The first phase of 5G is now becoming a reality with mobile network operators (MNOs) launching 5G services in 2019. Ericsson, as one of the world’s leading cellular infrastructure and solution providers, is at the forefront of making this happen.
Ericsson has a full suite of 5G offerings ready to go, including non-standalone (NSA) 5G and standalone 5G New Radio (NR) solutions.
The initial wave focuses on enhanced mobile broadband to provide faster and higher data bandwidth. Later waves will deliver enhanced mobile broadband in vehicles, tablets, gaming devices, plus immersive AR and VR, along with lowcost, low-power wide-area (LPWA) Internet of Things (IoT) and support for massive IoT with dense concentrations of edge-connected devices.
A later wave of 5G will provide ultrareliable, low-latency communications (URLLC) to support mission-critical applications for public safety and other mission-critical communication users, enhanced vehicle-to-everything (V2X), non-terrestrial networks (ie, 5G over satellite for extended and back-up coverage), and Industry 4.0 applications, such as highly automated factories with dynamic robots, self-healing smart grids and perhaps autonomous vehicles.
When does Ericsson expect to release URLLC solutions based on 3GPP Release 16 standards, given the R16 freeze date of March 2020? “Traditionally it takes 18 months for our industry to develop products after the standard has closed,” says Javier Garcia Gomez, chief technology officer for Europe and Latin America at Ericsson. “But what we have seen with 5G is that we have shortened the cycle quite dramatically. My expectation is that we will start trial applications of URLLC by the end of 2020 and then see a more wider implementation in 2021.”
URLLC has two chief characteristics, according to Gomez. “What we are doing is creating new signalling control channels with what we call mini-slots, where you can basically signal to the network faster. Then you have an additional wave to improve the reliability of the channel that you are using, so you will have latencies that go beyond the five milliseconds end-to-end. You will also have assurances that even in very hightraffic conditions, the channel will never be compromised.”
New use-cases
5G, and URLLC solutions in particular, will support a wider range of use-cases and business models than previous cellular standards, including the Industry 4.0 revolution. If realised, these use-cases will reach far beyond the traditional MNO business case of supplying mobile broadband and voice services to consumers.
5G threatens to disrupt traditional MNO models. Network slicing, the availability of more shared spectrum, licensed spectrum for private 5G networks, and the use of artificial intelligence (AI) to enable self-healing, self-organising and self-configuring networks should make it easier for other players to enter the RAN market and challenge the dominance of MNOs.
Who will these players be? “It’s a very good question,” says Gomez. “It will vary depending on the type of industry and geography.”
Gomez thinks it unlikely that many industrial players will build their own networks. “The depth of expertise required is not the core competence of industry.” Specialist connectivity partners will still be necessary for many years, it seems, even if AI and other innovative solutions make network management easier. “But what I cannot say is who those partners will be,” says Gomez. “Will it be an operator, or will it be somebody coming from the industrial side who has specialised in building connectivity networks for industry?
“For an operator it’s really a challenge. They know how to build wide-area networks for consumers, but building a network for an industry is quite a different thing. You have an indoor environment, a lot of [signal] reflection, a lot of machines – a lot of characteristics that you have to respect.”
Gomez points out that mobile operators will have to invest in learning what type of network deployment each specific industry will need. “But do they have a model that can scale from car manufacturing to shoe manufacturing?” he wonders.
This dilemma is similar to the one facing the IoT sector where each different use-case requires a bespoke solution, which is difficult to replicate in another market.
Yet Gomez is sure that 5G networks will be required to deliver the necessary levels of digitalisation if the ambitious goals of Industry 4.0 are to be met. Traditional cost-cutting measures will not provide anything above marginal gains any more.
“Most of the industry people we talk to see the only way they can improve productivity by another 20 to 30 per cent is if they go to a new model of automation where the robots are dynamic. This way they can have flexible product lines, which they can change according to demand,” says Gomez.
The idea is that highly secure, available and resilient wireless solutions with low latency and high throughput will replace cables to enable fully automated factories. 5G connectivity will support smart machines, dynamic robots, predictive machine maintenance, remote management with advanced QoS features, precise location services, rapid material handling and detailed asset and ambient monitoring of the factory floor.
Gomez does not believe Wi-Fi will provide an adequate connectivity solution for Industry 4.0 (although he concedes it might one day). He argues that Wi-Fi, being a shared medium, is too prone to interference. In addition, Wi-Fi’s Listen Before Talk (LBT) protocol can introduce delays, particularly in a congested RF environment. “If you have a factory with thousands of personal devices and everyone is trying to communicate to the network, controlling the latency of those devices is uncertain,” points out Gomez.
If the factory has robots which have to be carefully co-ordinated to ensure that the direction and timing of the programmable logic controllers (PLC) – the brain of the robot – are directed towards the factory’s goal, product or process, you need very low latencies and guaranteed connectivity.
“You have a latency budget, which is truly short, typically 5-10 milliseconds,” explains Gomez. “If you have interference, or you have too much load on the radio network, then you may exceed the latency budget you have to communicate with the robot’s PLC. The robot will stop and that is something the factory cannot afford. 5G will guarantee the latency, even in a highcapacity situation. That’s the fundamental difference [compared with Wi-Fi],” says Gomez.
Fresh offerings
One way MNOs can offer industries a dedicated service is network slicing, whereby the operator provides a portion of its network to a particular business or application. But how easy will it be for MNOs, wireless equipment and software vendors to fulfil the required servicelevel agreements (SLAs) for mission- and business-critical users?
“There are many technical things you can do, but you have to translate it to a certain level of investment in the purchase life,” observes Gomez. He cites the FirstNet mission-critical broadband network which Ericsson is helping AT&T to build in the USA. “That network basically uses Band 14 (700MHz). But that band is shared between police and consumers, so you need to ensure police, firemen or ambulance crews will always have preferential access to a channel.”
But he points out that if a firstresponder is in an area where they are depending on one radio and it fails, they will have no service. “So, what do I need to have? Maybe duplication of the boards in that radio, or you need to have two radios and then to have two channels, two possibilities to connect,” says Gomez. “It’s the same with the core network. The core is duplicated, so you always have two servers in two geographical locations, which are distant. Then you need to have a really dense conversation between the two core networks.”
In short, if you want an ultra-reliable, mission-critical network, you need high levels of redundancy in terms of functionality to ensure that in the event of equipment or power failure, the mobility management entity (MME) is able to securely switch transmissions to the other core in milliseconds, preferably without dropping any transmissions.
URLLC and Industry 4.0 use-cases will rely on having more computing power at the edge of the network to keep the latencies low. This becomes particularly acute in highly mobile situations such as moving vehicles and vehicle-to-everything (V2X) use-cases. Just how easy will it be to enable the ultra-swift radio handovers at the edge?
Gomez replies that this depends on the definition of the edge. Is the edge use-case fixed or mobile, local or global? Is it more appropriate to have the computing power locally onpremise or centralised remotely in the cloud? Ericsson believes collaboration is necessary between CSPs, public cloud service providers, tower companies, IT firms, operation technology companies and system integrators.
“The edge can be a regional data centre,” Gomez says, “or it can be a computing device that is embedded in the speaker. If the edge computing resource is as close as a few microseconds then we [the network] may be able to make the decision for you and prevent your car colliding with mine. But if the edge is really far away, then the cars will have to have autonomy.”
For the latter scenario there has to be a more efficient way of communicating between the vehicles in real time, notes Gomez. “That’s one of the things that the future of the 5G standard will solve: device-to-device (D2D) communication. We will see that in Release 17 [due for completion in June 2021].”
What does Gomez envisage 6G might bring? “It’s too early to say, but we will see a progression of the issues 5G is trying to solve. Fundamental challenges remain.” He points to the continuing need to deploy more computing power at the edge to enable new AI functions. “To do that efficiently, you need to improve the capacity density of the network and you need to improve the efficiency of the devices that we are now using.”
5G and 6G will also be using very high frequencies in the terahertz range, so Gomez says the industry will need to find new ways of connecting every device.
Further evolving D2D functionality is also key, says Gomez, taking us into an era of ubiquitous, high-capacity radio with even lower latencies and extensive Gbps capacity. This will support highly intelligent haptic IoT technologies at hyperscale; new forms of vehicle-to-infrastructure, vehicle-to-vehicle, vehicle-to-pedestrian and person-to-person connectivity.
