Inetlab platform: full-scale laboratory for connected objects

While the Internet of Things (IoT) promises to revolutionize many uses, its applications today remain limited in relation to the technology's potential. All over the world, researchers and manufacturers are continuing to develop new systems based on networks of connected objects. However, once a model has been developed and tested on a computer, how can it be put to the test in real-life conditions? Such experimentation requires a large infrastructure, with multiple pieces of equipment communicating with each other, which is often too costly for a laboratory.

Over 1,000 connected objects accessible to all

The national SILECS initiative aims to fill this gap, through its IoT-LAB service, by providing an infrastructure of over a thousand connected objects. These are spread across seven sites in France, including one at Strasbourg University's ICube laboratory, operated by the Inetlab platform and comprising several hundred pieces of equipment. "It' s important to note that our main activity is to contribute to a national infrastructure network," stresses Guillaume Schreiner, technical manager of the Inetlab platform at the ICube laboratory. "What' s more, the service will soon be part of a Europe-wide project. SILECS will then become the French node of the European SLICES (anagram of SILECS) initiative.

The IoT-LAB service is now open to all, and free of charge: simply create an account on the dedicated portal to reserve equipment available throughout France and carry out large-scale experiments. In particular, the infrastructure offers "constrained" connected objects, i.e. low-power, energy-hungry devices equipped with standard sensors.

The platform's users can therefore come from a wide variety of backgrounds: individual IoT enthusiasts, start-ups, major corporations... But in practice, the majority of the community is drawn from academic circles worldwide. Researchers see it as a way of sharing their experiments with their peers and validating their results scientifically.

Monitoring performance and energy consumption

IoT-LAB is particularly well known for testing communication protocols between connected objects. In particular, the tool can be used to evaluate their performance through various measurements: radio power, data transmission speed, latency, message loss rate, etc. " Our research community frequently uses it to test mesh networks of connected objects," notes Guillaume Schreiner. " For example, within a building, it's possible to design a network comprising a single gateway connected to the Internet. And within this mesh, each object will pass through its neighbors to repeat its message, right up to the point of exit. " This design facilitates the deployment of such a network of devices, since only one of them needs to be connected to the Internet.

The service also offers users the possibility of tracking the energy consumption of their equipment. This is a key factor in the design of networks of connected objects, and not just because of their ecological impact. After all, some devices are designed to operate autonomously on battery power. It is therefore vital to optimize their energy consumption, so as to maximize their lifespan.

Rainfall in a remote region of Uganda

In addition, for concrete applications of networks of connected objects, the Inetlab platform team works directly with researchers in various fields: industry, health, meteorology, etc. It then puts its network expertise at the service of multi-disciplinary collaborations, as in the case of the Strasbourg Eurometropole's connected city.

For example, the Inetlab team was approached by two researchers from the University of Maryland (USA), Inbal Becker-Reshef and Catherine Nakalembe, respectively World and Africa Director of NASA's Harvest program, which groups together projects relating to agriculture and food security. The aim: to deploy a network of very long-range radio weather stations to measure rainfall in a remote Ugandan valley on the border with Kenya. " For this project, we built on the work we'd done in Strasbourg," recalls the technical manager. " The notable difference here was that there was no Internet connection point close to the measuring instruments. So we deployed our own nomadic antennas, based on LoRaWAN autonomous long-range radio technology, and connected the measuring instruments within a radius of 10 to 15 km. " This network made it possible to collect weather data from remote points in the valley and transmit it, in real time, to servers in Strasbourg.

From the point of view of NASA's Harvest program, this deployment primarily met a scientific challenge: to cross-reference telemetry data with that collected in the field via weather stations, in order to assess rainfall intensity more accurately and reliably. A crucial piece of information in this part of the world. First and foremost, it helps identify where and when it is necessary to irrigate a field, and when it is possible to do without, in a country where water is a precious resource. It can also help to protect the population, by alerting them to possible landslides, depending on the amount of water absorbed by the soil.

Towards data lifecycle monitoring

This example illustrates the wealth of possible applications for connected objects. In order to encourage the development of such projects, the Inetlab platform has no shortage of ambitions, always in keeping with national and European initiatives. " Our next objective is to be able to track data over its entire lifetime," presents Guillaume Schreiner. " Today, we have the producers of data: connected objects. Now we want to look at the rest of the way, starting with the major Internet networks, which have to adapt to a new traffic of small messages arriving en masse. And we also want to add to our experimental suite the final link in the chain: data centers, which harvest, store and process data. " The idea is to reproduce, at each stage, the experiments already in place, in order to monitor performance and energy consumption throughout the life of the data. The aim is to address new issues, such as the centralization or decentralization(fog computing) of data processing.

A development supported by Carnot TSN, which the technical manager considers important: "Today, we enjoy a solid reputation and credibility in the eyes of the academic world. We'd like this to change, with even greater visibility thanks to Carnot TSN, to create more links with the industrial world, by providing guarantees of the seriousness of our work".