The low-power wide-area (LPWA) Internet of Things (IoT) market is still in its infancy in terms of actually realising the large-volume deployments confidently predicted for it over the past few years.
The overarching reason for this sluggishness is the sheer complexity of pulling together a full LPWA IoT solution, according to Professor William Webb, telecoms consultant and part-time CEO of Weightless SIG.
Webb believes the business case for adopting IoT is compelling just in terms of the productivity improvements, let alone the new revenue-generation potential. “The problem is everyone underestimated how complex it is to create the IoT stack,” he says.
“A farmer does not just want 50 tags for his cows. He wants a management system, data analytics and a dashboard providing easy-to-understand visuals that work on his PC. For that to happen you need all the bits in place, including a systems integrator on top to pull it all together, and that takes time because it is so complex,” argues Webb.
He cites a number of specific problems also bedevilling IoT, principally the multiplicity of LPWA wireless radio technologies. First, there are the technologies that use unlicensed spectrum including SigFox, LoRaWAN, Ingenu and Link Labs’ Symphony Link (based on LoRa, but with a higher-performance feature set) – all of which are proprietary to some degree – and the fully open Weightless Technology standard.
ZigBee, Z-Wave, Thread and 6LoWPAN are short-range technologies, but they can cover a large area using mesh technology, as can Wi-SUN, which uses the IEEE 802.15.4g standard. Wi-Fi is of course another possible unlicensed option. This proliferation of technologies has only got worse with the commercial availability of the 3GPP cellular LPWA standards of LTE-M (CAT-M1), NB-IoT (CAT-NB1) and EC-GSM last year.
Webb argues that for short-range connectivity applications, Bluetooth or BLE is the obvious standard as it is embedded in so many devices, while Wi-Fi is the obvious choice for smart home applications. “LPWA IoT needs to get to a place where there is the same obvious choice,” says Webb.
“But if you are an end-user today, what chip will you put into your cow tag?” he asks. “The trouble is, if you get it wrong, you may have a system that cannot be supported for very long. So, unless you are desperate you will probably wait for the de facto obvious chip.
“I’ve not come across many areas where it is necessary to move yet,” notes Webb. “Smart meters are perhaps an exception because their use has been mandated by government regulation in many countries.”
The LPWA opportunity
Vikrant Ghandi, IoT and digital transformation industry director at analyst Frost & Sullivan, notes that all the technologies are at different stages of deployment and real-world testing. “While evaluations continue over technical suitability and licensing approaches, there is definitely a growing need to standardise around a smaller set of LPWAN options to achieve the required scale and drive growth.
“SigFox and LoRa-based implementations do have an early mover advantage. However, we expect 3GPP-based deployments to eventually capture a majority of the market share in the LPWAN IoT wireless market in the next five to six years. The upcoming 5G networks will also help mobile operators increase the level of service delivery capabilities for the IoT,” says Ghandi.
“The opportunity in this market is large, and multiple network technologies will co-exist and collectively support the connectivity requirements,” he adds. However, at the moment he observes that customers want longevity in their deployments and right now there is some uncertainty on which technology to bet on. “Customers are looking for guarantees on this front.”
Lee Ratliff, senior principal analyst, connectivity and IoT, technology, media and telecom at IHS Markit, does see some market leaders emerging. “There are essentially four technologies vying for the LPWA IoT market: SigFox, LoRa, LTE-M and NB-IoT. There are others, but it seems likely the market is gaining consensus on which will be the high-volume runners. Even in those four it is looking like NB-IoT and LoRa will be the most popular.”
As to speculation that NB-IoT will kill off LoRa, Ratliff comments: “Maybe one day, but at the moment LoRa is highly differentiated and it can definitely coexist with NB-IoT, and both can grow at a fast clip.”
Nicolas Damour, director of technology partnership development at IoT module and gateway/router manufacturer Sierra Wireless, says the biggest difference in the LPWA IoT market compared with even two years ago is that cellular LPWA networks are now live. “I would say the Tier 1 European operators are about 75 per cent ready at the moment,” says Damour. “But if you are a potential customer, you can engage with the technology, as LPWA IoT devices have been available for a while now and customers are conducting tests.”
Nonetheless, both Ratliff and Damour believe the main barrier to the mass adoption of IoT is the lack of availability of nationwide public LTE-M and NB-IoT networks. “It is hard to market the technology when you have the chips, modules and end devices but not the network,” points out Damour.
But he thinks most of the mobile operators should have full networks up and running across Europe at least, if not by Christmas 2018 then in about six months’ time. However, there are a lot of regions of the world that have no 4G networks yet, let alone LPWA LTE networks, so this may keep the door open to unlicensed solutions.
LTE-M vs NB-IoT
Mobile Network Operator (MNO) strategy around LPWA IoT has varied considerably around the world. Webb observes: “I think there is some tension among MNOs about how far do they deploy bespoke proprietary solutions, which involves more cost, but maybe it is worth doing if there is a good revenue stream, or do they stick with cellular?”
That choice is further complicated by the multiple licensed LPWA solutions. “It didn’t help the cellular market having two [three including EC-GSM] technologies,” says Damour. “LTE-M and NB-IoT are in fact very complementary, but the very aggressive marketing that was done by MNOs in Europe didn’t help customers understand that.”
“I don’t see any consensus moving towards adopting one more than the other as yet, it seems to be about 50:50 split between the two at the moment, with some regional preferences – LTE-M in USA and NB-IoT in Europe, with some MNOs doing both,” observes Webb.
Damour says the only Tier 1 European MNO not doing NB-IoT is Orange (with the exception of Orange Belgium). “Orange is keen on keeping LTE-M, but that may be because they have a LoRa network as well. But it is hard to tell yet if the market will coalesce around any one particular technology, especially as it is still early days for LTE-M and NB-IoT.”
This is confirmed by a GSMA Intelligence report, The Mobile Economy 2018, published in February 2018, which stated: “To date, 23 mobile operators [out of a global total of around 800] have commercially launched 41 mobile IoT networks worldwide using the NB-IoT and LTE-M standards.”
The picture is somewhat confusing with the MNO community either hedging its bets or genuinely believing that there is a place for both unlicensed and licensed LPWA networks. But this makes sense to Ratliff, who says: “We are firm believers in hybrid networks and we think that MNOs will support several technologies including NB-IoT and LTE-M.”
LPWA revenues
But just how much revenue will there be in low-power wide-area IoT? “MNOs should ask themselves: how much money will I get from this?” says Webb. He cites Europe’s first flat-rate service of 500Mb for every IoT device over 10 years for a one-time fee of just €10 announced by Deutsche Telekom and 1NCE in February. This works out at €0.083 a month per device.
“That is nothing compared with smartphone ARPU for a consumer. Plus you have all the set-up costs and long-term service management costs, so I suspect operators are thinking revenues might not be great. So, for me there is a question mark over how extensively the MNOs will commit to the LPWA IoT market,” says Webb.
Damour agrees that the projected revenues and ARPU for LPWA IoT are not that attractive, but points out that MNOs have the opportunity to sell other services to their LPWA IoT customers, including consumer mobile phone SIM cards, broadband and mission-critical IoT services, which will be quite expensive subscriptions.
More competition
The MNOs, SigFox and LoRa networks are also facing competition from other sources. David Parker, senior analyst with Beecham Research, points to the largely France-based Wize Alliance, which promotes industrial IoT solutions using 169MHz VHF spectrum.
Wize is mostly aimed at wireless metering of water and gas, but has expanded into waste management and other smart city applications. GRDF, one of the largest distribution system operators in Europe, is committed to hooking up 11 million gas meters in France by 2022.
Then there is Wi-SUN, which uses the IEEE 802.15.4g standard to provide wireless mesh solutions for field area networks (FAN) for smart metering, distribution automation and home energy management. “It is being used for street-light monitoring in Paris, Copenhagen and Bristol, and it is also being used for HVAC systems,” says Parker.
Wi-SUN claims it has 85 million connections worldwide (LoRa is predicting 100 million by the end of 2018), making it one of the largest IoT standards globally. It may not be regarded as a wide-area network as it uses mesh technology with peer-to-peer jumping, but like Wize it targets many of the same large-volume LPWA applications as SigFox, LoRa and NB-IoT are aimed at, so they are serious competitors.
New business models
“LPWA is stimulating a lot of different business models,” observes Parker. He cites Senet in the USA which began as a propane and oil tank monitoring outfit for the residential heating sector but has morphed into a public network-as-a-service (NaaS) provider. It claims it now owns and operates the largest public, carrier-grade LoRa network in North America with cloud-based OSS and BSS platforms.
Outside of its own network area it offers RAN provider services to operators along with managed services for IoT (MNSi) – a cloud-based ‘turn-key’ M2M network service. This enables MNOs to deploy a LoRa network on their existing infrastructure, and Senet will operate and manage it on their behalf.
Parker also points to cable giant Comcast’s machineQ network. Comcast now includes a LoRa transceiver in all the connectivity assets it owns and operates, so it is almost inadvertently providing a LoRa network.
Comcast boasts it has “one of the most robust data backhaul networks and the largest Wi-Fi network in the US” and uses Semtech LoRa gateways and transmitters, STMicroelectronics microprocessors, Murata modules and Microsoft’s Azure IoT cloud platform for data ingestion and analytics.
Parker believes these developments point to a change in the LPWA market. “It is now more about which business model wins out and not so much which technology standard,” he asserts. SigFox has the handicap of having to pay upfront to roll out its network and only then sell its services. “LoRa doesn’t have that problem, as there are lots of companies building LoRa gateways, switches and sensors [although chipsets are still restricted to Semtech or its licensees].”
SLAs
Parker also thinks that service-level agreements will be as influential as the technology and business models in determining who wins business. Customers will first assess a particular technology for what Parker calls its essential attributes: battery life, coverage, transmit modes, message delivery, latency, scalability, data rates, security and so on. Then there are application-specific attributes such as whether it provides in-building coverage, roaming/ubiquitous connectivity and geolocation.
Also there are what Parker refers to as service attribute ‘wraps’. These include things like service availability, message delivery guarantees, order fulfilment processes, customer portals, help desks, fault discovery and reporting procedures, SLAs/KPIs – and, of course, pricing structures.
“What the market needs now is to understand what services are being offered,” says Parker. “This may be the antidote to a fragmenting technology market. The service wrap around the technology takes away the worry for the customer. Customers may be prepared to pay £2 a month, rather than £1 if they get good SLAs in return.”
Ratliff believes the barriers to mass adoption for IoT have moved up the value chain to the IoT platforms and the data analytics in particular. “How to deal with the mass of data, make sense of it and draw useful conclusions is the problem. Solving those issues needs
co-operation. Somebody needs to take the data and put it into a format that is usable. How do you deliver an automated irrigation system that can water a field on an as-needed basis? It is no use just having the soil moisture data if you don’t have a solution that can capitalise on that data.
“The solution has to be able to abstract the data and use it in real time to ensure the automated irrigation system irrigates the one part of the field that needs it more than other parts, or turns the water off if the system is turning another area into a bog,” argues Ratliff.
Damour believes that large, experienced customers now have the capability to deal with all the elements that make up the full IoT stack. “But overall the integration and complexity of the IoT stack remains a challenge, so there is a lot of value in trying to provide and market unified end-to-end solutions to make it easy for smaller customers.”
The LPWA IoT market is still finding its feet and is not yet clear which radio technologies and business models will win out. Security remains a major issue too. But it will be a few years before the market begins to mature and the winners and losers are revealed.
“The predicted connectivity numbers are right, but IoT adoption is not a consumer-style hockey stick adoption profile. It will happen, but it will just take longer than predicted,” says Webb.
What impact will 5G have on LPWA IoT?
Frost & Sullivan’s Vikrant Ghandi believes the IoT market will benefit significantly from the introduction of 5G. “The key benefits will be: programmable service delivery; network slicing capabilities; support for a massive number of IoT endpoints; high bandwidth and low latency; and edge computing capabilities.”
Weightless SIG’s Webb believes 5G will make up only a small proportion of the overall number of IoT connections. “It requires huge amounts of capacity and lots of hardware at the edge, and those use-cases will be very expensive.”
IHS Markit’s Lee Ratliff adds that it is the low latency required for vehicle-to-vehicle (V2V) and vehicle-to-everything (V2X) that is important in IoT for 5G. “You need 1ms latency to communicate to your car that the one ahead has hit the brakes. But a sensor in a farmer’s field detecting moisture levels does not need that kind of performance, and this is the kind of app that will make up the bulk of IoT, and the technology for that is available now.”
