Enhanced mobile broadband and fixed wireless access will lead the way, but IoT applications and use cases requiring ultra-reliable and low-latency communications can also be supported more efficiently on 5G networks than on previous-generation infrastructure.
There is one overriding question when examining the case for new 5G business models: is there an elephant in the room? It's the old fable of three blind people with the putative pachyderm, each describing quite different animals depending on whether they're at trunk, flank or tail. If 5G is that elephant, then operators and suppliers will see multiple sources of revenue while having just the one mouth to feed: if those use cases need very different networks, the benefits will look less elephantine.
The industry expects new services fall into three camps: enhanced mobile broadband (eMBB), ultra-reliable and low-latency communications (URLLC), and massive machine-type communications (mMTC). Each class makes different demands of bandwidth, latency, and intelligence.
5G is designed to handle this mix of services better than LTE can. 5G's Radio Access Network (RAN) specification manages all the radio functionality -- coding, antenna management, errors and retransmissions, and so on. The RAN connects to the Core Network (CN), which sets up connections, and controls bandwidth management, authentication and other classic network functions. What makes the 5G CN so special is that it can handle multiple different configurations of RAN simultaneously, the so-called 'multi-access scenario', selecting those that match the needs of different services without needing separate network functionality. These different sets of functions are called 'network slices', and it's how operators build and target network slices that will define their success.
LTE has a nominally similar basic design with its own Evolved Packet Core (EPC) handling network functions, but lacks 5Gs flexibility both in radio and core sectors. However, 5G's RAN can interoperate with the EPC, as the CN can with LTE radios. This allows 5G to move from the initial Non Stand Alone (NSA) mode, where it supports and requires LTE radios and infrastructure, through the three-to-five year plan which will see the emergence of pure or Stand Alone (SA) 5G. The lack of this blended interoperability transition hindered the move from 3G to 4G, not least by requiring much larger investments in infrastructure before returns were realised: 5G's new business cases have somewhat better engineered financials to go with the improvements in technology.
Enhanced mobile broadband
The first new business case looks very familiar: enhanced mobile broadband (eMBB) is faster data to handsets. This is already being deployed -- EE has started putting 5G in high-density areas of six UK cities, with a further ten promised for 2019. It is partnering with two handset makers, OnePlus and another to be announced, and the company says that the major use case is to relieve congestion at venues and areas where 4G is overloaded at peak times, such as major railway stations.
5G's use of higher frequencies than 4G will necessitate more and smaller cells, the cost structure of which needs a new topology. Instead of most traffic going through cell sites that handle lots of baseband processing (turning data into radio and vice versa), a large macro cell will handle the baseband needs of around twelve sub-cells, which EE calls the 'baseband hotel' model, and it simplifies control and management.
Fixed wireless access
Another new business model that builds on the old is fixed wireless access (FWA). Already quite well advanced, with LTE links providing cable/DSL speeds to businesses and homes, the biggest upside for 5G is its new radio bands, especially in the higher, millimetre-wave 28GHz and upward spectrum. One company that's already building out 5G FWA is CBNL, a Cambridge UK-based manufacturer that has pre-5G systems with hundreds of operators in more than 50 countries. It considers the fibre-like speeds achievable match well with the halo effect of 4G's mobile broadband becoming widely accepted. With fixed installations, there's a lot of flexibility about antenna size and power usage, which can push the usable range of millimetre wave links to kilometers.
There may be a natural end point in the speed battles for broadband delivery. At 1Gbps, home 802.11ax wi-fi and affordable wired Ethernet are saturated and even with multiple independent HD video streams there's not much need for more. 5G can deliver that while DSL can't, and many cable systems have problems; wireless data needs no new cabling to the end point. Switching to gigabit services without having to dig up the street is an appealing option.
Massive machine-type communications
5G's IoT-focused massive machine-type communications (mMTC) technologies are inherited wholesale from the latter revisions of LTE. There are two main classes: the low-power, narrowband NB-IoT and the medium power, medium bandwidth LTE-M. The first is aimed at the classic IoT model of static, embedded sensor and control nodes; the second is geared towards supply chain and more flexible uses. 5G absorbs these standards as they are and adds the new bands for greater capacity and the new core network capabilities to add more custom network slices for different use cases. Future releases of 5G will add novel IoT features, such as direct device-to-device communication, but for now it's existing functionality in new wrappings.
5G's IoT will compete with a whole slew of technologies -- LoRa, Ingenu (formerly OnRamp), Sigfox, Telensa, and others in unlicensed bands. Plus, the 802.11ah wi-fi standard -- a.k.a. HaLow -- is due this year. While each is bullish about its prospects, the market is both small and fragmented with a lack of common standards and off-the-shelf complete networks making deployment risky and expensive.
For example, the UK government's utility smart meter push has been going since 2013, but is only around a third of the way towards the 53 million target for the 2020 deadline. With a high failure rate and cost overruns, a glum Parliamentary report said it may cost more than the £16bn total expected benefit: its dependence on 2G networks for communication to utilities and the almost-moribund ZigBee for home controllers has made integration with smartphones and home networks very difficult, and left the system dependent on 2G not being phased out.
5G's big hope in IoT is that it can provide competitive services on the back of infrastructure that's already in place for other uses, and -- unlike older systems like 2G -- the network has a very great deal of flexibility. It uses licensed spectrum, which is managed and not susceptible to third-party interference, and inherits cellular levels of authentication and security.
Operators are also building on experiences of how IoT is paid for. Even when an IoT node is used by a consumer, they don't pay for service as they would for mobile access. Instead, operators, device makers and service providers use a variety of cost- and benefit-sharing revenue models, including Right To Use -- where the operator provides access to defined ports, capacities and bands, and then gets paid according to use. Subscription, permanent licence and other mixtures of connectivity-as-a-service plus support are also used as IoT revenue models. Again, 5G's ability to create new service types on existing infrastructure is seen as advantageous.
Ultra-reliable and low latency communications
Ultra-reliable and low latency communications (URLLC) opens up use cases where safety and life-critical tasks are involved -- not the sort of things that sit well with LTE, with its typical latencies of 50ms to several seconds and block error rates before retransmission of one in ten. 5G, by comparison, will be able to achieve sub-1ms latencies and error rates in the one in a billion. Three typical use cases become feasible at these levels.
Telesurgery is in some respects an established procedure, where a specialist surgeon in one hospital can operate on a patient in another, provided there is excellent connectivity between the venues. Sub-millisecond latency is needed to remove lag between the actions of the surgeon and the video they see of what's happening. That video needs to be very high quality and stereoscopic, and other channels are needed for haptic feedback from instruments.
Providing wireless connectivity extends the range of telesurgery to those patients who are too ill to be moved, who are too far away, or where physical communication between them and the appropriately equipped hospital is disrupted due to natural or manmade disaster. Mobile surgical units could also provide services for whole populations.
URLLC fits in particularly well where there is good low-latency, high-bandwidth communication to population centres surrounded by a large rural contingent. This models work in India, where fibre links have been laid alongside the rail network, providing telemedical services near stations. Wireless can extend this considerably.
Autonomous driving becomes much more efficient with low-latency, high-reliability networking. When cars can cooperate with each other and share information with roadside infrastructure, tasks like automated overtaking, cooperative collision avoidance and high density 'platooning' -- vehicles forming up into road trains for more efficient driving -- become possible and desirable. This sort of task needs latencies below 10ms and block error rates in the tens of thousands. That's not as stringent as telesurgery by any means, but at much higher densities and with a much more dynamic environment.
A final example of URLCC is factory automation. This is an already highly automated environment, but mostly via wired networks. Robotic production lines, industrial processes and factory-wide monitoring and power management requires the same level of performance as telesurgery if safety and efficiency are to be maintained, and wireless networks haven't been able to deliver the required reliability or performance. Once they can, though, the advantages in flexibility, increased reliability due to no cable motion-related breakage, speed of deployment and lowered maintenance costs will make wireless the preferred infrastructure.
Not all new use cases for 5G fall into the three industry-sanctified groups. A good example of a very different business use comes if you look up. Dronesthat fly out of line-of-site of their controllers rely on data networks, but no terrestrial networks have been designed with drone control in mind. Drones can and do use LTE networks, but the radio environment at flight altitude looks very different to that on the ground. Multiple cell sites will have unblocked line-of-site routes to a drone, which will see many more strong interfering signals than at ground level. Also, the drone will fly into and out of 'sidelobes' -- small areas of strong signal created by antenna configurations -- which will have a similar effect to a drone in motion as seeing a lakebed through the surface of rain-spattered water.
Those are radio issues, but other problems exist, as recent events have illustrated. Drones have significant security-threatening and disruptive capabilities, but are notoriously hard to identify and shut down.
5G has multiple enhancements to provide explicit support for drone usage, including specific drone identification and authorisation, height and location reporting, interference detection triggered when a configured number of cells is reached, power control enhancements managing interference versus signal ratios, and signaling of flight path information from the drone to the network. Other aspects of the 5G specification that are not explicitly designed for drone support -- such as multiple-cell delivery of a composite signal to a single terminal, beam-steered control channels and enhanced inter-cell handovers -- also provide a useful set of tools for supporting flying terminals without disrupting ground-based services on the same infrastructure. If high reliability, ultra-low latency links can be added to the mix, then who knows -- remote drone hire from home may be plausible, with drone sport or extreme adventuring thrown in.
One network, many services
It isn't clear how big a business model drone support will be, but the idea is that it becomes possible to provide a variety of network services for novel functions without having to make major changes to physical infrastructure or across too much of the software stack on which the network runs.
Ericsson, which sells a lot of the infrastructure the core 5G network will be built on, says that the best mental model is of a cheese shop that used to sell just cheddar, but now has to offer hundreds of different kinds of cheese. The old core cheese business may not be so important, but with lots of new customers prepared to pay for unique offerings, the whole business remains profitable.
Whether network operators are ready to make that change remains to be seen. There is a long list of technically possible and genuinely useful new ideas in telecommunications that either never happened or died in infancy due to lack of corporate flexibility. With 5G, though, the tools are there if the imagination requires them.
5G market predictions for 2019 (TechRepublic) Find out how fast experts anticipate 5G rolling out next year, as well as what to expect and where to find it.
5G technology: A business leader's guide (Tech Pro Research) It's still early days for 5G services, but as they're being trialed and deployed, they're on track to have a far-reaching impact for both consumers and businesses.