There have been a lot of superlatives thrown around about 5G, up to and including it being the most important invention since electricity. While that is almost certainly not true, it is one of the hottest technology topics today. 5G offers a number of significant improvements compared to previous mobile generations. Those involved with the Internet of Things (IoT) have been intrigued to understand what capabilities might be used for connecting things other than phones, tablets and PCs.
5G is the Fifth Generation of the technology standards for cellular communications. The capabilities required for 5G have been defined by the International Telecommunications Union. The standards have since been developed by the Third Generation Partnership Project (3GPP), which is a consortium of the major global standards development organisations. Its name stems from its creation to develop the 3G standard in the late 1990s and it has stuck ever since. All subsequent cellular technology evolutions from 4G (i.e. LTE) onwards, have been standardised by 3GPP.
3GPP adds in new features on a regular basis, through a series of ‘releases’. The latest was Release 16 which was frozen in July 2020. Release 15 included some elements related to 5G including ‘New Radio’ (5G NR), and IoT-related elements. Release 16 included substantial refinements to these, effectively providing a full 5G system. Further evolutions will occur in future, including refining capabilities and adding new ones, for instance Release 17 includes significant work on ‘Non Terrestrial Networks’, e.g. satellites.
5G has significant changes to architecture and capabilities relative to 4G. With regard to 5G NR and the access network, there are three major capabilities delivered by 5G that will be of interest for IoT:
The following sub-sections explore each of these capabilities in turn.
Theoretically 5G New Radio offers speeds of up to 10Gbit/s but the reality is that the experienced maximum speeds by a single user will typically be 100-200Mbit/s. This represents a significant improvement (about five-fold) over LTE. The predominant benefit of this capability is to give a richer experience for mobile broadband usage. Most of that relates to video, gaming and other high bandwidth streaming. It also makes mobile a viable alternative to fixed line broadband for more households.
In the context of IoT there are a few immediately valuable use cases. Augmented/virtual reality (AR/VR) is one, and there are also other applications where higher bandwidth will improve the experience, such as for connected car and for video cameras, for instance for CCTV. While higher bandwidth is certainly appreciated, and history suggests there is an almost unquenchable desire for more bandwidth, no-one has yet come up with a game-changing high bandwidth IoT use case that means that 5G is anything other than a welcome incremental addition courtesy of delivering superior throughput in a more cost-effective way. Growth of the metaverse, with associated demand for AR/VR may prove to be the critical use case, but today there is no killer app awaiting these higher data rates.
This capability set is aimed specifically at IoT. The headline functionality is the ability to support at least 1 million devices per square km (up from 100,000 with LTE).
The 5G standard also incorporates, and builds upon, a set of existing standards aimed specifically at supporting low bandwidth, low power, IoT devices, including features such as power-saving mode (PSM) and extend Discontinuous Reception (eDRX) which extend battery life. These LPWA (Low Power Wide Area) technologies, NB-IoT and LTE-M are discussed at length in another of our Hot Topics pages: Low Power Wide Area Networks. In addition, within 5G Release 17 is 5G RedCap (Reduced Capability) which sits somewhere between full 5G NR and the LPWA technologies.
Regarding the ability to support massive deployments, we are currently nowhere near needing this additional capacity. According to the Transforma Insights IoT Connected Devices Forecast there will be 6.7 billion connected devices in use across wide area and campus networks worldwide. Most of these will be simple sensors, reporting only occasionally. The volumes of connections do not particularly indicate an overwhelming demand for cells supporting thousands of devices, let alone hundreds of thousands, all active concurrently.
Under the banner of URLLC 5G promises two changes that will open up certain use cases, particularly associated with IoT. The first is reducing latency. Latency refers to the delays in getting data packets from point A to point B. This is typically not a major consideration for, say, video streaming, but it is for gaming, where responsiveness is important. The time it takes for a message to traverse the network from the games console to the server and back again is critical to the user’s enjoyment. It is also a critical component of some of the most sophisticated 5G use cases that are proposed, such as remote surgery or managing autonomous vehicles. Another example is energy grids, where split second control might be necessary. The responsiveness of a device to the messages that are being sent to and from it needs to be very close to real-time. Clearly, in many of these examples, the requirement for low latency is associated with critical systems where reliability is also paramount. Hence low latency is bundled with ultra-reliability as a requirement.
Latency is measured in milliseconds, i.e. the number of thousandths of a second that it takes for a ping to travel across the network, or part of the network. Historically, mobile networks had quite poor latency in comparison with, say, fibre. LTE, for instance, has latency of over 50ms for the hop from the device to the cell tower. This compares to 20ms delay added by a 1,000km round trip from the tower to a central server via fibre. Plus there’s likely some other delay added by moving through various network elements along the way. But, with an LTE network it’s safe to assume for most use cases that it is the radio access network part that is the delaying factor.
5G promises latency as low as 1ms, although in reality it will be more like 10ms for most applications. While a 5x improvement might not appear transformational, it may well be. This is less because of the capabilities that it enables and more because of the relative latencies of the various parts of the network. Getting total latency below 100ms, and ideally below 30ms, is needed for gaming and AR/VR. 5G certainly delivers that. However, demand for those types of services is, at best, unproven. A sweet-spot for this kind of low latency connectivity occurs in the context of 5G Mobile Private Networks supporting industrial automation. The change that 5G creates is that the core network, rather than the access network, becomes the slow part. Historically the radio access network (RAN) was always the bottle-neck. The 20ms delay added by a 1,000km round trip to a central server was not the limiting factor. With 5G as the RAN, the core network suddenly represents the majority of the delay.
The overall impact of a 5G NR access network would be to shake up the relationship between device, access network (i.e. the 5G) and transport network (i.e. the part that connects the base station to servers). In the past 10 years or more we have seen an inexorable move to the cloud, based on faster and faster transport network speeds. But, being limited by the speed of light, those fibre networks are unlikely to get much faster. There will increasingly be an advantage to putting more processing and storage at the edge of the network, for instance with 5G base-stations doubling up as mini data centres. Such a reduction in latency also encourages the shift of compute power for IoT applications out of the devices themselves and into the base-station; why have smart edge devices when the network latency back to a more cost-effective network edge compute function is only 10ms? The result of 5G deployments therefore is that compute shifts to the edge of the network, either from the cloud, or from the device, or both.
For more explanation of Edge Computing, including both network edge and device edge, see the Hot Topic Hub: Edge Computing.
There are numerous other features included as part of the existing 5G standard and on the roadmap. These include: vehicle-to-everything (V2X), NR for unlicensed spectrum, enhanced location services, satellite access and network slicing (which is effectively the capability to offer different user groups different grades of service over a shared network).
In terms of volumes of devices, the greatest impact of 5G will come from the use of the mMTC technologies NB-IoT and LTE-M, and subsequent evolutions of them in forthcoming Releases. Hundreds of millions of smart meters, environmental sensors, consumer goods and asset trackers (as well as numerous other applications) will be connected over the next ten years, far in advance of use of high bandwidth and/or low latency connections, at least in terms of volume of devices. For more on the two LPWA technologies, see our Hot Topics Hub: Low Power Wide Area Networks.
While the LPWA technologies may account for the majority of the volume, they won’t necessarily account for the majority of the revenue delivered by 5G. It will be the high bandwidth and/or low latency applications that will be the most impactful, and generate the most significant revenue.
The earliest ‘full’ 5G deployments in IoT will focus on private campus networks. These are the quickest and easiest to deploy and have a tremendous amount for focus from the network operators and infrastructure vendors today. Use cases include factories, warehouses, hospitals and oil refineries. For more on Mobile Private Networks, see our Hot Topics Hub: Mobile Private Networks.
Other applications will make strong use of 5G’s high bandwidth capabilities. We expect strong adoption of 5G in connected cars, where the automotive OEMs are always keen to future-proof their products, as well as other high bandwidth applications like CCTV and consumer electronics.
The applications relying on ultra-reliable and/or low latency communications will probably be slower to arrive as they represent more of a sea-change compared to existing use cases. Today there is a lot of testing the waters of potential use cases that will demand URLLC capabilities, such as autonomous vehicles or remote surgery. We await the killer use cases, although we expect that these will first emerge in the context of enterprise Mobile Private Networks.
For more details on which applications will see the greatest adoption in IoT using either 5G NR or 5G mMTC, see our Forecast Highlights page, which includes top level data drawn from our TAM Forecast Database.