Research and preparation to bring the internet to Idjwi

In developed countries, the internet has become like a utility, but our research told us that introducing the internet into the developing world requires providers to become more service like – content sometimes need to be curated, and access needs to be gradually introduced. This is because:

  • People often don’t have the skills to extract value.
  • Connectivity is so sparse and expensive that there is a high risk of exploitation in terms of fees and devices.
  • Cultural influences can prejudice information access.
  • Safeguards are not in place to protect children or vulnerable others.

“People don’t really know what to do with the internet when you give it to them… You can’t give people a clean Google page, it’s absolutely meaningless for most of the communities we’re dealing with.” – 


Screen Shot 2016-06-16 at 21.21.28

In the west we get our internet directly but invisibly from ISPs.  In Africa there’s an interim layer of local infrastructure owners – whether it’s the internet café that brings a connection to a slum, or the single person in a household who owns a smartphone. There’s more interaction and exchange between the user and the owner.

On Idjwi, the internet needed to become a service.  We needed a solution to involve and include those without devices and with low literacy levels.  We needed to provide curation and internet skills education, and when bandwith is prohibitively expensive, provide curated content.

Internet services for Idjwi

Ensemble helped us ideate solutions for local coffee growers, miners and traders, along with educational and entertainment providers. We hypothesised that with better weather and news information, people could protect themselves better. We also knew that we must provide some kind of wider-reaching service including SMS broadcasting and public displays for wider penetration.

Screen Shot 2016-06-16 at 21.22.05

We had many questions:

  • What forms of content, information and communication feel most compelling to local people?
  • Who are the people we think could benefit from internet value, and what kinds of services would be required?
  • Does anyone there have a smartphone or laptop?
  • Who most needs to send out information, and who most needs to receive it?
  • How and where might people use a wireless mesh network?
  • How could the service be monetized and maintained?  It’s vitally important that a community owned service can be self sustaining.  We knew there could be many maintenance and operational tasks.

So we prepared a raft of research material and questionnaires for Mike to take to explain our potential services to the people of Idjwi.

Screen Shot 2016-06-16 at 21.22.20

Screen Shot 2016-06-16 at 21.25.09

Understanding the technical challenge

Huge thanks must be given to Arjuna Sathiaseelan and Jon Crowcroft of Cambridge University’s “Networking For Development Lab” for their initial advice and steer on the steps we would need to take.  I sincerely recommend that if you are interested in connectivity in remote and developing locations, that you review their work.

Screen Shot 2016-06-16 at 21.21.45

We were advised to firstly understand the lay of the land in terms of placement of antennas and realistic procurement of equipment. We decided to download terrain data and a 3D model of the island so we could plan positions for antennas near villages and houses.  We needed to research radio transmission frequencies, fresnel zones, weather and other natural barriers.

Screen Shot 2016-06-16 at 21.22.40

Directing WiFi long distances required us to develop an understanding of the physics of point-to-point WiFi communication, develop solar power requirements along with gaining an understanding of antenna mast construction, and lightning protection.

Screen Shot 2016-06-16 at 21.23.07

The island if Idjwi lies directly in the lightning capital of the world, so we knew we would need thorough protection in the system. Sometimes there is not much you can do about the forces of nature other than protect people as best you can.

Screen Shot 2016-06-16 at 21.23.15

It’s likely that equipment will be damaged and needing to be replaced.  This illustrated to us again how important it will be to create a sustainable business and operating model to keep the network maintained by the community. Modelling the kiosks in 3D helped us illustrate to our Ensemble colleagues the kind of public access point and solar power requirements we would need.

Screen Shot 2016-06-16 at 21.22.58

Anders Olsson from the charity Falling Whistles, was able to scout these locations on Idjwi and together with Patrick & Mike from Ensemble, discuss and arrange the support of land owners. It was vitally important that we had line-of-sight between our links and the source of our Internet, Liason Congo Internet Cafe in Bukavu, 50km away across the lake.  Water is the enemy of wireless,  making the connection especially challenging.

Screen Shot 2016-06-16 at 21.25.25

Using the Ubiquiti networks mapping tools, we could evaluate the right antenna equipment and point-to-point links. We discovered we would need four antenna masts to get the internet to Bugarula, the bustling market town we had decided to position our kiosk in, strategically place near the town’s only bar as this was a natural meeting point for the locals.

Screen Shot 2016-06-16 at 21.23.30

Sourcing the equipment

We were be able to source solar panels and batteries, and build the antennas and kiosk locally, but we knew robust networking equipment would be harder to locate. We decided to use the funds from the Fjord Innovation Fund grant on quality Ubiquiti equipment for these initial links. These links form the backbone of a future internet mesh on the island, so we felt it was important to make that backbone as robust as possible, before expanding on our network with home-made or lower cost antennas in the future.

For our longest link we purchased two Ubiquiti RD-5G34 RocketDish antennas, with Radomes to protect against wind load.  Our studio manager, Britt, then had to contend with receiving and storing some oversized deliveries to our office in central London for quite some time.

With the equipment in the building for prototyping, we knew we needed to test everything.  But for transmitting WiFi at 5GHz long distance you need regulatory permission.  You certainly can’t test it in Oxford Circus.  With visions of us blacking out the BT Tower, we thought it best to find a more rural location to understand how all this stuff works.

We decided to take them north, to test in Scotland.