451 Research: MWC 2015 – the brewing battle over low-power wide-area IoT networks

InsightaaS: Not all IoT traffic is created equal – or so it would seem from the explosion of interest that is building in low power, long-range radio technologies that can take advantage of underused spectrum to support IoT applications. In the research report below, 451 Research analyst Brian Partridge delivers the buzz from Mobile World Congress 2015 on low power IoT WANs, including analysis of key players and emerging business models, and most importantly for the broader IoT ecosystem, why this “brewing battle” matters.

As Partridge explains, while low bandwidth IoT applications such as utility meter reading or industrial monitoring have relied on 2G/GPRS cellular, satellite, and fixed networks or distributed mesh network systems for connectivity, and high bandwidth applications such as video surveillance have required support from 3G/4G networks, there are issues with the latter – “maintaining device clock synchronization and the asymmetry of power consumption” – which put strain on operational costs in high bandwidth scenarios. As MNOs wait on LTE networks to solve for these issues, a new variety of IoT WAN technologies has appeared that allow mesh networks, where devices work together sending data from one onto the next and on to the next, to transform into star networks that enable the device to connect directly to the cloud or to an aggregation point, an approach that essentially reduces power consumption.

The evolution of the IoT space is apparent in the range of activity around these networking issues: Partridge discusses prospects for individual players such as SIGFOX, which uses Ultra Narrow Band (UNB)-based radio technology to connect devices to its network at a much lower cost than even 2G and recently received VC support from many of the key European telecoms, as well as the LoRa Alliance, which is developing LPWAN standards and looking to create an ecosystem around the LoRa protocol (IBM and Cisco are on board), the Huawei-supported Weightless Special Interest Group that placed its bet initially on abandoned analog TV whitespace, but has since built support for devices to work across a variety of unlicensed and licensed spectrum bands, and other groups. Why does this emerging battle around Low Power Wide Area Network (LPWAN) matter? Will ongoing standards warfare around long distance networking protocols hamper broader deployment of IoT applications, or as commentators at the Mobile World Congress suggested, are these emerging technologies simply a “placeholder” for LTE? Partridge’s analysis below provides an insightful and detailed look at the various emerging network technology alternatives, which should be required reading for MNOs and IoT implementers alike. It’s never good to fall on the wrong side of the standards divide. (ed.)


Brian Partridge, VP, research & consulting, 451 Research
Brian Partridge, VP, research & consulting, 451 Research

As expected, the Internet of Things (IoT) was a dominant theme across Mobile World Congress 2015. While the full glitz and glamour of new connected cars, more useful wearable and connected devices, and smart cities were on full display, there was also serious discussion surrounding the less sexy networking plumbing of it all – off to the side. At issue: the growing potential for alternative low-power, low-cost, ‘SIM-less,’ long-range radio technologies to supplant or augment 3GPP licensed alternatives and mesh connectivity architectures. The $115m funding round announced in February by low-power IoT specialist SIGFOX (which included the support of three top-10 global tier one mobile network operators) has sharpened focus and debate on this topic. Many in the traditional 3GPP infrastructure crowd would like to believe that LTE-M can solve a majority of the low-power/long battery life challenges in the long run. Telecom operators aren’t so sure, and are making some bets behind a wave of disruptors led by SIGFOX, the LoRa Alliance and Weightless SIG. These players make a strong case that entirely new architectures and standards are the way forward to support a huge volume of IoT devices that demand low cost, sip bandwidth, and support several years of battery life.

The 451 Take

LTE-M presents a viable alternative for MNOs and their customers to fully leverage their investments in LTE to support low-bandwidth, low-power implementations over managed spectrum, but we are realistically two or three years away from the technical and ecosystem maturity to match the low-cost/low-power characteristics via alternatives using ISM bands today. In the meantime, SIGFOX, LoRa Alliance, Weightless Special Interest Group and a few others will attempt to fill the existing market gap and prove their mettle. Any wins by these stakeholders will go far to prove their case for the viability of technology, ecosystem and business models at scale. The stakes are high as these firms reach for the brass ring of winning standardization and a long-term seat at the IoT connectivity table.
Why this matters

Prior to the introduction of alternatives such as Low Power Wide Area Network (LPWAN) – promoted by the LoRa ecosystem – and Ultra Narrow Band (UNB) technology – developed by SIGFOX – the long-range wide area networking for IoT fell to direct 2G/GPRS cellular, satellite, and fixed networks or distributed mesh network systems backhauled from an aggregation point. These systems were built to support low-bandwidth data gathering applications, such as utility meter reading, industrial monitoring and asset tracking. The higher-bandwidth IoT applications, such as video surveillance, digital signage and connected cars, are supported by more expensive and capable 3G/4G network/module combinations. With mesh networks, a series of devices work together – device data is sent from one device to another until a master node is reached and the data is then backhauled via satellite, 3G/4G or fixed line. The challenge with these mesh architectures is that they become unwieldy at larger scale – issues such as maintaining device clock synchronization and the asymmetry of power consumption (devices closer to the master node must act as repeaters, and therefore consume more power) can cause operational costs to expand beyond viability. The availability of low-cost, low-power networking technology creates the opportunity to evolve mesh networks to star networks, where each device establishes its own connection directly to cloud services or to an aggregation point.

The low-power, long-range upstarts

SIGFOX – moving rapidly

When SIGFOX announced that it closed the largest VC round ever for a French company, the size of the round was impressive, but even more interesting is who participated. Global telecom heavyweights Telefonica, SK Telecom and NTT DoCoMo participated in the round, which strongly signals that these operators see complementary value in the SIGFOX offering, or at least want to create a hedge bet.

SIGFOX uses an Ultra Narrow Band (UNB)-based radio technology to connect devices to its network. To put this technology choice into context, it was able to cover 90% of the population of Spain with only 1,500 base stations at a cost of $15m – at least 10x lower than similar coverage built with 2G technologies. As a result, SIGFOX offers an alternative in places such as the US, where the decision to rapidly shutter 2G networks to re-farm spectrum for LTE and LTE-A raises serious questions on how to fill the need for machine connectivity that will never require the bandwidth capacity of 3G/4G, or where the power-drawing characteristics of 2G/3G/4G edge equipment doesn’t fit the application or equipment lifecycles. The typical SIGFOX customer pays the company on average $5-8 per year for connectivity services. For far less than $1 per month, data consumption is rigidly controlled – only 140 12-byte messages are allowed per day upstream and four eight-byte messages downstream. These controls are partially driven by the rules that regulators have set for the use of ISM frequency bands, and partially by the design attributes of the SIGFOX network architecture itself.

SIGFOX is as interesting for its underlying business model as for its technology. The company has ambitious plans to build out its global network coverage by signing up strategic network operator partners that take care of the local ‘heavy lifting’ of network deployment, using SIGFOX technologies and back-office systems. If SIGFOX can execute flawlessly, it should continue to attract customers and partners that are hungry for the low-cost and power-sipping characteristics that its UNB technologies promise to deliver. MNOs may be wary of SIGFOX for some deployment scenarios, but they may conclude that SIGFOX is more friend than foe – as they re-farm spectrum allocation to higher-bandwidth technologies, such as 4G and ultimately 5G, but still aim to maximize connectivity options for enterprise customers deploying IoT. The major downside to the SIGFOX architecture is that data is able to travel in only one direction – from device to cloud.

  • Strengths: First-mover advantage, market traction, low costs, innovative supporting business model, expansion cash with tier one operator support (Telefonica, SK Telecom, NTT DoCoMo).
  • Weaknesses: Dependency on network operator partners, proprietary, only supports limited one-way communication path, no support for mobility.

LoRa Alliance

LoRa Alliance represents a serious competitor to SIGFOX, and its approach stands in some contrast to the approach SIGFOX is taking. The LoRa Alliance mission is to drive the adoption of LPWAN standards and develop a vibrant ecosystem around the LoRa protocol (LoRaWAN) by sharing knowledge and promoting interoperability. It’s taken some solid first steps in building a coalition of the willing. Initial participation included technology industry heavyweights IBM and Cisco. Additional participants include MultiTech, Sagemcom, Semtech and Microchip, along with endorsement from major operators such as Bouygues Telecom, KPN, SingTel, Proximus and Swisscom. At the center of LoRa is Semtech, the analog and mixed-signal semiconductor company that developed the LoRa RF platform – a two-way wireless transceiver and gateway it positions with ranges up to 10 miles on sub-GHz spectrum bands.

Actility is the farthest along with this technology – it has actually built networks using the LoRa gear from Semtech. Semtech provides the edge technology, which is based on the new generation of bidirectional LoRa transceivers. Actility provides the wireless base station network, core long-range controller (LRC) servers, associated operations support systems (OSS), and application development environment and app store. The two companies consider themselves natural fits for low-power, low-bandwidth usage scenarios such as smart parking, digital signage monitoring, electric vehicle charging, smart metering and safety sensor metering. Actility benefits from Semtech’s existing ecosystem of sensor vendors addressing the verticals mentioned above. With the combination of Actility and Semtech, a typical smart city or smart building project would require only a small, low-cost antenna on a rooftop to connect sensors in a 2-5km radius for dense cities, or a 15km radius for rural and less densely populated applications. Semtech’s LoRa technology claims 5-10dB more link budget than cellular technology and strong in-building penetration characteristics where cellular and GPS cannot reach. At Mobile World Congress, Actility announced a technology partnership with both Swisscom and nke Electronics.

  • Strengths: Strong in-building penetration characteristics, relatively broad industry support from operators.
  • Weaknesses: Still relatively unproven technology.

Weightless: Huawei (Nuel)/NWave

In September Huawei announced that it had acquired IoT specialist Nuel, the silicon company behind the invention of the Weightless connectivity platform, for $25m. Neul was formed in 2010 by former executives and founders of Cambridge Silicon Radio. The company had originally targeted its efforts at utilizing abandoned analog TV whitespace, but later iterated its approach to work across a variety of unlicensed and licensed spectrum bands. The Weightless platform supports bidirectional data flow between edge and core if the application is also expected to be supported via legacy GSM base station infrastructure. Weightless expects to be able to achieve battery life of 10 years at the edge device, a normal range of 5km and chipset costs below $2 over the next few years (initial Weightless modules will cost around $12). The Weightless-N standard (expected to be released in Q2 2015) has generated some impressive industry support, driven by the Weightless Special Interest Group, which includes more than 1,000 members. The Version 1.0 standard will support UNB uplink connectivity, frequency hopping (to mitigate interference), mobility support and shared key security protocol support (central server/edge).

At this point, the strategy for Huawei to fully leverage its investment in Neul and the Weightless standard efforts is unclear; we expect some of the technology to underpin its own offerings for SPs or DIY IoT networks. NWave Technologies is another company in this mix – it builds systems that can operate in the unlicensed industrial, scientific and medical (ISM) spectrum. In October 2014, NWave announced that it was contributing its own proprietary technology to the Weightless SIG to support Weightless-N, and has become a vocal leader of the movement – likely sensing a much larger opportunity if Weightless can catch on as an industry standard.

  • Strengths: Broad industry support, security, support for device mobility in Release 1.0.
  • Weaknesses: First iteration will only support uni-directional data flow, lack of clarity on Huawei’s strategy.

On-Ramp Wireless

Perhaps the least well known but still active player in the bunch is US-based startup On-Ramp Wireless, which has supplied end-to-end infrastructure in several installations around the globe – it, too, has built its infrastructure to operate in the unlicensed ISM band. The company has patented technology that it calls RPMA – Random Phase Multiple Access – which has been designed to minimize interference in the noisy unlicensed spectrum. The company touts the ability to support up to 16,000 endpoints from a single access point while covering more than 400 square miles per access point, if proper elevation requirements can be met. On-Ramp is a founding member of the 802.15.4K working group, which has set out to develop open standards for low-power, long-range critical infrastructure monitoring. On-Ramp is also offering a fully managed service offering backed by a 99.99% uptime SLA to spur deployments.

  • Strengths: Real world deployments, managed service options with SLA support.
  • Weaknesses: Still relatively unproven technology beyond private deployments.

The case for LTE-M

One of the refrains heard at the stands of larger operators and cellular module makers is that the pre-standard activity around LPWAN is simply a placeholder until LTE-M (3GPP release 13) comes online in 2016/17. LTE/4G is the global mobile broadband wireless standard of record. In many parts of the world, spectrum assets are actively getting re-farmed to support LTE in lieu of GSM/GPRS – LTE supports much better spectral efficiency, and therefore higher bandwidth at lower costs. LTE supports both frequency division duplex (FDD) and time division duplex (TDD) modes using a common sub-frame structure of size, 1ms. Having such a short sub-frame length allows latency to be minimized, ensuring a good user experience. LTE in its current form is ‘overkill’ for most IoT usage scenarios, and can’t support the cost and edge-device battery-power requirements of a majority of low-bandwidth IoT use cases.

Anticipating these market dynamics, the 3GPP has specified new system characteristics to support low bandwidth and power requirements of IoT implementations. The new releases are called Category O and Category M. Both of these new specifications will strip away unnecessary functionality and cost to support reduced bandwidth of 1Mbps FDD and 375Kbps HD-FDD, with a single receiver, less memory, reduced number of decoding attempts and power savings mode. CAT-O LTE modules are expected to be commercially available from companies like Sierra Wireless later this year. LTE-M modules are expected to show up in 2017.


That the need and corresponding demand for low-cost, low-power WAN options will increase sharply as the IoT market matures is not in question, but the ‘how’ certainly is. Just as standards and architecture debates heats up within the local-area and personal-area zones of IoT between ZigBee, Wi-Fi, Bluetooth LE and 6LoWPAN, the same dynamic is underway for LPWAN. While there are technical differences among the different approaches – for instance, SIGFOX and Weightless-N networks only support unidirectional flows in their current iterations – the real question is how far these new technologies can get before LTE-M and its subsequent iterations can catch up on the cost and power-reduction curves. Global mobile network operators are hedging their bets; on one side they are participating and even funding the new unlicensed spectrum alternatives, but many view these to be placeholders for their licensed networks of the future. The race is now on for specialists to take full advantage of the slowly closing window of opportunity to prove their mettle across the continuum of quality, cost and security.