The imminent death of Sigfox and LoRa at the hands of cellular NB-IoT has been much exaggerated. There are two important points in the favor of unlicensed spectrum solutions for LPWANs (low power wide area networks) – they have an existing market presence, and they enable non-spectrum owners to play in the growing market for digital machine-to-machine (M2M) services, including in smart cities.
Delays and dilemmas in NB-IoT commercialization – seen last week in Orange’s decision to defocus on the standard for now – point to a multi-access picture persisting in LPWANs, with the power shifting to platforms that can support seamless interoperability between different cellular and non-cellular solutions (see previous item).
When 3GPP finalized its own LTE-based LPWAN standards last year, proponents argued that there would be only small corners of the market left for the unlicensed alternatives. The mobile operators had key assets such as sites and relationships with city authorities, and the twin standards – LTE-M and Narrowband-IoT – could be implemented easily and cost-effectively as updates on existing LTE networks.
All true, but there were various flies in the ointment. The cellular standards might be relatively simple to deploy, but they will still take time to develop the developer and services ecosystems which are essential to the uptake of any network. As in wireless broadband, where unlicensed WiFi has increasingly become an integral part of the connectivity jigsaw, so in LPWAN, where Sigfox and LoRa are working to integrate themselves into the cellular-driven environment with new levels of coexistence, rather than present MNOs with an either/or decision.
Also as in WiFi, they additionally provide non-spectrum owners – such as cablecos, increasingly hungry to play in the wireless and multiplay worlds – with an entry pass, as seen in the US where Comcast is expanding its LoRaWAN roll-out to a dozen cities so far.
For MNOs, the pressure from new rivals such as cablecos drives new urgency to colonize LPWAN territory as quickly as possible. Even with NB-IoT and LTE-M out there, this will be more effectively done using a combination of unlicensed and licensed technologies, especially if an operator has already made investments in Sigfox or LoRa, as many have. Going cellular-only would involve those networks becoming orphans, or being ripped out before they had achieved return on investment. And LTE-M and NB-IoT are still at an immature stage, with even their greatest supporters, like Vodafone, experiencing delays to their roll-out targets.
France is one of the most interesting case studies of LPWAN decisions by MNOs. Home to Sigfox and to several large-scale MNO roll-outs in both licensed and unlicensed bands, it has provided many useful lessons for less advanced markets. Sigfox has been working with TDF and other partners to build a national IoT network using its technology and ‘as-a-service’ model. Altice, the powerful group which owns both the second MNO, SFR, and the leading cableco, Numericable, is also working with Sigfox. The third MNO, Bouygues, has set up an autonomous business unit based around its LoRaWAN network and services.
Orange, like other MNOs such as SK Telecom in South Korea, has been following a multiple network strategy. It is one of the few operators, outside the US, to be deploying both LTE-M and NB-IoT in parallel. US carriers like AT&T have been rolling out LTE-M first (AT&T on a national scale). This standard was part of the 3GPP Release 13, so it pre-dates NB-IoT (Release 14) but has higher cost and power consumption, making it suited to M2M data devices such as wearables rather than ultra-low power sensor networks like those enabled by Sigfox, LoRa and NB-IoT.
Orange also has a significant LoRaWAN deployment across France. According to Mobile Europe, it is now pressing the pause button on NB-IoT to concentrate on LTE-M and LoRa.
Orange’s Jean-Pierre Casara told the 5G World conference in June: “We are not against NB-IoT and we also plan to roll out NB-IoT as a complement to the others. Obviously NB-IoT doesn’t make sense in France because LoRa covers half of the use cases but it does in other markets.”
Olivier Ondet, VP of IoT and analytics at Orange Business Services, is going further, telling the journal that plans to expand NB-IoT on a broad basis in France and other countries in 2018 are now on hold, except in Belgium, where the technology has already been deployed. For now, Orange can satisfy the demand it sees for IoT connectivity using LoRa and LTE-M, he added. The operator expanded its LTE-M trials in Spain last week and plans launches in Belgium and France by the end of the year, having been piloting a smart metering application.
This change of direction reflects a similar reappraisal of LTE-M in other parts of the world. Once considered of little interest outside North America, the technology is now regarded by some operators as more mature and interoperable than NB-IoT. And for those which have unlicensed LPWANs in place (useful to keep the cablecos out of the picture), these are complementary to LTE-M. They can support very low power sensor applications, while LTE-M can help them add new services requiring higher data rates. By contrast, NB-IoT would be a replacement for LoRa and Sigfox, rather than serving new customers and services, making its business case more challenging in the short term.
As well as Orange, Telefonica and KPN have belatedly fallen in love with LTE-M, and KPN also has a dual LoRa/LTE-M strategy, with no timelines announced for NB-IoT (or whether they will support it at all). SK Telecom is another LTE-M/LoRa supporter, using the former for connected car and other mobile M2M applications and the latter for low mobility, low power services. Some analysts even believe LTE-M could evolve to cover far more of the IoT bases. Tom Rebbeck, head of the enterprise practice at Analysys Mason, told LightReading: “The price difference between NB-IoT and LTE-M modules is marginal and there is not much difference in terms of battery performance. And the performance of LTE-M is a lot higher and might be more flexible for future applications.”
The split among European operators – between the NB-IoT skeptics and its most ardent supporters, such as Vodafone and Deutsche Telekom – will hurt efforts to build a broad ecosystem and significant economies of scale around a single technology. This was a key argument in favor of NB-IoT as opposed to the fragmented state of unlicensed spectrum LPWANs (not just LoRa and Sigfox but many other alternatives, some mainly operating in the private network field, which is very important in the Industrial IoT world).
But Ondet says NB-IoT lacks those economies of scale and remains limited in its openness, either to non-cellular technologies, or even multivendor solutions within the platform. “Obviously if we see Narrowband IoT taking off wonderfully, we will be there, but today we see LoRa as the ecosystem with momentum,” he said. Orange has 140 LoRa partners in France and 14m IoT connections across all its technologies (including GSM and the third new 3GPP LPWAN standard, Evolved GSM).
This is not just hedging bets. There are many IoT services which can potentially be supported by a long range, low power M2M network. But unlike mobile broadband services, which have a high degree of commonality in how they use the network, the IoT applications will have diverse requirements, and dramatic contrasts in revenue and profit potential. Some will require high security and availability, and therefore are likely to command a premium price. For others, reliability will be the top priority, or ultra-low power consumption, or low latency – and so on. Some city services will have little or no revenue, so cost of ownership will be a top consideration.
Ondet praises LoRa’s simplicity for low power, low cost deployments, saying: “The LoRa approach is very basic. When you start talking, you send a message and the network will sort it out.”
This can be complemented by the greater functionality of LTE-M, which is a “better option when the objects are moving or you want them to work everywhere in the world without too much bother. This is why when we work with our customers, it’s with both technologies. We believe that LTE-M is a perfect match for a wide range of IoT use cases in Industry 4.0, fleet management and asset tracking.”
In other words, a variety of connections, each optimized for a different set of performance criteria, and each coming at a different price point, makes good sense. It is very much the 5G vision for wireless broadband, but can potentially be implemented in the 4G era in the LPWAN at least.
That will only happen if there is seamless interoperability between the different connections. That has taken years to achieve between WiFi and cellular, and remains imperfect and complex to deploy. In the IoT, there are various innovations which may smooth the path – multimode chips; virtualized network orchestrators which assign functions across any access technology; hub platforms which support interconnection via common middleware. These coordination technologies may be run by an operator, or by a cloud provider, neutral host organization or vendor (as Nokia is offering, on a multinational scale, via its WING services).
Andres Escribano, IoT global connectivity business director at Telefonica, believes that scale and interoperability is essential at all three layers – network, devices and software. He said in a debate at the recent IoT World Expo in Czech Republic: “Globalization is one of the key topics we try to solve. If different components are designed for specific use cases, and we are unable to replicate this around the world, we have a problem in justifying that it really is a global solution.” Telefonica signed a global IoT connectivity deal with Sigfox earlier this year.
Georges Karam, CEO of Sequans, one of the vendors of LTE-M and NB-IoT chips, echoed this theme in a recent interview. He said it was simple to roll out IoT networks, and to upgrade from one technology to another within a few months, but the real problems are in the fragmentation of the device and developer ecosystems. He said: “You need to see things made more simple on IoT because of the fragmentation of the market. You need to allow any developer to use this connectivity, in the same way they would do if they were connecting a device to Bluetooth for example. And so, you have some challenges there related, for example, to the certification, the ease of use, the data plan of the carrier, and adapting the data plan for IoT application.”
Sigfox and LoRa are certainly making efforts to play nicely with cellular technologies so that they remain appealing to MNOs as well as non-spectrum owners. At the Czech event, Sigfox unveiled a multimode chipset, made by GCT Semiconductor,, which supports the three cellular standards (NB-IoT, LTE-M and EC-GSM) plus Sigfox itself. The French firm said “several OEM customers” are working with the chip.
CEO Ludovic Le Moan said: “GCT provides a single-chip innovative solution that further proves that Sigfox and cellular are complementary technologies by leveraging the best of both worlds – ultra-long battery life using Sigfox and high throughput cellular connectivity.”
Milan Haba, program director of integrated network strategy at T-Mobile Czech Republic, said: “I can only say that I welcome this initiative because it eases the life of customers. Life would be a lot easier if there were fewer LPWA technologies available”.
Another important advantage of cellular technologies is the roaming mechanisms which have been embedded in their platforms over the years. These have taken many years to replicate in WiFi, but the LoRaWAN community is determined not to wait so long. In June, Orange, Bouygues and Proximus of Belgium said they would start testing on interconnect and roaming between their respective LoRa networks by the end of this year, and hoped to involve other European players. The trials will be carried out under the auspices of the LoRa Alliance, which will also seek to use the work as the basis for a standards framework for roaming.