Satellite internet: orbiting the obvious

Oddly, or perhaps not, I know quite a bit about rural internet – not as a technician but as a user, since I actually live in a rural area, as opposed to most politicians who pretend they do, but don’t. 

When I first moved into my rural digs near Prince Albert in the Karoo, I was desperate for internet connectivity – so desperate that I became a client of Telkom. Horrors!

Telkom, though it is partly privately owned, still had an air of the Post Office about it at the time; getting anyone to actually perform a service, even one that the organisation depended on financially, was bizarrely difficult. But, like the Post Office, it was oddly functional in weird ways. 

People forget this now, but Telkom had a satellite internet service for domestic use long before Starlink was a twinkle in the eyes of starship Musk. It was called SpaceStream and I was a customer. The service promised a download speed of 512KBps, which was around the speed being offered by urban networks circa 2003.

Anyway, I thought at the time, well, job done. With SpaceStream employed, I’ll be no worse off than I was in the city. Hoo boy. It was then that I learnt about latency. Data speed, measured in kilobits or megabits, describes the bandwidth or throughput. It’s how much data the connection can carry each second once the transfer is under way. But latency is the delay between sending a request and receiving a response, and it is as important from the point of view of the user.

A high-speed highway

Think of bandwidth as the width of a road: a provincial road carries fewer vehicles at once than a multilane highway. Latency is the length and difficulty of the journey – how long it takes one vehicle to reach its destination and return. A road can be enormously wide but still take hours to traverse. For a useful internet connection, you need both capacity and responsiveness.

Latency basically made the SpaceStream service unusable for practical purposes. My huge SpaceStream dish – rather like a DStv dish but larger – eventually blew off my house in a Karoo storm. How completely unsurprising that a state-owned enterprise (SOE) could offer a service that was technically brilliant but actually defunct. 

The reason for the latency is that it operated off geostationary satellites 36,000km in the sky, and bits of data travelling that distance take a bit of time. This is why Starlink is a low earth orbit system, and its satellites are only about 550km from earth. But because they are closer, you need many more of them to make the system usable. There are currently more than 10,000 active Starlink satellites in low earth orbit, but that is only a fraction of how many the company intends to put out there. 

So, to get usable internet to rural dwellers, the obvious solution is Starlink. But the political class in South Africa is so outraged that Elon Musk objects to being racially pigeonholed that Starlink has been effectively banned. There are all kinds of other arguments, but IMHO the fact that Musk is using his considerable prominence to highlight South Africa’s racially discriminatory laws is really what irks our racially obsessed politicians. 

Enter Amazon

The government’s answer has been to insist that Starlink is not the only satellite in the sky. Trade, industry and competition minister Parks Tau has invoked the European system Eutelsat OneWeb and alternatives as proof that other low orbit systems exist. OneWeb does have roughly 650 satellites in orbit, and lowish latency. But it’s essentially an enterprise option and it’s not cheap. And it doesn’t come close to Starlink’s heft. 

Now Amazon Leo has arrived as the more plausible alternative, punted recently by DA communications minister Solly Malatsi.

Amazon announced this week that Leo would launch a South African residential service in 2027 with Herotel, under the brand name “evry”. Herotel will sell, install and support the service through its network of 120 offices, while Amazon supplies the satellites and technology. Financial details have not been disclosed, and it’s actually not commercially operational anywhere at the moment. But Leo solves the political problem that Starlink has so far refused to solve: it arrives wrapped in a local partnership.

This is all great, but to me it’s like a parable. The preponderance of ideology is preventing the logical solution, which would be Starlink.

Telkom’s slow progress demonstrates somewhat the flat-footedness of SOEs and their gradual decay in the hands of cadre deployment. In desperation for alternatives and to not look like they are technological troglodytes, ANC government figures cite other possibilities, like OneWeb, without acknowledging their weaknesses. The DA favours the private sector “dealmaking” solution in Amazon and Herotel, as one would expect.

They are all, essentially, fantasising.

It all just fits. It is like a snapshot of politics and technology in South Africa, and it partly explains the country’s sad decline.

ALSO READ:

• What Ramaphosa, and the ANC, get wrong about Starlink
• How to play Africa’s satellite revolution
• Starlink: don’t debate Musk – debate rural connectivity

_{Top image collage: Pexels/Susan Hartzenberg; Rawpixel; Currency.}