Engineers looking to connect their “thing” to the internet face a tyranny of choice when it comes to selecting a wireless protocol. From WiFi to cellular and from Zigbee to Bluetooth, the established technologies all have their applications. Devices participating in the Internet of Things (IoT) however, typically require very low power, low data rate and are widely distributed. When supporting these unique requirements, the plethora of established protocols are found lacking.
WiFi is power hungry, and despite the impressive efforts of groups like GainSpan, it remains problematic for battery powered devices. Cellular is wonderfully prevalent, but it requires every device to have a SIM card and an on-going subscription with a telco to support power-sapping features that are mostly irrelevant.
ZigBee promises adaptive, sprawling mesh networks with low power devices, ideal for the IoT. But manufacturers have discovered that battery powered Zigbee routers are unreliable. Practically speaking, extending a Zigbee network requires routing wired power or cumbersome solar panels, fundamentally undermining the network’s goal. Meanwhile the Bluetooth standards body (the Bluetooth SIG) recognised that a low power version of their popular standard could launch it into the IoT realm. They ultimately adopted Nokia’s Wibree and released it as Bluetooth Low Energy (BLE). Unfortunately, the core limitations on node count, pairing procedure and periodic connection maintenance all limit its widespread adoption as an IoT protocol.
A Paradigm Shift
Around the world many have recognised these and other shortcomings, and have rushed to try to fill the gap. A smaller number have realised that the fundamental limitations require a fundamentally new solution. The notion of a network router or coordinator or master is irrelevant to devices in the IoT. The ability to move megabytes of data is useless. The requirement to check in with the network every 20 seconds, or even every day, can be an energy-sapping killer for IoT devices. Complex protocol stacks, multi-antenna designs, SIM cards, repeaters, pairing and other concepts associated with WLANs and PANs are massive roadblocks to the world of cheap, prevalent, low-maintenance and quietly operating IoT things.
Enter the Low-Power, Wide-Area Network (LPWAN).
Enter a world of inexpensive base stations with kilometres of range.
Enter a protocol stack that can be implemented on the most basic of microprocessors. Imagine small wireless devices, measuring their local environment (be it car park occupancy, electricity meter reading or a bridge vibration signature) and intermittently transmitting a morsel of data for immediate propagation to the Internet, where smart applications collate and compute to produce the services of the future.
The good news is that that world is here. Telensa and Sigfox have developed an LPWAN based on Ultra Narrow Band (UNB). Link Labs and Actility's ThingPark use a LoRa based solution. These solutions are in the field and enabling real IoT solutions.
The bad news is that most IoT projects end up spending an inordinate amount of resources on building the LPWAN to support their own things. Having invested so heavily in the network, they then guard it jealously.
This leads to the situation we see now - one network per manufacturer.
It’s absurd when you think about it - Telensa wanted to deploy a smart street lighting project. They ended up deploying smart street lights, plus their own proprietary network called PLANet(R) to support it. VIMOC Technologies wanted to deliver smart parking solutions in New South Wales and in California, and ended up running extension leads out of nearby shops to support a proprietary network infrastructure. These pioneering efforts are highly commendable, and wonderful demonstrations of the power of smart city solutions. But they’re not scalable. They do not promote multi-vendor solutions. They do not achieve the goal of many cities participating in Smart City concepts - an efficient, vendor agnostic, open ecosystem of city data, shared with citizens and available for smart applications.
An Open Ecosystem
The solution is simple - vendor-agnostic networks. And fortunately, the world’s startups have delivered. Leading the pack are Sigfox and several LoRaWAN installations. Hot on their heels are: nWave and the Weightless derivatives; the upcoming cellular LTE-M; a low power version of WiFi (working moniker is IEEE P802.11ah); the Dash7 Alliance protocol; Huawei’s “clean slate” cellular CIoT; Ingenu (formerly On-Ramp) and probably a dozen others by the time I finish writing this article.
So which one is best? Naturally, the answer is “it depends”. Weightless is still the only truly open standard but may get gobbled up soon. Sigfox’s UNB technology is very lean but limited. LTE-M can leverage existing infrastructure but is still a year or two away. And Huawei's business practices may not be winning hearts but it is hard to ignore their shear muscle and willingness to use it. Ultimately the best choice is less about technological superiority and more about positioning in the market. If we narrow our attention to the immediate Australian market, things become a lot clearer.
The LPWANs that are technically ready to deploy in Australia probably amount to Sigfox and LoRaWAN. Sigfox has for years been plotting world domination and their website shows them closing in on Oceania via New Zealand. Despite several planned “announcements" and claims, Sigfox is yet to make a physical mark on the Australian continent. In the end, their strategy of appointing a national operator, which has worked so well in the small, dense countries around France, may be difficult to adapt to the Australian environment.
For a variety of technical and non-technical reasons, the LPWAN most likely to first find widespread adoption in Australia is LoRaWAN. Telstra announced a trial in Melbourne recently, and prior to that National Narrowband Network Communications (NNNCo) selected LoRaWAN to underpin their national IoT network. The budding NNNCo was founded by influential industry heavyweights David Spence and Robert Zagarella, and they are already set to launch a 50-100 square kilometre network in Sydney.
2016 is set to be an exciting and pivotal year for the Internet of Things, and Australia is playing an important role.
While those with vested interests in the status quo downplay the importance of LPWANs, the deployment of a national “thing” network will be transformational in the adoption and proliferation of IoT solutions in all industries.