phases. This has potential to enable a range of new applications and services, for both conventional MNOs and new network owners/operators.

But the timing and sequencing of 5G phases is crucial.

In recent years, the mobile industry has sometimes oversold the near-term opportunities of 5G, or confused them with longer-term possibilities. Given the complexity of the technology this is perhaps understandable but nonetheless unfortunate. It has also sometimes overlooked alternative technologies that can satisfy the same requirements – perhaps sooner.

For example, various commentators have falsely asserted that “autonomous vehicles will need 5G”, or have made cliched comments about sci-fi applications such as “robotic surgery”. Coupled with misunderstanding about when and where ultra-low latency may be available or needed (hint: only in very specific locations) this has meant that some weary observers have perceived the whole industry has overhyped.

The irony is that this obscures some genuine good news.

Certain applications and services are much more 5G-centric, or even 5G-dependent. Some of these are still quite a long way off in the future, as they need later versions of the radio and core technology, but others are more viable in the next 2-3 years.

Disruptive Analysis believes that MNOs (as well as governments, vendors and industry bodies) need to focus on what is actually realistic to achieve with 5G, given its strengths, weakness and – especially – market timing. The industry needs to stop “chasing unicorns” and instead talk about realistic, credible goals.

Which aspects of 5G enable unique applications / services?

The previous blog in this series [link] already considered options for monetising the earliest versions of 5G. The general conclusion was that it offers "more of the same", as most current deployments use non-standalone mode, which links the new radio network into the existing 4G core network.

5G NSA does not give additional features beyond higher speed and capacity – but it allows for new pricing models and bundles that MNOs can use as an initial step to earn a return on their new investments. Examples include:

• Flat-rate mobile broadband (MBB) offers, given potential lower costs for the operator and access to new spectrum
• Faster fixed-wireless access (FWA), expanding the addressable market previously open to 4G FWA and other technologies. In some circumstances, 5G FWA can compete against cable or even fibre offers from rivals.
• Opportunities for marketing and bundling of certain applications such as gaming or AR/VR – even if they do not currently need 5G connections to function well, they can be used as a lever to encourage subscribers to upgrade to higher tiers or new devices.
• Basic mechanisms to differentiate service or experience levels – for instance some form of SLA (service level agreement) with 5G subscriptions, or unlimited UHD (ultra high definition) video-streaming rather than just SD.

However, nothing here is unique to 5G. All of these applications can work well on good 4G networks, or indoors on well-engineered Wi-Fi, at least for now. It is future versions of 5G that start to make a difference. Most readers will be familiar with the acronym URLLC (ultra-reliable low-latency communications), for instance – which is at the core of many later 5G release features.

So for instance, future AR/VR might be used outdoors, in vehicles or other locations where 5G is preferable – for example by field engineers for utility or aviation companies. The history of "wide area" games such as Pokemon Go also suggests continued evolution could lean on the abilities of 5G, once coverage is sufficiently predictable. Having an early "starting point" for 5G could offer useful experience to both operators and developers.

Key aspects of genuine 5G-centric applications include:

• Throughput / speed: 5G can – given enough spectrum – offer multi-gigabit data speeds. This high bandwidth is important for uses such as fixed-wireless access, enterprise SD-WAN and selected demanding applications such as streaming video from public safety cameras on vehicles or worn by personnel.
• Capacity: The corollary of network speed is the density of bandwidth that can be delivered to multiple users or devices in crowded areas, such as stadiums or busy streets. 5G enables more aggregate data downlink or uplink, per square meter of coverage. This is already happening.
• Latency: Current versions of 5G can enable "ping times" of perhaps 20-30ms, which can improve the performance of some applications such as gaming or some transactions. Over time, this will reduce with subsequent technology updates, although realistically not a sudden or ubiquitous fall to much lower levels. 15ms will become more widely achievable, then 10ms and so on. In well-engineered deployments such as private or campus networks, reliable ("deterministic") latencies of 1-3ms should eventually be available.
• Reliability: 5G networks are being designed to offer greater levels of redundancy for critical applications and services, as well as improved performance at the edge of cells' coverage. It is worth breaking down the familiar URLLC into the UR and LL components separately.
• Location: Later versions of the 5G standards allow increasingly accurate positioning of users to be provided from the network – from 10m to 1m and eventually 10cm or less. This may also work indoors, providing significant differentiation from GPS or other methods used by applications today. For example, this would mean a robot in a warehouse could accurately gauge which shelf it is next to, or a car's navigation system would be reliably aware of which lane it is in, or even how to avoid a hole in the road.
• Security: Considerable extra work has gone into ensuring the security of 5G RAN and cloud-native core networks. While any new technology also brings additional possible threat-surfaces, there has been considerable emphasis on cybersecurity built-into 5G technology, as well as the more recent focus on the vendor-driven aspects. Numerous governments are developing test and monitoring capabilities for 5G that go beyond 4G activities.
• Private networks: Although enterprise-run 2G/3G/4G/5G networks are already proliferating, later versions of the 5G standards provide some more "official" guidelines for what 3GPP terms "non-public networks". Overall, 5G brings some significant opportunities for non-MNO network to be deployed, especially for demanding on-campus applications such as IoT and industrial automation.
• Vertical-specific features: In 3GPP Releases 16 and onwards, there are various feature-sets intended for the needs of specific industry verticals and applications. The most well-known are probable C-V2X (cellular vehicle-to-everything) for automotive applications, as well as critical-communications enhancements for public safety and similar services. Later versions should also have special features for broadcasting, aviation, railways, medical and other use-cases.

That said, certain other aspects of 5G are less-important for new applications. The "massive IoT" capabilities of thousands (or even millions) of devices per square kilometre sound impressive, but are largely deliverable with 4G-era technologies such as NB-IoT, or non-cellular alternatives such as LoRa and Wi-Fi.

Edge computing is also an interesting general trend for applications, but is not dependent on 5G or specifically linked to its deployment – it is access-neutral, and indeed most Edge functions will be connected with fixed/fibre, or even embedded directly into devices. There may be some unique 5G + MNO-owned edge applications, although at present this is mostly hypothetical.

What determines 5G applications' realistic possibilities?

While 5G technology offers these various theoretical new application and service opportunities, operators need to consider more carefully when (and where) they can actually be offered to customers.

They need to understand that what will determine the real-world opportunities for their specific circumstances. Not all MNOs will be able to offer the same 5G portfolios. There needs to be a mechanism for applications to assess what capabilities are actually realistic at a given place and time.

Key criteria will include:

• Spectrum: This spans both a particular MNO's specific assets, and the national regulator's plans and rules for future releases. Frequency allocations determine coverage, capacity and ability to support features like URLLC.
• Infrastructure: What sites, fibre and other supporting infrastructure does an MNO have access to for 5G, either in-house or via third-party providers? Some IoT and V2X services have much wider coverage requirements (for instance in rural areas), compared to consumer applications that can be constrained to dense urban zones.
• Indoors / outdoors: Indoor coverage for high-performance versions of 5G is a major issue, as it involves offering public services on private property. High frequencies will not go through walls well, so specialised (and perhaps costly and time-consuming) specialised systems – and perhaps partnerships - will be needed for some services.
• 3GPP versions: Many of the more advanced 5G applications rely on features in later phases of 5G - especially 3GPP Releases 16, 17 and 18. The timing of deployment and commercialisation will strongly drive an MNO's 5G application opportunities. Release 16 was finalised in mid-2020, with the earliest commercial deployments expected to start in late 2021, with more MNOs adopting it during 2022 and 2023. Release 17 – which include features such as industrial URLLC - is likely to face greater challenges with timing because of the pandemic. Originally expected to be completed in 2021, it is more likely to be early/mid-2022, which then implies a late-2023 initial commercial launch or later.
• Devices: Many 5G possibilities will depend on availability of suitable end-user devices – from smartphones to IoT-suitable modules, AR/VR headsets or fixed-access modems. These too need to support the right bands, 3GPP Release features – and be available at a price to make the solutions economical. Operators can wait for the market to evolve organically, or can choose to more actively source, design bundle and perhaps subsidise certain product groups. (Timing may be impacted by semiconductor supply chain issues as well).
• Skills and understanding: Many operators have only a superficial understanding of applications – how they are designed, how they are bought, the behaviour of users and developers, or their paths to market. There are many skills gaps that need to be filled for 5G applications to become real, both on the network side (for example with slicing or capability-exposure) and in terms of sales and support.
• Partnerships: Many 5G applications and services require complex partnerships and ecosystems. MNOs need to develop new relationships with application developers, systems integrators, different branches of government/regulators and perhaps sources of finance for end-users. It will not be possible to cover all verticals and horizontals, from sports to agriculture, or from military to retail sectors.