On September 25th and 26th, the new edition of TheThingsConference was held, the must-attend event of the LoRaWAN ecosystem and, more broadly, LPWAN. Even though the event, which was very open, is naturally becoming increasingly focused on TheThingsIndustrie’s solution and thus LoRaWAN, it remains an opportunity for numerous discussions in many other areas.
The organization was, as always, perfect, with ample space for side events that provide great opportunities for networking.
There were few announcements this year beyond the arrival of TheThingsIndustrie’s “plug & play” gateway and the introduction of trackers in the form of solar-powered tags, which seem to confirm the trend of IoT in this form factor, following RAK Sodaq and many other ongoing projects.
So, if we start with the usual figures, TheThingsIndustry (TTI) announces 2.3 million devices operating on their network, which is very impressive growth. They reported reaching the first million by the end of 2021, representing a growth factor of 2.3 over 30 months. TTI is experiencing substantial growth. However, it should be kept in mind that a network like Sigfox (Unabiz), which many people mistakenly believe has exited the race, has over 20 million devices and signed a deal in June for 15 million devices in South Africa. Nevertheless, it is a very impressive performance by Weinke and his team for a company that originated from a crowd-sourced network and had to build trust, which is now well established.
Another figure has been updated: the number of messages processed is 4.5 billion per month, which is 1712 per second—a fairly impressive volume. This implies an average of 2.6 messages per hour per device. It is not clear whether these are unique messages or if they include duplicates, but in any case, this gives us an idea of the average emission period of objects in large fleets and helps to put into perspective the capacity issues on LoRaWAN. Indeed, theoretical studies generally rely on usage at the limit of the Duty Cycle, which is 1% of the time and roughly several hundred messages per hour. With this data, we can clearly see that the reality for each object is well below the 1% duty cycle, closer to 0.01%, although for a realistic calculation, information on the average spreading factor (SF) used would be necessary.
Focus on IoT Solutions TCO
This year, the central theme was the reduction of the Total Cost of Ownership (TCO) of solutions, a theme I really like even though I tend to think that if we focus on costs, it means we still don’t fully understand the value creation brought by IoT deployments. I would like to believe that after having exploited all the value, we are so mature that we are tackling costs, but honestly, the industry still seems so far from understanding the capturable value of IoT. Moreover, what percentage already understand what IoT is and know LoRaWAN?
There are a few key points to understand that are strategic for TTI. On this subject, the first, of course, is that to start (and even in the long term), relying on a public network is much more economical than building your own network internally. Everyone likes to innovate internally, but like electricity and telephone, managing a network is a specialist’s job and involves significant fixed costs. I described this many years ago, concluding (and being very optimistic) that managing your own network was not economically efficient below 500-1000 devices in a coverage area (a site, a city). Based on that, entrusting the management of your LNS to TTI makes sense, especially since this year, TTI is announcing the release of a 100% autonomous and managed gateway, a plug & play gateway that reinforces this notion of reducing network costs.
This gateway is somewhat the nth re-edition of the specific hardware provided by TTI, starting with the very first one they launched with a Kickstarter in 2015, which was truly a breakthrough in terms of cost, allowing the democratization of LoRaWAN usage among a large tech community. Then came the TTIG, the first gateway priced below €100, and now a connected and managed gateway.
This gateway is for indoor use only but with radio performance that should be better than the TTIG. It includes “4G” communication, allowing it to be deployed without interfering with the company’s network. It is pre-commissioned on TTI and updates itself automatically. The evolutions of TheThingStack’s NOC solution also enable its management.
Honestly, I am curious to see how it will spread and the size of the associated market. For large corporations, I see too many limitations: the lack of an outdoor solution necessary for wide industrial coverage and especially the lock-in with TTN. For small structures, the presence of recurring costs of €5/month or $6 is still a challenge in IoT, in addition to a very reasonable initial price of €130.
However, it seems to me to be a good solution for deployments in medium-sized entities with potentially numerous sites, such as retail chains, and especially for service providers who need to deploy such solutions at their clients’ sites without connecting to the network. The recurring costs are then absorbed by the service itself being recurring, and the communication deal made by TTI is rather good for communication that works globally. For reference, with the same operator, one of my LoRaWAN gateways on 4G in a rather isolated site costs me $7/month, and TTI includes NOC/management services. In any case, it seems to me to be a good thing to have in the catalog today.
Protect our eco-system
Beyond the announcements, there was an intervention by the movement Liberté, Égalité, Connectivity (Freedom, Equality, Connectivity) which aims to defend the use of ISM bands—free and open radio frequencies—that enable the advent of LPWAN networks. As strange as it may seem, these frequency bands, accessible for decades, no longer appear to be a secured and community-owned asset. Indeed, in the United States, a private company has taken the liberty to propose the privatization of a portion of these frequencies for its own interests. While this request is unlikely to succeed, it subtly highlights the inability of policymakers to grasp the stakes surrounding IoT technologies and the resources they require. The movement, originating from the LoRa Alliance, thus aims to inform and protect our rights. A movement to watch. The happening was quite nice and funny during the event.
AI on the trends
For the rest, we had tons of discussions with the various participants. In the developments related to LoRaWAN technology, we naturally saw AI making its way into the keynote and the booths, particularly at Seeed Studio, whom I must thank for helping me attend the conference. As I have been saying for almost a decade, IoT is the engine for capturing AI data to feed its real-world data models. Concretely, there are two approaches: the older one being the massive data collection to feed models operated in data centers, and the newer one, with the development of edge computing and especially the energy optimization of data capture and local model execution, we are seeing solutions like those from Seeed Studio. Presented in a workshop, their solution enables a low-cost and energy-efficient device capable of analyzing a video stream, identifying patterns it has been specifically trained for, and transmitting the presence of these patterns via a LoRaWAN network. From there, counting people, identifying vehicles, detecting presences, preventing risks for isolated workers… all these use cases become accessible. While this is not necessarily new, Seeed has significantly improved the user experience, whether with Watcher, an exceptional tool for testing and searching for business use cases for embedded AI, or with its fully integrated platform for building and deploying models online, such as accessing a large library of pre-trained models. In short, the world of AI is finally converging with that of IoT, which seems like good news to me since the industry appears to better understand the stakes of AI.
LoRa (Wan) Mesh discussions are growing
The theme of LoRa Mesh, which I have been addressing regularly lately, was also at the heart of many discussions outside of the official program, demonstrating a growing trend and need. The relay technology proposed by the LoRa Alliance is struggling to gain traction; it must be said that functionally it does not seem fully developed. As a result, Rak and Chirpstack have developed a rather interesting solution based on standard hardware incorporating Chirpstack’s LNS configured in a specific mode where it will repeat the received frames, including all metadata, ultimately allowing a relayed device to appear as if it were directly connected to a conventional gateway. This enables seamless integration into an existing network where the challenge is to extend the network without necessarily having accessible 4G connectivity (or to save on this connectivity) with gateways that can, for example, be solar-powered and thus completely autonomous. Information can be relayed over LoRa 2.4GHz, which is also quite clever to avoid packet loss (non-full duplex) and manage the duty cycle at the gateway level. Finally, the use of 2.4GHz 😉 Besides this solution, currently, between Meshastic, which is gaining a lot of attention for its commercial success (more than technological), and Orama-net, a very promising startup in this field and for industrial applications, mesh is in the spotlight.
Well, during the conference, I also talked a lot about Mioty, but in this case, it’s because I’m working on it and articles will be published soon. However, it was interesting to confront opinions that are quite convergent on both poor marketing, the lack of ecosystem maturity, and the general desire to extract too much value too early, greatly impacting the image of a technology that is otherwise quite nice. Even though personally I am still looking for its sweet spot, but we will return to that in future articles.
New products on the way
Throughout the discussions, I also learned about the upcoming release of new Kerlink gateways, operating on solar energy with very impressive specifications. They have apparently made a significant effort to reduce the gateway’s power consumption, which is quite good since they were not the best in this aspect previously. However, what I was told is better than what I have seen so far, and this allows them to announce autonomies that should enable a 100% SLA with solar power in most parts of the world.
Lacuna Space, a satellite-based LoRaWAN solution, continues its development and is building partnerships. We are expecting more satellites to become operational next year, allowing for a more usable service. Currently, we can only rely on one message per day on our end, which limits use cases. Therefore, I am eager to talk to you more about this. In the South America region, their partnership allows the use of the S-band with satellites orbiting around the South Pole, enabling a service that is almost geostationary but with much less distance, which is also interesting. In any case, I got the impression that the team is growing, which is already encouraging.
Finally, another development in technology that seems to be following a promising trend is the arrival of mono-channel LoRaHub gateways. I discussed this at the Munich LoRa Alliance show a few months ago. In my opinion, this is very important for LoRaWAN, which does not offer satisfactory global coverage for B2C deployments (also applicable in B2B for services such as room management, environmental management…). The Munich demo, based on a Heltech module, was technologically interesting but required better integration to be usable. In Amsterdam, I saw an implementation that still seems to be in beta by Seeed Studio in its product named Indicator, which clearly demonstrates the potential of this solution. By offering local LoRaWAN coverage with a quality touchscreen display and WiFi connectivity, it becomes possible at a very low cost to provide both an IoT gateway and a synthetic data consultation dashboard for individuals, while allowing integration with the cloud to enrich this data and offer advanced services based on massive data collection feeding AI engines. Thus, a missing link in LoRaWAN for B2C, which is a gigantic and emerging market, is beginning to appear.
As always, I leave Amsterdam with 1,000 ideas and things to test or do, and I could still elaborate on the potential market for Truvami’s solar tracker tag. However, it seems that the article is getting long, so I will save that for a dedicated article a bit later when I have had more time to test it in the field.