In the '80s, satellite comms showed promise – soon it'll be a viable means to punt internet services at anyone anywhere
We've come a long way from Terry Wogan's celeb interviews
Feature Terry Wogan has a lot to answer for. From 1982 to 1992, he presented an eponymous chat show on BBC1 where he would often interview celebrities via satellite in front of an unconvincing backdrop of the Hollywood sign.
Back then, everyone had a single word to describe satellite communications: slow. The interviewer would ask a question, two or three seconds later you would see the interviewee's eyebrows twitch as they heard the first word, and then a few more seconds later the mouth would begin to move and the words spouted forth.
Satellite communications in the late '80s and early '90s were cumbersome, but they worked, and they were pretty much the best you could get for TV-quality video communication. The first transatlantic fibre cable, TAT-8, didn't exist until 1988, with a princely 280Mbit/sec bandwidth, so flinging TV over wires hadn't yet been done.
Satellite had the bandwidth for half-decent data speeds such as those required for video or other lumps of ones and zeroes – but latency was the killer. So someone had a bright idea: what if we use satellites for non-interactive traffic streams that were largely one-way and weren't a slave to latency?
In 1980, a thing called Usenet was established. It was a bulletin board based on Unix technology, which was becoming more and more popular at the time, and used UUCP – the Unix-to-Unix Copy Program – to transfer articles from system to system. But it was realised that Usenet traffic was highly asymmetric: for every kilobyte you uploaded there were probably several megabytes of inbound data from everyone else. Connectivity was either dial-up (measures in kilobits, not megabits, per second) or contemporary wide-area, fixed-link tech such as X.25 (tens of kbit/s at most). Simple answer: use satellite for the downstream service as the bandwidth was better... but nobody actually wanted it, so it never became a thing.
Wind on to the mid-1990s and the internet had been invented – then the World Wide Web happened. And guess what... someone realised that like Usenet, web access was highly asymmetric. A few tens of bytes of an HTTP request would elicit many tens of kilobytes of response in the form of the requested page. Unlike Usenet, there was an element of time-criticality – rather than just squirting stuff in any old order, the response had to come shortly after the request – but the idea was that one could use satellites to provide the downstream responses for web requests. And you could see the point: new 56Kbit/s modems cost per-minute money to be connected. This correspondent, in my time as technical editor for a networking and telecoms newspaper of the '90s, was offered a chance to review such a solution. As it would have involved mounting a 1.6-metre satellite dish on the roof of our office – which would've required planning permission – we politely said "no, thanks." And that wasn't a bad thing, given that this product didn't take off any more than the Usenet solution of previous years had.
You need sea for sub-sea cables
At this point, more and more undersea fibre was being laid, with increasing amounts of bandwidth. With traffic traversing thousands of miles under the sea, rather than tens of thousands of miles up to a satellite and back, latency fell like a stone and the average round-trip time between the UK and the US became – and remained – around 100 milliseconds. Modern sub-sea cables abound, and although they cost a few quid, the price is sufficiently affordable that large companies lay their own (such as "Firmina", between the US and Argentina, announced by Google in June 2021). The tech that used to be the plaything of big national and international telcos is now affordable and open to all.
- The coming of Wi-Fi 6 does not mean it's time to ditch your cabled LAN. Here's why
- We're terrified of sharing information, but the benefits of talking about IT and infosec outweigh the negatives
- Monitoring is simple enough – green means everything's fine. But getting to that point can be a whole other ball game
- Tolerating failure: From happy accidents to serious screwups … Time to look at getting it wrong, er, correctly
Yet with this all said, there is still a problem: you can only run sub-sea cables where there's sea. Running cables across landmasses is expensive and often difficult (and the more difficult, the more expensive). So satellite is back in the race as a communications technology for the average person. And this time it will work.
If you consider satellites such as the ones that deliver our Sky signals, they're in orbit a tad under 36,000km from the surface of the Earth. And they're expensive. The radical new approach uses Low Earth Orbit, or LEO, satellites, which orbit at a 20th of that height. And with less height comes much lower latency if one bounces a data stream off it – 20 to 40 milliseconds, according to Starlink, the LEO satellite network being built by Elon Musk's SpaceX.
And this is what makes a difference, along with the bandwidth (20-50 megabits per second). This isn't up to the kind of speeds you can potentially get with, say, 5G mobile telephony, but that isn't the point: the point is that it's absolutely fast enough for most users, and it's particularly attractive for those who live in sparsely populated land where no telco can afford to implement a decently fast cabled or mobile broadband network.
The vast majority of people with cabled broadband services have performance that's reasonably good. Just as we pointed out that, in an office setting, we're silly to try to use Wi-Fi when cabled networks are achievable, the same applies outside the office. So where cabled networks are fast enough, we will stick with cables; and where they aren't – or where they are not available at all – satellite has finally reached the stage where it's ready to step in and fill the geographical gap.
This said, though, there's a problem with LEO satellites providing data services: you need a lot of them to bring decent speed service to all the populated areas of the Earth. Amazon's planned offering, for example, is reckoned to comprise a little over 3,200 satellites at heights of between 590 and 630 kilometres (as a comparison, Amazon boss Jeff Bezos' Blue Origin passenger craft tops out at a height of 100km). SpaceX has a bit of a head start with about 1,700 satellites deployed by the late spring of 2021 – and could end up with tens of thousands.
SpaceX's rocket advancements make it economical
Why has this only just happened? Because it has only recently become possible or, more accurately, economical. One estimate of the cost of a SpaceX satellite is around a quarter of a million dollars – which is a lot to most people reading this (and certainly to the person writing it) but is relatively small change in a telecoms setting. With reusable, inexpensive spacecraft now in existence to launch relatively economical satellite tech, this is the first time such an approach has been possible.
And yes, the bill for all this stuff will run to tens of billions for SpaceX, Amazon, and anyone else who thinks they can find gaps between all the little shiny boxes that Jeff and Elon throw into orbit over the next year or two.
But think back to the days when 3G mobile technology was new in the UK, and the government put the radio spectrum up for auction back in 2000... and raised over £22bn. And that was for a mobile network in a country with a population (back then) of just under 59 million people – we're now looking at not much more for a global solution with a market of, theoretically, over six billion. With Musk reckoning that he'll have half a million users of Starlink by the middle of 2022, there seems to be a market for such services too.
There's just one thing, though. Yes, we are finally at the stage in history where satellite technology is (pretty much) available to the masses, and at long last actually works at a speed and latency that's usable, even in the 2020s. But one regime that is considerably more developed than the next generation of satellite data services is green technology – along with the general awareness of the environmental impact of the tech we buy and use. The useful life of these dinky LEO satellites is sub-10 years, according to submissions to the US Federal Communications Commission, and LEO satellites are not really the kind of thing one can throw on the compost heap or take to the council recycling centre.
So, at long last, satellite technology is finally something that really is bringing proper, high-speed, usable internet connectivity to the masses. It will change the lives of the millions of people worldwide who simply cannot make the most of cabled or land-based mobile connectivity.
But expect the debate about green technology to ramp up... along with the moans from all the space agencies around the globe whose kit will have to tiptoe around all this shiny stuff some kilometres up in the sky. ®