Transforming Industries at the Edge with 5G and AI
Published on : Wednesday 02-09-2020
Companies must contextualise the benefits of 5G and prepare for a 5G future, as it is inevitable, says Rob van den Dam.
The fifth generation of mobile networks – in short 5G – is poised to shift the dynamics of virtually every industry. It will become the foundation platform for many new applications in industry verticals – such as manufacturing, transportation, and healthcare – and enable entirely new groundbreaking use cases simply not possible today. However, this ‘revolutionary’ aspect of 5G remains beyond the near-term horizon. Though the technology strategy is fairly mature, it will still take years to migrate to a 5G environment that makes such a transformation possible. It depends on having a cloud-native 5G core network, leveraging extensive network virtualisation, and implementing AI-based automation.
Most telecom operators are still in the early phases of developing this envisioned 5G network infrastructure. They are proceeding with caution because of the high risks and investments involved in moving toward ubiquitous 5G networks. For example, according to a report on 5G spending from Greensill, the world’s biggest non-bank provider of working capital solutions, infrastructure upgrades will need global investment in the neighbourhood of USD 1 trillion. Rolling out 5G requires denser (than 4G) arrays of masts and small cells, an intensive fibre-optic backhaul, and many other new infrastructure elements. Furthermore, additional – mostly high-priced – 5G-spectrum has to be acquired. And these are only the anticipated telecom investments. They don’t include, for instance, upgrading Internet of Things (IoT) devices.
Telcos will need to make these large investments during a period of relatively flat revenue growth. In 2018, the global annual year-over-year revenue growth in the telecommunications industry was 1.4 per cent – less than half of global GDP growth. And this growth is expected to slow to around 1 per cent by 2025, according to the GSM Association (GSMA).
Furthermore, between 2010 and 2018, telcos spent globally well over USD 1 trillion on upgrading their networks to 4G. They still have to amortise their 4G core networks. Many of them are even still upgrading their 4G networks to cope with growing demand for bandwidth.
As a result, most telcos will not prioritise the core network in the first stages of 5G deployment. Instead, ‘islands’ of 5G New Radio (NR) coverage will be added to the 4G core, mainly to solve the capacity issue. This ‘evolutionary’ approach will be the natural approach for most telcos, allowing them to lessen initial investments with the revenue potential remaining small.
Dependency on 5G spectrum allocation
Timing of 5G rollout in each market largely depends upon the availability of 5G-compatible devices and 5G spectrum allocation, with each country having its own schedule for either reserving, auctioning, or making formal plans for 5G spectrum.
The spread of Covid-19 has made various regulators around the globe decide to postpone planned auctions of the 5G spectrum, in particular in many European countries. Also in countries like India, where 5G auctions may not happen to start until at least 2021, which would push back the launch of commercial 5G services until at least 2022.
India is special as the timing of the 5G auction is also influenced by other factors. The poor health of the Indian telcos and the high reserve price for spectrum – one of the highest in the world – are key obstacles to attract interest on short notice. All the more because the completion of a full-fledged 5G network will require an additional investment of around USD 70 billion, part due to the country’s lack of a widespread fibre-optic backhaul infrastructure.
India is also special for another reason. American tech companies – such as Facebook, Google, Microsoft and Amazon – seem to be increasingly keen in investing in India’s mobile sector. These tech giants are very focused on India as the fastest-growing internet market with 650 million users, most of them being serviced by the mobile telcos. By partnering with these telcos, they are better placed to monetise their vast customer base, while it can help the Indian telcos to (partially) fund the 5G roll-out and to monetise today’s 4G and future 5G networks.
Creating private 5G networks
In the 5G era, telcos can no longer be assumed to be the sole parties obtaining spectrum. 5G could see explosive growth in private enterprise networks, according to a GSMA Intelligence report. This nascent movement toward allocating spectrum for enterprise – as opposed to operator – use will likely fuel demand within the next five years. Manufacturers, for instance, might opt to build private 5G networks for their in-factory applications.
A number of countries have carved out spectrum for private licenses in its 5G allocations, some even without auctioning but instead assigning spectrum to applicants on request. German airline Lufthansa, for example, has gained a spectrum license for operating 5G in the 3.7-3.8 GHz band. And BMW, Volkswagen, and Daimler have already expressed interest in operating private 5G networks for their German plants. As another example, Fujitsu was granted Japan's first commercial Private 5G radio station license from the Kanto Bureau of Telecommunications.
Blending 5G, edge computing and AI to create the perfect storm
5G supports 3 broad categories of use cases: enhanced mobile broadband (eMBB) for increased throughput and improved coverage and capacity, massive machine-type communications (mMTC) to support mass adoption of IoT services across many industry verticals, and ultra-reliable low-latency communications (URLLC) for mission-critical applications.
However, 5G is only one of a set of technologies coming together to potentially reinvent entire industries. It is particularly the blending of 5G, edge computing, and AI that will impact the industries and will enable many new enterprise and industrial applications (see Figure 1).
Delivering the promise of network slicing
With a full-fledged cloud-based virtualised 5G network available, telcos can deliver services to specific verticals through network slicing, which provides the ability to deploy any number of virtual end-to-end networks on a single physical infrastructure, with each virtual network representing one slice.
It enables telco to reserve a slice for a specific customer or application at a guaranteed quality of service. Network slicing allows operators to dynamically balance disparate requirements – such as availability/reliability, throughput, and latency – among different applications, such as those for remote healthcare and autonomous connected cars (see Figure 2).
Enterprise and industrial applications will especially benefit from 5G network slicing in combination with edge computing. Deploying network slicing and edge computing in a 5G core network environment creates the possibility for previously unimagined latency-dependent wireless solutions, including those for connected autonomous vehicles and smart factories, among others.
Teaming up: Edge computing and 5G
Without edge computing, 5G applications and services will rely upon connecting through the core network to centralised cloud resources for storage and computing, losing much of the positive impact of the latency reduction enabled by 5G. The basic idea behind edge computing is that by storing/accessing data and performing high-demand processing tasks closer to the network edge – that is, on the network gateway, customer premises or edge devices – network congestion and latency are significantly reduced. These reductions are crucial for latency-critical applications dependent on continuous corrective actions.
Organisations are already capitalising on edge computing’s capabilities to enable real-time insights from connected devices and systems, real-time equipment monitoring, and real-time inventory management. For example, sustainable agriculture companies – such as New Jersey-based vertical farming company AeroFarms – are equipping plants with IoT-enabled sensors and using edge computing to monitor the growth needs and ideal harvest time for individual plants.
Automotive companies are essentially making cars into edge devices, equipping them with internal and external sensors that generate data. Tesla, as an example, manufactures cars that make intelligent autonomous driving decisions via powerful on-board computers using edge computing for making decisions and actions in real time – from braking to steering and lane changes.
Organisations are counting on edge computing to help them be more responsive to changing conditions and customer interactions. In IBM’s edge computing survey, 84 per cent of executives said edge computing applications will have a positive impact on their organisation’s operational responsiveness in five years. Executives from the automotive and telecommunications industries are particularly confident in edge computing’s capabilities in this area (see Figure 3).
Ramping up AI in 5G networks
Edge computing will require more and more processing power at the edge of networks. Gartner predicts that, by 2022, more than half of enterprise-generated data will be created and processed outside of data centres, and that by 2025 about 75 per cent of data will be analysed and acted upon at the edge. AI technology is crucial to process the large amounts of data at the edge and provide the insights to drive quick, local, data-informed decision making. In fact, AI has become the key driver for the adoption of edge computing.
Locating AI close to the edge is vital for applications where near-real-time feedback and optimisation are a priority for applications – such as machine control, equipment monitoring, and remote surgery. Faster AI-based decision-making and responses at the edge also result in better user experiences – for instance, in the areas of AI-assisted driving and multi-player gaming. In addition, edge computing supports improved security because sensitive data can be analysed at the edge and doesn’t need to be sent over the core network to a central cloud, in this way limiting the amount of data stored in any one location.
Many telcos are already evaluating edge computing on 5G networks. For example, Verizon is developing new 5G edge technology that could revolutionise mobility for virtual reality (VR), augmented reality (AR) and extended reality (XR) in enterprise applications. As another example, SK Telecom has launched a 5G edge computing open platform to third parties and enterprise customers. And Bharti Airtel recently announced to work with IBM and Red Hat to launch new edge computing services and solutions, backed by a broad ecosystem of partners to help enterprises and telcos speed their transition to edge computing in the 5G era.
Industry 4.0 – Moving from massive production to massive customisation In industrial IoT, 5G is the crucial element in the Industry 4.0 evolution toward increasingly smarter and more dynamic manufacturing. Indeed, manufacturing is expected to be the largest beneficiary of 5G services. The unique attributes of the technology could unlock USD 740 billion of value in manufacturing by 2030.
The combination of 5G, network slicing, edge computing, and AI will be a driver of Industry 4.0 with billions of machines, devices, and sensors just waiting to be wirelessly connected. This will become the backbone of manufacturing and related services in the future. It will be powered by robotics, AI, IoT, 3-D printing, AR, and cloud technologies, all of which will use 5G technology to allow machine-to-machine communication.
Industry 4.0 is completely changing the concept of manufacturing from massive production to massive customisation. Industrial connectivity today mainly uses wired connections, which provide the high performance and reliability required for automation, but lack the flexibility to meet changing production demands. A flexible and programmable environment based on highspeed, low-latency wireless 5G connections, the use of edge computing and AI – will link machines, processes, robots, and people. This environment sets up the opportunity for more flexible and dynamic production capabilities to meet rapidly evolving market needs and massive customisation.
The concept of 5G with network slicing and edge computing is being evaluated and tested in several Industry 4.0 scenarios, with telcos as one of the ecosystem partners. For instance, Vodafone, Ericsson and the German electric microcar company eGO have launched a 5G Industry 4.0 car manufacturing plant, helping deliver security-rich, near-real-time data networking across the production chain, from digital material management to autonomous vehicle control. As another example, Singapore-based operator M1 is collaborating with IBM and Samsung to develop and test Industry 4.0 solutions using 5G and edge computing, among others to improve the control of robotic arms and other smart devices on factory floors.
Preparing for a 5G future
It will still take several years before we begin to see a significant uptake of 5G. But a 5G future is inevitable, it isn’t really a choice. Companies must contextualise the benefits of 5G and prepare for a 5G future. Sooner or later the envisioned 5G networks will become available and new products and services will be launched. It is important to prepare a solid foundation for seamless 5G on-boarding whenever the time comes.
Rob van den Dam is the global industry leader Telecommunications, Media and Entertainment (TME) for the IBM Institute for Business Value (the business think tank of IBM). He leads TME strategic thought leadership and is a contributor to IBM’s global telecom strategy. He has 25 years’ experience in this industry and has worked in a range of advisory roles for major TME organisations.
LinkedIn: https://www.linkedin.com/in/robvandendam/
The full research can be downloaded for free from https://ibm.co/5g-telecom
