Sigfox’s CTO on where satellite fits in an IoT-only network


This article originally appeared in the Nov. 20, 2017 issue of SpaceNews magazine.

Sigfox is forming a network uniquely designed for the Internet of Things (IoT), essentially building the infrastructure for legions of devices to share digital information. French global satellite operator Eutelsat is among Sigfox’s backers, participating in a $115 million capital raise in 2015 that ranked as the largest venture capital round in France’s history.

Last November Sigfox raised another $159 million, furthering its development of a worldwide network for devices. Raoul Mallart, Sigfox’s chief technology officer, envisions a future where devices are so connected that they will enable problem solving on a level unthinkable today.

As far as new technologies go, IoT is still young and impressionable. Satellite companies are vying to play a large role in this greenfield environment. Two satellite operators, Inmarsat and Thuraya, joined Sigfox competitor LoRa Alliance to help define IoT standards.

Sigfox covers around 80 percent of the population of Europe with its IoT network, Mallart said, adding that 36 countries in total have Sigfox networks operating, with 17 of those continuing to scale. Outside of Europe, Sigfox reaches Australia, Japan, Singapore, South Africa, Brazil, Mexico, Colombia, Canada, the U.S. and others, he said. Sigfox’s goal is global coverage.

From Mallart’s perspective, there is need for satellite in making the Internet reach everything it can, but satellite won’t be the primary connectivity most of the time.

Before becoming CTO of Sigfox, Mallart donned the title of vice president of imagineering, responsible for synthesizing imagination and engineering, as the word suggests. Mallart said he ditched that title to avoid giving the impression that “innovation” was limited to one person or department.

SpaceNews spoke with Mallart about how satellite communications fits into Sigfox’s vision for the future Internet of Things.

A Sigfox IoT antenna, located in Anarctica. (Credit: Sigfox)
A Sigfox IoT antenna, located in Anarctica. (Credit: Sigfox)

How does one build a network that is optimized for the Internet of Things?

There are three approaches I can think of. First is you use an existing network, like a cellular network. That is what started at the beginning as M2M. The idea was to use the same cellular network that you use for your mobile phone and equip objects with cellular modems. The issue with this is that the price of the modem is quite high and the power consumption is high. If you have applications which must run on battery, it becomes very problematic because you need to change the battery very often and it becomes very costly.

There are two other ways to solve that: one way is to go to local area networks, like a home Wi-Fi network. You can use the local networks to connect objects. This is O.K. for some use cases, but as soon as you have a mobile use case where objects need to move from one place to another, it doesn’t work anymore because you don’t have continuous service.

What we have done at Sigfox is we developed the company on technology that allows us to have a very wide area network and low power consumption at a low cost of deployment. The idea of low cost of deployment is that we are able to deploy the network in a country with something like 20 to 30 times fewer base stations than a cellular network. In a country like France, you can use 2,000 base stations to cover the full country. In a country like the U.S. it is like 15,000 base stations. You can have much lower cost than the cellular network. Also the base stations are much less expensive.

Why would IoT base stations be less expensive?

The reason why they are much less expensive is they are dedicated to the Internet of Things. They do not require huge data rates like you have on a base station for a cellular network. And then all the operation is specialized.

Given the technology, we started to deploy the network in countries and equip and build the network so we are able to receive IoT [traffic] from many, many places.

Eutelsat is an investor in Sigfox. What do you think about this interest from a satellite company?

There are two reasons why it is interesting that we have Eutelsat.

Each of our base stations is connected to the internet, and it’s connected via a wired connection like ethernet or DSL. But in many places when you are far away from anything, the only way to backhaul the base station is through satellite, so there is a huge interest to have a good connection with satellite companies to do the backhaul.

There is another interest — and many satellite companies are interested in doing this — which is IoT from satellites. The idea is that IoT is a global market and a lot of use cases require connectivity from very remote areas. It doesn’t make business sense to put Sigfox antennas in every single place in every country because it becomes too expensive, and there are some areas where there are very few objects to connect. It doesn’t make sense to have antennas over them. Nevertheless if you could have some kind of global coverage of these wide spaces through satellite, it could be interesting. And of course there is an economic equation to solve between how many satellites you need to launch, what is the cost of the satellite, and how many base stations on the ground. This is a complex situation and working with the satellite companies can help solve these equations.

The Internet of Things is mostly about connecting sensors to the internet so we are able to sense activities on the ground, whether it is sensing soil moisture, or heat or chemicals in the ground. In logistics you are sensing the position of the vans. It’s more data on the field and sending it to the cloud. It is mostly an uplink activity.

In the Sigfox network we have the opportunity to do uplink and downlink, but the majority of the traffic is uplink, which means that in the case of the satellite, we are talking about objects sending data to the satellite as opposed to the satellite sending data to the objects as the major activity. It’s a bit the reverse of the natural satellite operations. I think it is at the same time a challenge and an interesting technological activity.

Satellite is often looked at as a last resort for connectivity. Is that the same for IoT?

My personal point of view, talking for myself here is it will not be the primary, it will be a last resort kind of approach. Regardless of if it is a cubesat or not, I think even if the cubesat is bringing the cost of satellite quite low, it is still not so low, and the challenge satellite has is the fact that even the [low-Earth orbit] satellite will be covering huge areas of the globe and has to deal with all of the objects in each area in parallel.

If we are talking about massive IoT, regions where you have a lot of objects, I have doubt that you are able to take all of the same objects at the same time on the same satellite because the capacity you need to do that is tremendous. Either you have a huge bandwidth to be able to do that, which is expensive, or you need to reduce the footprint of the satellite. There is a tradeoff that is not so easy. Having satellite will be more for areas where you have less density of objects and where it is more economical to go by satellite.

What does a typical Sigfox customer look like?

It can be anything. We have hundreds of different [kinds] of customers. We have people in office information, trucking and logistics, in factories, even some reservations in Africa. Wildlife reserves use our technology to track animals. It’s quite a huge number of use cases whicha re in many, many industries.

You’ve written that IoT today is about transforming things into better versions of what they already are. How does that evolution shift from individual optimization to interconnected optimization?

Today we achieved the first steps. We have not yet arrived at the interconnection I believe will happen.

IoT is about the digital transformation of companies today. It’s about improving a process through using digital tools like the internet. We have done a lot with the internet already as companies, but you can go beyond with the Internet of Things by optimizing and generalizing the collection of data from the field. You have a direct input of data from your information system and you can make decisions based on that.

There is one example with a customer of ours in e-commerce. They want to use our technology to track all the parcels that they send to customers so that they are able to provide the service to customers where the customer can track the parcel in a much better way than what is offered today by the FedEx’s and DHL’s of the world.

What I believe is that in the future, when you have many many different systems which are connecting data and bringing all this data to the cloud, you will be able to use data which was created, which was meant to solve a specific vertical problem, and mix and match all this data to create more value.