Understanding the Cellular and LPWAN Landscape for IoT Devices

When you’re designing an IoT network, you’ve essentially got two choices when it comes to how devices at the ‘edge’ connect to the internet: cellular wireless carriers and low-power wide-area networks (LPWANs). There are multiple LPWAN standards – some proprietary – while in cellular, the focus is on the LTE standard.

With wireless carriers and LPWAN providers charging for every device connected to their networks, it’s clear why IoT connectivity is such big business, and why there’s so much competition. IoT device numbers are forecast to hit 20 billion by 2020. Not all will link directly to the internet, of course, but it only needs a fraction of them to do so for wireless service providers to be looking at huge revenues.

As an IoT designer, it’s important to understand the ins and outs of cellular and LPWAN, their respective strengths and weaknesses, and when to use each one.

Different applications, different requirements

IoT applications vary enormously: connecting up a city of 100,000 residents with smart utility meters is a very different challenge to connecting up a factory with sensors. And that’s before we even consider the unique IoT environment that autonomous vehicles will create. Given the variety of applications, it’s not surprising there’s no one standard that perfectly suits every need.

Figure 1: Already common in the USA, smart utility meters send usage updates over-the-air, typically via LPWAN, but will soon also be using cellular.

That said, there are some broad similarities between what LPWAN providers and wireless carriers offer for the IoT. Ultimately, the aim is to enable remote sensors and actuators to communicate with other devices, using so little energy that they can run for years at a time on a coin cell battery. To minimize energy consumption, LPWAN and wireless carriers employ a range of techniques, including limiting how much data is transmitted, limiting how long each transmission lasts, and using very low data rates that only need narrow bandwidths.

Both also require very sensitive base station receivers, given the weak signals transmitted by the sensors. The base stations also use techniques such as Multiple Input Multiple Output (MIMO, see figure 2) and directional antennae to help maintain constant connections.

Figure 2: MIMO takes advantage of multi-path propagation to boost network capacity and minimize errors.

Lastly, there’s the need for large numbers of base stations (the ‘small cells’) to minimize how far signals must travel. This, in turn, enables designers to deliver the near-instantaneous communications that certain IoT applications require.

Cellular and LPWAN: The current state of play

With widespread LTE coverage in the USA and elsewhere, enabled by a huge network of macro base stations and many more small cells, the cellular industry is already in a strong position when it comes to the IoT. What’s more, a software upgrade (as opposed to an expensive hardware change) is generally all that’s required to get this infrastructure to work with IoT devices. And let’s not forget that even in the days before the IoT was a widely-recognized term, cellular carriers were linking up earlier generations of wireless sensors with 2G technology.

There’s been a big push in the cellular space to accommodate the IoT, not least in 3GPP Release 13. With IoT-related development continuing as 3GPP prepares the standards for fifth-generation cellular, it’s clear that wireless carriers have an ever-strengthening foundation on which to offer IoT connectivity.

LPWAN, on the other hand, is building from the ground up. If a provider wants to offer coverage, it first needs to put in the infrastructure – and time is of the essence if it’s to compete with cellular players’ IoT offerings. That said, LPWAN systems are typically cheaper to deploy (partly because they don’t necessarily need to lease tower space), and are able to cover a larger area using a smaller number of base stations.

Analysts are generally optimistic about the prospects of LPWAN, with some suggesting more than half of all IoT applications can be served in this way. So even though cellular will remain a big player, LPWAN will likely have its place too, with price wars highly probable in some markets.

At a glance: Cellular IoT choices

The cellular industry is developing IoT solutions based on LTE, refining its capabilities, cutting complexity and driving down cost. This will make cellular a more attractive option for a greater variety of IoT uses, culminating in the launch of 5G in the next few years.

To help achieve these goals, three standards are coming to the fore: LTE-M, NB-IoT and EC-GSM-IoT. To help signals travel better over longer ranges and penetrate into buildings, solutions should be implemented at frequencies below 1 GHz.

LTE-M, or enhanced Machine Type Communication (eMTC), is an evolution of the LTE standard that’s found in 3GPP Release 12. NB-IoT is a narrowband variant of LTE, specifically for the IoT. And Extended Coverage GSM for IoT (EC-GSM-IoT) is an IoT-optimized version of GSM.

5G will further enhance both NB-IoT and EC-GSM-IoT. It’s anticipated these future developments will focus on improving battery life, cutting complexity and reducing device cost. Other focuses are expected to include lowering deployment costs through carrier-capacity sharing, and improving coverage, partly by increasing the signals’ spectral density and partly through more sophisticated coding.

At a glance: LPWAN options

In the LPWAN space, there’s a choice of open standards (such as LoRaWAN) or proprietary solutions (including Sigfox). Both Sigfox and LoRaWAN use the unlicensed spectrum. Sigfox claims it’s ‘the world’s leading IoT services provider’, and offers connectivity in around 50 countries. At the same time, LoRaWAN, with its 500+ alliance members, has the widest industry acceptance. This scale is helping drive down the cost of LoRa RF and baseband hardware: it’s already fallen by more than 50%, and this trend will continue as volume rises.

LoRaWAN

When discussing LoRaWAN, it’s important to understand the difference between LoRa, LoRaWAN and offerings such as LinkLabs. LoRa is the physical layer of the LoRaWAN standard. LoRaWAN is the media access control (MAC) layer that delivers the networking. And LinkLabs is one of the suppliers offering IoT connectivity services using LoRa technology.

Sigfox

Unlike LoRaWAN, Sigfox is entirely owned by the French company of the same name. All technology in a Sigfox network, from the edge to the server and endpoint, is also owned by Sigfox. It effectively supplies the entire ecosystem, or may even take the role of network operator.

Any organization that agrees to Sigfox’s terms can use Sigfox endpoint technology for free. This is one way it’s built up relationships with IoT device suppliers and wireless carriers. And it’s gaining market share, especially in Europe, where its transmission length meets EU guidelines. To meet US FCC rules, Sigfox in the USA is very different.

Weightless

Sigfox and LoRaWAN aren’t the only shows in town, of course. Weightless is an open standard, managed by its own special interest group. Designed as an exceptionally lightweight protocol, a transmission can contain just a few bytes of data. This makes it ideal for IoT devices that only need to send or receive very small amounts of information, such as utility meters or certain medical and industrial equipment.

Operating in the sub-1-GHz spectrum, Weightless typically delivers wide coverage with low transmit power, as well as the ability to penetrate structures that generally pose RF challenges.

Weightless comes in two flavors. Weightless-N is a uni-directional, ultra-narrowband technology. And Weightless-P is bi-directional, offering carrier-grade security and performance with very low power consumption.

Nwave

Nwave, an ultra-narrowband technology, uses software-defined radio (SDR) techniques capable of operating in both licensed and unlicensed frequency bands. You can connect up to a million IoT devices to a single base station, over a 10km range, with RF output power no higher than 100 mW and data rates of 100 bps. And when you operate in sub-1-GHz bands, Nwave benefits from the favorable propagation properties in this region.

Ingenu RPMA

Ingenu is the company behind RPMA, which is designed to deliver secure, high-capacity, wide-area connectivity, using the 2.4 GHz band. A single access point can provide up to 176 square miles of coverage in the USA – much more than LoRa or Sigfox. Latency is low and commands can be sent simultaneously to large numbers of devices. However, since RPMA is proprietary, hardware and software all comes from Ingenu.

Making the right choice

Even on this quick tour of the main LPWAN options for the IoT, the differences between the technologies – and therefore the importance of choosing the correct one for your application – become clear.

Part of the choice will come down to which technology is available where the devices will be used. Your budget might also shape the decision: how much will it cost to connect each device in your design? And that’s before you get into the technical ins and outs of each technology. 

To help make the right decision, map your technical and commercial requirements against the capabilities of the available technologies. This will quickly help narrow down your options, leading you towards the best wireless technology for your IoT devices.