Connectivity choices for IoT, particularly cellular-related, are changing rapidly with new options now coming to the market, writes Robin Duke-Woolley, the chief executive of Beecham Research. This report reviews these with reference to the European market and taking into particular account views expressed by Mohsen Mohseninia, the vice president of International Market Development for Europe at Aeris Communications, a long-established mobile virtual network operator (MVNO) based in San Jose, California that specialises in IoT connectivity and operates globally. Aeris manages more than 14 million cellular IoT connections, including within the auto sector, transmits over one billion IoT messages per day through its dedicated network and has operational connectivity reach in 190 countries.
Is 5G ready for IoT?
5G is very much in the news at present and IoT users may be under the impression this means the technology is imminent for IoT. This is not generally the case though.
At present, mobile network operators (MNOs) are focusing on getting 5G into the mobile market and that raises very different challenges. Essentially, although it is good to know what is happening with 5G and what it can do for IoT users, the timing is not right yet for designing it into new IoT products unless those products will be out in the field for the next ten years. That is the case for the auto industry because of their long product cycles but not typical of IoT devices in general. It is all about the economics versus the benefits and, at present, the cost of 5G modules is high. These costs will decline as production increases and 2022 is most likely to be when 5G becomes a real opportunity for IoT. First implementations are likely to be for applications using fast routers and gateways, where the demand for ever-higher data speeds is strong.
Is there a Brexit opportunity for IoT?
Border controls are an obvious area for increasing use of IoT. The UK’s departure from the European Union (Brexit) has brought into sharp focus the need for real time analysis of border control data in order to minimise trade disruptions.
This centres around freight and transport, including borderless entry to and exit from Ireland and high volumes of movements around Dover-Calais. A number of companies have been approached by the government to see if they can use technology to enable faster movement of goods across borders. Extra delay means longer queues building up quickly and these technologies can assist.
The other related area is applications using cameras for freight, which is sharply in focus now following recent tragedies with people smuggling and closed containers. But that is not just cameras. This is about collecting data from transit assets and being able to apply machine learning and artificial intelligence (AI) to pick up things for inspection. There will be a major drive as the government is putting increasing pressure on freight companies to make sure they have done their homework well. All of this means more broadband is required, all of it 4G not 5G.
How does NB-IoT compare with LTE-M for deployment?
Narrowband IoT (NB-IoT) has been in prospect for a while, but with later market entry than LTE-M. NB-IoT represents the first cellular-based technology that is aimed at static or semi-static assets rather than mobile ones and that means it is essentially lower cost than those alternatives focused on mobile assets. Battery life is also expected to be longer. As a result, there is some pent-up demand for NB-IoT. One of the main drawbacks of the technology is use for international solutions. At present implementing NB-IoT in different countries requires different variants – it is not a case yet of one device, many markets. There are even different variants of NB-IoT within countries, with different operators using different infrastructure providers. That means it may not be possible to easily move an NB-IoT device from one network to another within the same country, for example in the UK where LTE-M is not planned to be available.
The Aeris view is that NB-IoT is now late and that it may well become the solution that never takes off in the way that everyone was expecting. Time has moved on and the technology was not out there early enough. Major customers are now looking at LTE-M, because it provides better solutions for global deployment, rather than just locally. Many customers are thinking that NB-IoT has very good properties like good battery and low cost but they cannot deploy it globally. LTE-M is a natural evolution for many network operators of their current LTE deployments, and availability of modules for LTE-M vs NB-IoT at the moment is driving many to say they will deal with NB-IoT when it is ready, but in the meantime proceed with LTE-M. In general, LTE-M does give them a significant amount of the benefits that they would likely get with NB IoT, although not all. Power saving on LTE is quite similar. However, the ability of NB-IoT to fire and forget – user datagram protocol (UDP) vs transmission control protocol (TCP) – provides additional saving as most LTE-M applications tend to be TCP based.
LTE-M is designed for roaming, with NB-IoT originally designed for fixed assets. Version 2 of NB-IoT is looking to address that mobility issue but there is no particular timescale for that.
From a module perspective, it means that in some markets there will be LTE-M, in others NB-IoT. Both are essentially evolutions from 2G and 2G will be around in Europe for some time. As a result, some module suppliers are now catering for NB-IoT, LTE-M and 2G in one device, on the basis that devices can be switched over from 2G to either NB-IoT or LTE-M when those technologies become available locally. In effect, this provides IoT users with the ability to deploy globally with one module – one stock keeping unit (SKU) that is future-proof. This avoids the need for multiple SKUs for multiple regions and multiple networks, which is potentially a major saving of supply chain cost.
When will LTE-M take off in the IoT market?
LTE-M is already standardised and network implementations are underway. So when will LTE-M take over from 2G/3G?
The Aeris view is that in Europe we are already seeing network operators announcing a switch-off of 3G. For example in Norway and Switzerland, sunsetting of 3G is now planned for 2021. This is not the case for 2G though. For 2G there are a lot of long-term contracts – in particular government-run contracts – that are dependent on 2G availability and network operators cannot easily extricate themselves from those. So last year at Mobile World Congress the CEOs of Vodafone, T-Mobile and Orange all hinted at keeping the 2G networks up and running until 2025. There are no consumer handset technologies that will rely on these 2G networks, so in reality the 2G networks are becoming essentially IoT networks. The operators do not have to spend any money for additional capacity because most customers are looking at newer technologies.
That is in Europe. Looking then at Africa, the spending power of individuals (consumers) is not high, so low cost 2G mobiles are still around and will continue to be used in African communities. 2G will be in Africa for a long time to come, as it will in some parts of South East Asia and in some parts of Latin America. 3G is likely to go, but 2G will be around for a while. On the other hand, it is likely that network operators will start to switch spectrum more aggressively so there is unlikely to be further 2G expansion in coverage.
That means attention will focus more on LTE-M and NB-IoT over the next few years and from 2021 it will most likely be the main area of interest for IoT users. Large LTE-M device deployments are then most likely to start happening during 2022 and 2023.
In comparing hardware costs, LTE-M is expected to be significantly lower than 2G, with NB-IoT being lower cost than LTE-M.
Other technologies – Wi-Fi and satellite use for IoT
Beecham Research’s own research has shown that Wi-Fi use for industrial applications is currently very high for on-site use but is challenged in large sites with many access points and particularly where there is a need for outdoor use and for use with mobile assets. The use of LTE for private networks in both licensed and unlicensed spectrum is then becoming of greater interest. The opportunity to use LTE as part of a longer evolution towards 5G in private networks – which would utilise network slicing – is a likely future trend.
In the meantime, the overlap between Wi-Fi and cellular is being utilised by Aeris. One of the key issues with Wi-Fi is the security management of access, which is that customers or providers regularly change the service set identifier (SSID), user name, password, and those are the handicaps of using a Wi-Fi based solution. Aeris is combining cellular with Wi-Fi, so that through its platform and SIM it can remotely update the SSID, user names and passwords to make sure the Wi-Fi route is working. This is envisaged to be in residential premises, for smart home type applications, where the best method would be to use the Wi-Fi that is currently available, but have the ability to manage the credentials using cellular. As such this is a light use of Wi-Fi as opposed to an industrial use. Another area that Aeris is now offering is satellite. Which is receiving attention at present partly because of new, low cost constellations being planned. Satellite has a role to play for niche IoT applications, as does combining satellite and cellular together as a dual mode – using cellular when you can for lower cost and satellite when cellular is not available.
What is the opportunity for high bandwidth cellular?
Figure 1 shows projected total cellular IoT connections worldwide to 2025. Within this there are three key elements – high bandwidth (4G, 5G); legacy (2G, 3G); and low bandwidth (LTE-M, NB-IoT). As outlined above, 2G/3G growth is expected to flatten quickly during the period whereas 4G/5G will grow quickly and NB-IoT/LTE-M will also grow quickly as indicated above. Beecham Research projections are for continued overall connections growth of more than 30% per annum. This indicates the growing opportunity for 4G/5G as IoT applications become more sophisticated, requiring more data from more individual sensors and other sources to provide a better overall service.
Of growing interest in this area at present is use of 4G (LTE). This has opened up new possibilities for innovative companies to obtain data coupled with machine learning and AI. This is streaming data that provides for real-time (or near real-time) decision making. That means transporting huge amounts of data in almost real time and processing that data very quickly. That includes applications like facial recognition, vehicle recognition and number plate recognition among many others.
At the same time, processing data at the edge versus in the cloud has matured significantly. In future there will be more of a mix of the two and where the processing is carried out will depend on the economics and requirements of the particular application. The growing trend is to put more focus on what can be done at the edge before it is passed to the cloud for further processing. While that reduces the amount of connectivity for each application, it is also tending to enable more new applications that then require new connectivity.
Aeris Fusion Platform
Fusion is about building intelligent applications on top of the networks that help Aeris customers to aggregate and enhance the data that they get at the edge, together with additional sources of data that Aeris can provide that can then go to the cloud for further processing.
For example, Aeris is focusing on machine learning and AI. An example is for a fleet management company, where machine learning is applied to learning places. The network data shows a particular pattern of the assets that are in transit where there could be places of interest. This is viewed directly from the network not from the GPS the devices have. Aeris is already offering this to a number of customers in Europe as one of the micro-services available as part of Fusion.
A further micro-service is about location-aware over-the-air (OTA) upgrades. In future, Aeris believes that OTA will become ever more critical to operations so they have developed a service specifically to manage that. In the world of LTE-M, for example, they see problems that customers previously have not experienced when using 2G that they will experience when it comes to LTE-M. Chief among those is congestion on the channels at the cell sites, where there are only a limited number of channels available to communicate to the devices. If you have a large number of devices – say 5,000 – that are communicating through a single cell site, then an OTA update to those devices will most likely fail.
Aeris has built a micro-service that allows their customers to sequence their OTAs rather than broadcast them. OTAs can then be scheduled into groups of devices as sequential activities, rather than allat-once. This significantly improves the success rate of OTAs and, as a result, decreases the cost of the overall OTA activity. This will be particularly significant in the future because considerably more data will be downloaded in future as OTA updates – from network profiles, to application updates and security updates. Any time a customer needs to update their end point devices, the end point firmware itself, the configuration of the radio module for a new security update, or the need to change the behaviour of a device regarding the amount of data collected or where it needs to go to. Even adding a new capability to it. All of that will require an OTA. Particularly with the high volumes of low data rate LTE-M and NB-IoT connections envisaged, OTA will become a central activity to operations far exceeding what is used today, and IoT users need to take into account the implications of that.
That is the nature of the Aeris Fusion network. It is to add intelligence – a set of micro-services to help customers reduce their costs, improve their service quality and take advantage of the new technologies. It brings intelligence to the application layer.
Fusion is an intelligent, purpose-built network for IoT that enables innovation on top of cellular networks to add real business value to enhance intelligence, security and also information. It is a network that enables customers to apply machine learning, to do artificial intelligence, and data processing on top of that network and enables innovation through micro-services.