Networks get care in the community

5 mins read

Today the IoT revolution is heavily reliant on short range devices so what can we expect from wide-area IoT solutions in the coming years?

Early predictions about the internet of things (IoT) certainly did not lack ambition. By the start of this decade, 50 billion devices were meant to be ticking away, measuring things and reporting their changes over the internet. As with many hype cycles, reality has turned out differently though it is not an order-of-magnitude discrepancy.

A survey by Ericsson published a year ago put the number of IoT connections at 14.6 billion. To a large extent, from wearables to environment sensors, the IoT revolution has relied heavily on devices using short-range Bluetooth Low Energy (BLE) connections as a way to get online. If you've attached something like an Apple tag to a set of keys and they are still nearby, no problem. BLE will give you a way to home in on their position.

Whereas Bluetooth and WiFi are near ubiquitous indoors if someone has an internet connection, network access beyond the garden fence has remained far less prevalent. The telecom giant’s survey estimated a mere 2.1 billion are accessing their hosts over wide-area network (WAN) connections, almost all using some form of cellular modem.

One important reason is that using wireless WANs presents a more difficult set of decisions: often none of the individual choices are entirely satisfactory. That problem has been worsened by a seemingly continual cycle of investment and retrenchment with, now, even popular choices such as SMS over 2G cellular now in danger of being retired. Despite this, Ericsson’s researchers believe the growth for wide-area IoT will be stronger in the coming decade than for short-range IoT devices: 19 per cent versus 12 per cent.

A clearer situation

In some ways the situation is clearer than it was a decade years ago where in addition to various forms of proprietary sub-1GHz narrowband protocols, choices included On-Ramp, SigFox, different flavours of LoRa such as Link-Labs’ Symphony and creator Semtech’s own LoRaWAN as well as the "white space" radio concept that proposed the idea of sniffing for free spectrum dynamically and popularised, at least for a while, by the Weightless-W protocol. Most of these still exist in some form or another, but a few have pulled ahead while others struggle.

SigFox's French owner filed for bankruptcy protection early this year, for example. Though Singapore-based UnaBiz bought the business with plans to support its remaining activities many analysts doubt the network's ability to scale in the way that is needed to support it long-term.

On-Ramp rebranded to Ingenu and has even begun a foray into space with plans to use 70 satellites in low-earth orbit to expand coverage. As more low-Earth constellations are filled out with an intense programme of launches, satellite access will open up another set of choices for designers.

Proprietary narrowband remains an option and suits projects where the supplier or user wants to maintain end-to-end control of the system and can afford the buildout to support it across a campus. For those who want to use public networks and standards, even geopolitics is playing a role. Because of Chinese manufacturer Huawei's involvement in the development of the protocol, the 5G cellular-based protocol NB-IoT has met greater resistance recently, which may focus more attention on the main cellular alternative LTE-M.

Though French cellular operator Bouygues Telecom decided to abandon plans to build a LoRaWAN-based IoT network in the country and instead switch attention to the 5G protocols, the low-power WAN protocol developed by Semtech has gained traction in several countries, achieving high coverage in cities such as Amsterdam and Berlin through a combination of crowdsourcing and private network operators that agree to act as peers for each other.

This form of roaming has helped operators such as The Things Network obtain greater coverage than would be possible if they had to invest in installing gateways by themselves.

The Things Network has over the past year introduced a packet-broker and an onboarding service that is intended to make it easier to not only exchange packets across private and public operators' networks but migrate between service providers.

Crowdsourcing approach

The crowdsourcing approach that LoRaWAN encourages has enabled a few experimental applications that have been used to test out different IoT concepts. At the Things Network conference in Amsterdam in September, Thingitude founder Mark Stanley, described how the company moved from a trial of systems to help prevent falls in the home by older people to a personal-safety network centred on the University of Reading campus but which extends across a large part of the city's inner region.

Thingitude fitted 120 streets with radio-connected sensors talking to council-supported LoRaWAN gateways that track activity and light levels, feeding the data to an app that students can use to plan journeys home at night. Of the 900 publicly accessible LoRaWAN gateways running in the UK, more than 10 per cent are in Berkshire.

“One of the hopeful effects was that the council would be interested in the data for managing streetlights or road planning,” Stanley says.

However, coverage of all networks, from crowdsourced LoRaWAN to cellular is far from ubiquitous. Even in Berkshire, population coverage hovers around 90 per cent, with gaps in more rural areas. The difficulty of guaranteeing wide-area coverage presents problems for designers of all kinds of IoT devices that need to be mobile, whether they are panic alarms, livestock trackers or in one example used by design house IRNAS at the Things Network conference: the expensive audio and lighting rigs needed for live concerts.

“We need to build solutions which [support] multiple wireless technologies,” says Luka Mustafa, founder and CEO of IRNAS.

Within a warehouse and even a supplier’s own trucks, one option is to use Bluetooth beacons to track individual boxes or pallets. Over time that might expand to ultrawideband technologies, which deliver better localisation. But what happens with rented vans or when the gear gets to the venue? That might mean access to LoRaWAN if coverage in the area is good and possibly adding NB-IoT on top. Adding radio transceivers adds cost, so it becomes difficult to justify fitting to every item. However, one expensive item could relay Bluetooth signals from devices it normally operates with: sound-reinforcement speakers may just report their continued presence to a central amplifier or mixing desk.

Sidewalk from Amazon

Amazon's attempt at building a ubiquitous neighbourhood network follows a similar philosophy. If there is no one radio standard that can do the entire job, why not combine them?

The delivery giant's Sidewalk pushes Bluetooth, narrowband operating with frequency shift-keying (FSK) and the physical layer from LoRaWAN into its smart speakers and doorbells so they can act as gateways to a household’s back-end internet connection.

Like Link-Labs’s and its Symphony Link, Amazon opted to manage the mesh itself through its devices instead of supporting the full LoRaWAN stack. Amazon has asked the Federal Communications Commission to approve a gateway design it will use to expand the network beyond what is possible with its consumer gear.

Silicon suppliers such as Silicon Labs have put together toolkits for OEMs to build their own devices that can connect to Amazon's back-end systems. Systems do not necessarily have to use all three radios: much will depend on the application.

“You could just have BLE for a pet tracker. Or opt to use FSK as well if the pet is likely to roam around but just around the neighbourhood,” says Dhiraj Sogani, senior director of wireless product marketing at Silicon Labs.

The question is whether a system that relies so intensely on a single supplier and its infrastructure-spending plans will win out against competition where a monopoly is not baked in from the start.

Though last-mile wireless remains one of the trickier choices for IoT implementors, the market may finally be moving out of its decision-paralysis stage to one which in which it is becoming clearer which will win long-term in both service quality and market penetration.