As this article is written, the internet - the world’s network of networks - is reliant on a backbone of optical links: very high capacity hookups working in or near the visible spectrum. The benefit of this, of course, is that such links can achieve extremely high bandwidth.
The downside of optical links is that almost anything can block or interfere with visible light; even clear air. This means that, normally, an optical link has to be carried along a fibre line. Fibre lines, often laid on the sea bed, make up that optical internet backbone. Provided your nation has enough undersea connections to cope with occasional breaks caused by weather, ships’ anchors etc, this is quite reliable.
Sometimes there are fibre links on the land all the way to actual individual users. In major cities, businesses can often get a fibre link of their own. In some countries and areas, even ordinary homeowners can have a fibre right into their own dwelling. However it’s more normal for the final stage or stages of the connection to be carried by legacy hard lines such as copper or coaxial cable: or, more and more, by one or another kind of wireless such as Wi-Fi or 5G mobile.
That’s the situation in the Northern Hemisphere, and quite a lot of the Southern, too. But things are a bit different in Africa - especially sub-Saharan Africa. Many African nations have only one or two undersea fibre connections to the wider internet, meaning that almost complete outages are routine events. Often the organisations controlling access to the fibre or fibres are ISPs and/or telcos in such a nation, meaning that they may set bandwidth prices very high and/or attempt to constrain rivals off the link altogether.
As a result, African internet bandwidth is some of the most expensive and unreliable in the world. Problems are particularly bad in landlocked nations where ISPs have to deal with neighbouring coastal states to get access to undersea links.
How much for broadband? How much money have you got?
The latest GSMA figures show that just 24 per cent of sub-Saharan Africans use mobile data: another 46 per cent are within mobile data coverage but do not use such services. Sometimes this is due to lack of electricity to charge up a device: but mostly it’s due to cost, and to a lesser degree unreliability. Many of those 470 million people have a phone or other device which could connect to the internet, but the capability goes unused. Another 30 per cent of sub-Saharan Africans don’t have mobile data coverage at all and can’t use it even if they wanted to.
The situation is even worse in fixed-line broadband. In some sub-Saharan nations it’s impossible even to assess what the price of fixed broadband is, as there are few providers and those few are prone to charging customers on a case-by-case basis.
Things are changing fast, of course, as the world’s investors are keen to do business with Africa. Many new cables are being laid around the coast and the connections inland are burgeoning. But for the next few years at least, Africa will remain very different from the rest of the world in terms of network service.
That could be an opportunity for companies in one particular sector: low-orbit satellites. Normal satellite bandwidth as seen today is delivered using spacecraft 35,786km above the Equator: geostationary orbit. Such a satellite makes one orbit every 24 hours exactly, and so from the ground it appears to sit still in the sky. It’s simple to point a stationary dish at it and achieve a fairly high-bandwidth link. If the use case is one-way space-to-ground omnicast, as with television, the ground equipment can be convincingly cheap and simple and there is only a moderate need for capacity behind the satellite.
TV is the dominant use case for satellite communications today: but you can also use a geostationary satellite to connect to the internet. You need an uplink-capable dish, the service is (very) expensive, and latency is bad enough to seriously affect things like video conferencing and online gaming. Even so, satellite internet gets used a lot in Africa.
Low orbit isn’t low-hanging, though
Low orbiting satellites are not so simple. To a user on the ground, such satellites pop up over the horizon, zoom rapidly across the sky, and then plunge beneath the horizon again. Keeping a dish pointed at one is not feasible: as a result, all the current low-orbit services can offer only low bandwidth, voice-grade at best, though at least the latency problems are much reduced. The onward links from the satellite to the wider network aren’t simple and aren’t very high-capacity either. GSM roaming has made this such a niche offering that most low-orbit companies actually operating today have been through bankruptcy and/or government rescue.
But there might be a better way. Military targeting radars used to use a dish in much the same way as mobile satcomms: the less sophisticated ones still do. But more and more, a flat phased array is used to generate a beam which can be pointed without any need for physically moving parts. It’s perfectly technically possible to build a pizza-box sized array which could simply be put on a roof - or maybe in a window - and easily pick up and lock onto low-orbiting satellites streaking across the sky. If there were enough satellites up there to have several in view at all times, the box could manage a constant high-bandwidth connection. If there were many in view at all times, many boxes could.
That’s one problem solved, perhaps. But this isn’t TV, it’s packet-based networking. We’ve got a link from the user to a node, in the satellite. But we need to hook that node up, in the end, to the backbone internet with its optical links. In Africa, as we’ve seen, the undersea fibre optical links at the moment are unacceptable choke points.
But this is the true reason for the current excitement around low orbit. The one place where an optical link doesn’t need a fibre to carry it is space, where there’s no atmosphere and nothing else to get in the way. It’s at least theoretically possible to make space-to-space optical links which would be at least as good as fibre: although, unlike phased-array radio, this is not proven technology.
Now, though, we have a vision which is especially compelling for Africa. The end user has no need to deal with anyone on the ground in Africa at all: the pizza box can reach the satellite constellation, and the satellite constellation has its own optical links and thus no need to negotiate with a possible coastal monopoly or duopoly, perhaps on the other side of a disputed or otherwise troublesome international border. Unlike normal satellite service, there’s no latency and no need to point a dish. It’s not mobile: but it could be a very attractive backhaul choice for someone thinking of setting up a 5G mast to serve that unconnected 76 per cent of sub-Saharan Africans.
The one place you don’t have to compete against terrestrial … Yet
There are hurdles for the low-orbit companies to jump. The pizza boxes can definitely be built, but they may be prohibitively expensive. The space-to-space optics are theoretically possible, but maybe not practically possible.
Then, there’s the matter of connecting the space optical network to junction points on Earth. Even if, as some space visionaries have suggested, the space optical net one day largely replaces the undersea and land fibres, things such as major data centres seem likely to remain on the surface. There will remain a need for high capacity, industrial grade connections from ground to space, and so far the new low-orbit companies don’t have an obvious worldwide answer here.
But their existing plans, for pizza boxes and space optical, could be enough to solve Africa’s problem today. There aren’t that many major data centres and junction points in Africa. In most of the world, analysts generally assess that space-based broadband will have to be able to compete for customers who already have fixed broadband options: that’s much less true in Africa.
The window is closing, as more and more fibres come ashore around the African coast. Elon Musk and all the other backers of space broadband don’t have a long time in which to get their services operational. But if they can do it, Africa might just be the springboard which lets them start to build a space backbone which could, indeed, one day replace the current undersea one around the world.