It is hard keeping track, because there are just so many possible options - this article aims to give a comprehensive overview of the emerging and established wireless technologies relating to the IoT.
This article aims to give a comprehensive overview of the different wireless technologies relating to the IoT – covering both emerging and established protocols.
Bluetooth 5 launched back in December 2016. It incorporates Bluetooth Low Energy (BLE) mode, and introduced an optional new interface that doubled the previous top data rate to 2 Mbps. Its faster transmission has meant there is a need for higher transmission power (+20 dB) in low energy mode – power is still conserved overall, however, as the data takes considerably less time to transmit.
Using the 2 Mbps mode is likely to reduce the range, but as Bluetooth 5 has a theoretical range that is roughly four times that of Bluetooth 4.2 BLE (up to almost 244 m), it still represents a major improvement.
The standard has now elevated beacon performance markedly. Though previously the data payload was limited to 31 Bytes, Bluetooth 5 supports a payload of up to 255 Bytes/packet. This means that beacons can send more information within a single broadcast message, so the user’s device doesn’t need an app or an Internet connection to access it.
Bluetooth 5 presents an eight-fold increase in broadcast efficiency, thus being capable of supporting far more devices. Beacons can, as a consequence, be used to cover whole buildings – for both domestic and commercial automation systems. Bluetooth 5 may also support low-quality video and audio streaming, plus short bursts of high-volume data (such as for transporting images).
Thread mesh networking
Thread was built to support robust IoT infrastructure, and was intended to become the standard for the smart home. Progress on this low-power mesh networking technology has been somewhat disjointed, though. The Thread Group, set up in 2014, attracted big names, but without gaining the corresponding market traction. Last August, Apple joined the group, and it was widely assumed that the company would include Thread in its HomeKit accessories and the devices that control them. However, there is no news on that front as yet.
The advantage of Thread’s mesh approach is that if one node fails, the other nodes can still connect to each other. Also, it can support over 250 devices, in multiple hops, and is so power-efficient that devices running on AA batteries will last for years.
This means a Thread network of battery-powered devices still works during a power cut, which would be a boon for security and monitoring applications. This is not least because all of Thread’s transmissions are encrypted end-to-end, adapted for low-power devices.
ZigBee and Z-Wave incorporate AES encryption too, but Thread also uses banking-class, public-key cryptography to back it up.
In January 2019, the ZigBee Alliance and Thread Group completed the Dotdot 1.0 specification and the Dotdot-over-Thread certification programme to make the two rival networks interoperable and accelerate their uptake in the currently fragmented home automation market.
Then in February, Nest was the first of Alphabet’s ‘Other Bets’ to be folded back into Google to become part of the hardware division that makes Pixel smartphones and Google Home speakers. Rick Osterloh, head of that division, said of the move: “The goal is to supercharge Nest’s mission: to create a more thoughtful home … built with Google’s artificial intelligence and the Assistant at the core.”
Thread’s mission is to enable devices to talk to each other without an intermediary hub, such as Google Assistant. Nevertheless, it is likely to remain part of the mix as it has a lot to offer, and Nest (which supplies two of the few products that directly support Thread) communicates with devices using a combination of Thread, Wi-Fi and Bluetooth.
It is true that only a handful of products have Thread built in – this is by design, though, as any device based on the 802.15.4 protocol can download software to gain support for it, and so it is not necessarily a hindrance.
LoRa stands for ‘long range,’ and like most of the pioneering, non-cellular low-power wide-area network (LPWAN) protocols, LoRaWANs run in the industrial, scientific and medical (ISM) spectrum band. These frequencies are reserved internationally for organisations in those sectors, and are free to use, as well as being positioned to avoid interference from mobile communication networks.
Typical applications are asset tracking, supply chain, agriculture, smart cities, intelligent buildings, home automation and smart metering. Installations do not need wiring or a power source, as the devices run on batteries – which can last up to two decades.
The LoRa Alliance is responsible for updating and defining the LoRaWAN protocol to ensure interoperability among devices and networks – it is a not-for-profit, collaborative association with more than 500 members involved. The alliance claims LoRa deployments increased by over 60% in 2018, with almost 80 million LoRa-enabled end nodes being installed over the course of the last year.
LoRaWANs are widely implemented within buildings, as they can penetrate dense building materials – reaching into basements and other below-ground-level locations – and have a transmission range of up to 10 km. The integral AES-128 encryption makes them more secure than Wi-Fi, and fewer gateways are needed than with Wi-Fi too (because a single gateway can cover a whole IoT system spanning an entire building or underground parking lot).
As LoRaWANs rely on different frequency channels and data rates by encoding packets, messages are less likely to collide, which increases the gateway’s capacity. One gateway can support millions of messages, making LoRaWANs suitable for public network operators simultaneously attending to many customers.
Critics say this is not really an open system – as the LoRaWAN stack relies heavily on chipsets from Semtech (which bought the technology from French company Cycleo in 2012). The proprietary nature of many of the established LPWAN protocols (such as LoRa and Sigfox) could count against them in the longer term, now that cellular IoT technologies have started to come onto the market. They have been developed by the telcos, and thus have the might of these companies (and their expansive global infrastructure) behind them.
LTE (Cat-M1) is usually referred to simply as LTE-M. It is the first LTE-based protocol designed for low-power, low-cost IoT applications, and employs the 1,4 MHz (as opposed to the 20 MHz) spectrum. Its average upload speeds are between 200 Kpbs and 400 Kpbs.
Using a transmit power of 20 dBm, batteries can last for up to ten years. Nodes can ‘deep sleep’ while in power saving mode (PSM), but remain registered with the network, or ‘wake up’ periodically – referred to as extended discontinuous reception (eDRX).
Numerous LTE-M industrial IoT use cases are currently being explored – including connected vehicles, fleet/asset management, smart pallets, container monitoring and smart shelving. In some instances, devices transmit many times daily, in others only once. For certain applications, the device will only ‘wake up’ when a threshold is reached, such as a predefined temperature.
End devices connect to the network without a gateway, thus reducing cost. Furthermore, operators do not have to replace antennas, just update their software. End devices come with lower price tags than ‘full’ LTE devices, because the chips are cheaper to make – being half-duplex for narrower bandwidth.
The service costs less too. This is down to the low bit rate and the fact that periodic traffic takes up little network capacity. Another deployment option could be using short-range connections, like Bluetooth, for asset tracking backhauled over LTE-M.
NB-IoT (Cat-M2) has similar goals to LTE-M, but uses low-bandwidth signals to communicate within GSM and LTE networks by exploiting bandwidth that is unused at any given time. Possible applications include smart parking, tracking livestock, smart metering, retail, vending machines, fire sensors and smart lighting, as well as the monitoring of pollution, soil acidity and moisture levels.
An advantage that LTE-M has over NB-IoT is that it supports higher data rates and greater mobility. In addition, it can carry voice over the network. That said, it needs greater bandwidth and costs significantly more.
The 5G future
LTE-M and NB-IoT are late arrivals compared to the non-cellular pioneers, but they are progressing fast. The GSMA claims that cellular LPWAN technologies are future-proofed for 5G communication – which should be regarded as a ‘network of networks,’ rather than a complete new overlay (like previous generations of mobile technology). They will support 5G and co-exist with it.
Conflating the availability of globally standardised cellular LPWANs and 5G is a good marketing ploy, reinforcing the notion of ongoing continuity – and that those networks will be here for decades to come. This is a tough argument for the proprietary technologies to counter.
ABI Research reckons that cellular and non-cellular LPWAN connections will ramp up globally at a 53% compound annual growth rate (CAGR) until 2023, driven by demand for smart meters and asset trackers (neither of which need 5G). From then onward it is anticipated that the prevalence of non-cellular LPWANs will wane, with NB-IoT and LTE-M accounting for around 55% of connections. Even so, assuming there is no sudden, massive collapse of confidence in the non-cellular technologies, they are likely to operate alongside cellular IoT networks for years.
The effect of advanced cellular technologies on IoT is hard to accurately predict. Many IoT applications do not need the attributes and (at least initially) the additional costs that will be associated with 5G. Others, like autonomous vehicles, will be dependent on its ultra-low latency and the flexibility that comes through network slicing. Making viable business cases will be critical and – as ever in technology – reality will take some time to catch up with the hype.