Now is LoRa: Building the Smart City bottom up
On Monday 16 November we organised a panel discussion about The Things Network and LoRaWAN in FabLab Barcelona (in Catalan, in English). We aimed to look at this from a commons perspective, which is often missed in the industry around the Internet of Things, which we consider particularly relevant now this week that the Smart City Expo World Congress is held in Barcelona.
The panel was moderated by Ricard Faura, head of the Knowledge Society within the Generalitat, the Catalan government. Mara Balestrini, who collaborates with the FabLab, presented their Smart Citizen project. Xavier Pi, who works in the municipal Knowledge Society department and coordinates the Embedded systems group at the Industrial Engineers College, presented his vision on Industrial Commons. Javier Creus, founder of Ideas4Change, inventor of pentagrowth and president of Ouishare Spain, shared his wisdom on creating, sharing and exchanging value. Marco Berlinguer, co-founder (with me and others) of the School of the Commons Barcelona and researcher at UAB/IGOPnet presented a more conceptual perspective of the digital commons and in particular its central production model of Commons Based Peer Production.
My role was to present the Things Network (see slides, 2,8 MB). This week that so many people come to Barcelona to see the latest developments in Smart City solutions at the Smart City Expo it seems a timely matter to discuss its socio-technical foundations. How can we build these smart cities so that they work for all of us, with a level playing field for all to participate, that we empower citizens and keep technological independence?
I presented the importance of distributed networks, especially for three fundamental reasons. First, as Edward Snowden has reminded us of with his revelations of the National Security Agency spying affairs, the only way to be able to maintain private our communications is through strong end-to-end encryption. Snowden is the living proof that that is so: he managed to escape and keep alive through it. Now we know that encryption as used by Google, Facebook and other private online platforms is not end-to-end and in fact is being decrypted by those companies to assure automated connections to the NSA through the backdoor. Now imagine the predictions of the exponential growth of the Internet of Things with billions of sensors and actuators all around us. If those are controlled through proprietary infrastructures controlled by those such companies and governments who wish to end encryption (in the name of terrorism!), we can start to imagine the surveillance state we are heading towards. George Orwell's big brother science fiction scenario will prove to be naif.
Second, centralised networks are known to be highly vulnerable to attacks. In fact if we go back to the origings of the Internet, when, in the 1960's, the USA's Defense agency DARPA started to design its foundations, it was with the objective to maintain a stable network in the case of a serious attack (like the A-bomb). And they concluded that a distributed network was the best way, the only one to keep operating despite attacks. While the Internet today didn't evolve really as a distributed network, at least its decentralised nature is clear. And that couldn't have evolved without three important foundations: 1) end-to-end communication between peers, 2) the use of open standard protocols and 3) the dominance of Free Software implementations. These are arguably the net's most important success factors.
The third argument for distributed networks is equality of opportunities. Who controls the technical infrastructure gets to control its social affordances or possibilities people have and under what conditions they can participate. If, as Wienke Giezeman (founder of the Things Network) says, one can only connect his/her sensors to the net by paying an subscription fee, this will leave out very interesting usecases. In other words: the owner of the network decides what we can do with it, what protocols we can use, whether and what kind of encryption we can use, whether we can stay anonymous or not.
A commons-based infrastructure allows us to self-manage the network collectively. And that is what the Things Network started to do. Last summer a small community in Amsterdam set up ten antennas providing free-of-charge access to anybody. Let's see how they did that and how dozens of cities started replicating the idea.
First we need to choose the most appropriate radio technology to connect our sensors to the net. While the Internet of Things is under development for years already, only now do we start to get the technologies in place that allow us to connect sensors and actuators in practical ways. One of the problems we had until now was the existing wireless and mobile technologies: they either don't work long with battery powered devices (Wifi, mobile, ZigBee) or don't cover wide areas / long distances (WiFi, Zigbee, Bluetooth). The last years several companies have been developing technologies for so called Low Power Wide Area Networks (LPWAN). They use unlicensed radio spectrum and the radio chips are cheap. That’s a big enabler as mobile licenses are generally considerably expensive.
The following three LPWAN technologies may be the most promising: SigFox, Ingenu and LoRa. SigFox is a French company deploying its private network around the globe based on a free radio frequency technology that they developed. Ingenu is a US based company that is similarly deploying its own private network based on its own technology. LoRa is the only radio technology for low power wide area networks that is developed by an open consortium based on open standard protocols. It is developed by the LoRa Alliance, founded by SemTech, IBM, Cisco and a long list of companies has become member. LoRa or LoRaWAN stands for Long Range Wide Area Network.
Like SigFox it uses the free radio spectrum to exchange its messages between sensor nodes and gateway antennas. While the free spectrum is unregulated and unlicensed there is no guarantee on the space available (or the amount of noise on that particular frequency band). However in recent years algorithms have been developed to transmit messages even in very heavily used frequency bands. Furthermore LoRa nodes automatically adjust the bit rate of their transmissions and can transmit up to 50 kbs. Given the typically small message size for sensor data, this means that the communication terminates quickly and the node can go offline until the next communication is planned, thereby saving its battery. Another relevant feature is that LoRa chips can be configured to work on different frequencies, as free radio spectrum varies between countries. Typically however the 868 MHz or 432 MHz frequencies are used.
Now LoRa networks can be deployed both as open or closed networks, controlled by private operators or by communities. The Things Network is unique in the sense that it works as a community following a manifesto that assures net neutrality, free of charge access to anyone, anonymous usage and no one can impose any restrictions on the network's users. It is this manifesto that is the basis for the network being a commons, or common pool resource. It can be compared to guifi.net or other free wireless communities. It is very different however for its low data transmission and its long distance antennas.
Reports have shown that LoRa antennas can cover up to 10 km in an open field. Others have shown that 5 km is a reasonable limit, while 3 km when the antenna is in a car. In cities where the transmission needs to pass through buildings, 2 km seems a reasonable limit. However in the audience we had one person who had been testing LoRa antennas in the city of Barcelona and was able to reach up to one km. Even a one kilometer radio makes more than 300 hectares, so a considerable Wide Area Network so to say.
In my presentation (and in this analysis) I argue that the Things Network can be seen as a set of four different commons. We have seen the LoRa protocol that is managed by the LoRa Alliance and can be considered an open standard. That would be commons number one. Second we have seen the Things Network providing a commons-based network, built up by peers who agree to the manifesto. This network doesn't require anything from its users. Possibly that may work, it has to be seen yet whether people have enough incentive to set up an antenna for the fun of it.
As Mara Balestrini pointed out about the Smart Citizen sensor kits, around 30% of the peers who bought that kit didn't put it to work. Interestingly though she explained that the Waag Society in Amsterdam, who participates both in Smart Citizen as in The Things Network, has introduced a peer learning experiment. Peers map their skills on an online platform and are matched to help each other and share knowledge with the final objective to keep the sensors operating and collectively gathering environmental data.
For the Things Network the motivations of people buying an antenna might be similar. In both cases peers contribute to deploying a collective network infrastructure. Some people might want to learn how LoRa works, and experiment with wirelessly connected sensors. Another motivation maybe to develop a business on top of such infrastructure, providing products that connect with it, or services, platforms and solutions. Public institutions may see the many useful services for remote monitoring, asset tracking, etc and consider that a commons-based infrastructure is the most sensible way forward.
While contributors may buy and put to work their individual antennas, I think it makes sense to organise locally in some form. That maybe an informal or more formal association in whatever legal form that's considered most practical. That would enable the collective infrastructure to be governed as a commons. In this sense we can learn much from communities like guifi.net, that have implemented the design principles as defined by Elinor Ostrom for commons governance. Costs that are made for the common infrastructure are being shared by its users. That could work on a donations based model, or with membership fees.
The third commons in the Things Network is the software that is developed as Free Software, under the GPLv2 license.
The fourth commons we can consider is the hardware design itself. While in Amsterdam the community set up ten antennas from commercial vendors (ca. 1200 €/antenna), the Things Network has developed a low cost plug and play antenna costing 200 euro in its kickstarter campaign. The designs will be published under the CC BY-SA license as Open Source Hardware and thus allowing the community to further develop and adapt it to ones' needs - as a commons resource. Notice however that the design of the LoRa chips is not a commons, and requires licensing from SemTech.
The kickstarter campaign reached its goal of 150.000 euro in its first ten days and has completed at almost 200%. More than 900 people have backed the campaign, purchasing a couple of thousand antennas, covering hundreds of cities around the world. Now is the time for LoRa! (Ara és LoRa! in Catalan ;-))
Xavier Pi illuminated us with a short presentation (PDF, 1,3 MB) of the industrial revolution in which we seem to be in and how things are changing from centralised to bottum up networked organisations. He referred to Jeremy Rifkin's book Third Industrial Revolution and its most recent one, The Zero Marginal Cost Society, in which he describes how capitalism is receding while the "collaborative commons" is on the rise.
Pi sees the converge of the physical and the digital world. In other words what we have come to call Atoms-Bits Convergence in the Digital DIY project. And likewise Pi sees how the maker movement is generating new business opportunities reconverting bankrupt electronic shops in Barcelona into maker shops that help Digital DIYers to realise their projects with components, tools and advice. In a more transformational way, Pi suggested that we're in the process towards an industrial commons.
Javi Creus reflected on the value of the Things Network, how it is not exchanged in the market but certainly appropriated by the users. For its success it is crucial to assure that the commoners contributing to the network also assure they appropriate at least part of this value. In this sense Javi pointed to the alignment of energy and communication, in so called smart grids. Even though Spanish public policy is currently very much against self-production, this will inevitably change in the near future. Moreover current legislation doesn't affect self-production for off-grid situations.
Marco Berlinguer shared a more conceptual view with us on the commons and commons-based peer production. In a research project where we have collaborated together, the P2PValue project, we studied almost 400 communities (see directory) that peer produce their commons. We kept some 150 variables and indicators on each case. This was the first study of this size and diversity done on Commons Based Peer Production so far. This has given interesting insights on how value is created in such diverse communities as the Linux kernel, BitCoin, Couchsurfing or AirBnB.
Marco pointed out that the rise of the commons relates to a reduction in market size for that particular market segment. Take Wikipedia: its dominance limits the possibilities for market-based encyclopedias to sell similar products. In the software market the case is clear: wherever Free Software solutions have been developed (as a commons), this has reduced the market segment. This confirms Rifkin's thesis for a decline of the market and rise of the commons.
Marco insisted on the importance of Free Software for building a commons, and in particular its underlying principles as stated in the Free Software Definition. For a free society we need the freedom to use the tools we need. For building communities we need the freedom to copy and share. For improving our tools we need the freedom to access the source code and share our improvements. Those are fundamental rights we should always keep in mind.
Mara Balestrini mentioned a new research project that makes sense of making: Make Sense. I'll be glad to hear more about it.
From the audience we had various interesting interventions. A few discussed the noise that one can experience on the free radio bands in the city of Barcelona (and I suppose in any big city?). One person suggested to request a dedicated frequency concession. The municipality of Barcelona already has a few of those. While the LoRa technology is designed to work with heavily used frequency bands, it might still be an interesting idea to pursue. However certain telecom operators might lobby against it. In any case, the LoRa technology seems to be capable of handling the noise, even in Barcelona.
From our community in Mataró there were two members present. There we have formed a group of around 15 persons to set up the Things Network collectively, and share the burden of the first antennas between us. It is worth noting that several of the guifi.net wireless community of Mataró participate in our group. That seems an interesting symbiosis that we could explore in Barcelona as well.
Another local group is being formed in Arenys de Mar, where one of the political parties is preparing a proposal for the municipality to opt for LoRa and the Things Network, and contribute to the community.
I look forward to see more people contributing to the Things Network. Now is the time to build this together. Especially as big telecom operators are planning their private network deployments - also based in LoRa - for 2016. We have the opportunity to build a commons infrastructure. Just to repeat myself, let's keep clear some of the key questions:
- who controls the infrastructure?
- how can we guarantee privacy and anonimity?
- how can we do this more inclusively? Do we guarantee the same opportunities for all?
- can we freely choose what technologies we use?
Summarising: do we go for a surveillance state worse than Orwell's 1984, or do we work together around a free technology, that is developed bottom up by all interested parties? Distributed networks, cooperation, empowerment, user communities, commons based peer production.
Let's make sure many of the school buildings, public offices, fablabs, hackerspaces and IoT related companies join us and set up their antenna. Now is the time for LoRa! :)