This is the third part of our three part series on the evolution and deployment of universal fast internet service. The first part talks about our growing need for faster internet, wherever we are, and the second part looks at Elon Musk’s new Starlink internet service.
We now look at competitors and alternatives to Starlink – other ways to get fast internet outside of population concentrations.
The concept of using many Low Earth Orbit satellites to provide internet service is not new, and Musk’s Starlink is not the only company scrambling to deploy this type of service. But it is the one closest to going globally live and offering service to significant numbers of users, all around the world.
There are two other major startups, one partly funded by the UK government, and the other an Amazon/Bezos venture. There is also a Canadian venture (Lightspeed, with plans for a 298 satellite constellation, to go live in perhaps late 2023, and promising to offer very high bandwidth speeds), and a Boeing venture too, with plans for a constellation of about 3,000 LEO satellites.
The startup with UK government backing is OneWeb, the British government getting involved as part of a bankruptcy rescue last year. Currently it has 254 satellites in orbit, and it seems it will need 650 in order to start service, and hopes to have about 2,000 when their constellation is complete. Their satellites will be at about twice the height of Starlink satellites, making for slightly greater latency. They plan to provide 50 Mbps download speeds, and expect to start limited commercial service late this year.
The Amazon off-shoot is Kuiper Systems. They plan to create a LEO constellation of 3,236 satellites at a similar altitude to Starlink. They plan to offer speeds in excess of 100 Mbps, and claim to have tested their satellites to 400 Mbps. They say they’ll need about 580 in orbit to start service, and as best I can tell, they currently haven’t launched any.
Just this week, Amazon announced it had bought out the satellite communications venture that had been started by Facebook. Facebook is turning away from the concept, and this is thought to help Amazon move ahead more quickly with its plans, which at present seem unlikely to see any degree of commercial service offered in the next few years.
Amazon’s satellite ambitions probably tie in to its Amazon Web Services division – the most profitable, but least visible part of the Amazon empire, and the largest provider of virtual/cloud computing services in the world. Indeed, even this web page you are reading now has been hosted on an Amazon Web Services computer for part of its life and journey from my computer to your computer. It seems to make great sense for Amazon to further expand their cloud computing services to have their own “internet backbone”, and we’re not sure how “end user” focused the Kuiper network might be.
More than one commentator, however, has speculated about it also tying in with Amazon’s video streaming service, and wondering if Amazon Prime might end up including internet access in the future, too. That’s a fun and obvious thing to speculate, but extremely unlikely, due to the same inability of Amazon’s Kuiper product, just like Starlink, to provide high density service to many users in a small area.
It is good to see these competitors evolving, but for now, Starlink seems to pretty much have the entire market to itself, much the same way that Tesla had the BEV market to itself for some years too, and only now is starting to encounter competition.
If you’re hoping for satellite service, it seems your only choice for the foreseeable future is Starlink, although it might be worth keeping an eye on what OneWeb achieves later this year, even though it seems like its bandwidth will be much less than Starlink. The Amazon product is also interesting, but unlikely to be available for several more years.
Where Starlink and its Competitors Don’t Work So Well
Elon Musk himself concedes that Starlink isn’t intended as a universal solution for everyone – he says it is intended as a viable option for about 5% of all internet users. As we calculated in the second part of this series, Starlink has limits on how many users it can handle per square mile (or any other unit of measure) of surface area.
While the exact number of users Starlink can handle per square mile is unclear, we guessed at perhaps in the order of ten. Certainly, a worst case scenario would be more than one per square mile, and a best case scenario might be as many as one hundred. While, over time, it is possible that the Starlink satellites will become more capable, it is also inevitable that people’s internet usage will increase.
While the growing connectivity in our lives will add some to our need for internet bandwidth, the main driver of increased internet usage will probably be video streaming – moving from 720P or 1080P HD video to 4K UHD HDR video will mean moving from to rates of 5 – 10 Mbps to perhaps 25 Mbps for each video stream. That’s a huge jump, and also, whereas the internet can handle “over capacity” demands for things such as email, chatting, and web browsing (because they are “bursty” in nature – a few seconds of usage, then a period of silence before another few seconds of usage, and if the few seconds of usage extend to a few more seconds, it is not too obvious), streaming a two hour movie represents a solid steady demand for an unbroken data stream at whatever streaming rate is required. It is not possible for a service provider to “cheat” and squeeze more than the rated capacity out of their service when each user has a steady load rather than an intermittent load.
So, let’s keep that 10 users per square mile. To put that in more meaningful terms, a square mile is 640 acres. So the service could support one user per 64 acres – that’s not too bad in areas of very low density farmland, but if you’re in a rural-residential area with 5 – 10 acre lifestyle blocks (ie 64 – 128 potential users per square mile), you are already at or above the maximum density that Starlink can handle.
And for those small country towns of a few hundred people – they’re never more than some fraction of a square mile in size. A square mile is huge – Manhattan is not quite 23 square miles (and has 1.63 million people in it).
What this analysis shows is there are actually at least three levels of internet service required, not just a simplistic binary high/low density model. There are truly high density uses, and there are truly low density uses. But what about the “middle ground” – country towns and rural lifestyle living – densities too high for Starlink to service, but too low to make it affordable to bring in a terrestrial solution based on cable or fiber or some other type of physical connection?
There are some existing and some new solutions.
Other Technologies and New Services
We’ve already spoken about the inadequacies of the current two geosynchronous satellite based internet providers, Viasat and HughesNet. They’re usually the worst solution in most cases, but, at present, sometimes they are the only solution, at which point you need to switch to seeing your glass as happily half full.
There is another type of service that is sometimes offered in rural areas – fixed location microwave services. The easiest way to think of this is as being a bit like a satellite based service, but instead of being hundreds, even tens of thousands of miles away, up in the sky, it is maybe five or ten miles away, on a transmission mast. This type of service tends to be limited to reasonably short range and are line-of-sight – if you can’t visually see the microwave station from where you’ll put your antenna, you’re unlikely to get much of a decent signal.
So this type of service works best in relatively flat rather than hilly areas. It is also limited to how many users it can serve per microwave site – sometimes you might be told “We have good coverage in your area, but we have no spare capacity”. While speeds can theoretically be very fast, usually they are slow to moderate because they are trying to find a fair compromise between the number of people they provide service to and the speed of service to each person.
A new and exciting type of service being rolled out now is fixed-mobile service. Sounds like an oxymoron, doesn’t it! This is where you have a fixed transceiver at your location, and it connects with nearby cellular towers – just like a cell phone, but from a fixed location, and on a different type of plan – sometimes giving you unlimited data per month, and perhaps even for a very reasonable fee.
This is becoming more common with the increased bandwidth and speeds capable via 5G wireless service. Here are a couple of articles about T-Mobile’s new 5G based fixed wireless service (one two); similar service is being rolled out also by both Verizon and AT&T.
This service is interesting – it seems they have contracts with the three major wireless companies (AT&T, T-Mobile and Verizon), allowing them to offer appealing unlimited service plans for $150 – $250/month. They’re a bit vague about connection speeds, referring to “as much as 80 Mbps” (and “as little as 1 Mbps”) and they do say it depends on your location and the towers nearby and how congested they are. But if you’ve 4G/LTE data available where you live (and no other options), this “Unlimitedville” service might be worth checking out. Be sure to compare their service and cost with what you could direct from the wireless company, of course, but don’t be surprised if it is cheaper to buy through the middleman rather than direct.
This article is also very helpful – it tells you how to replace the standard antenna in at-home units with a more powerful antenna, enabling you to get better/faster signals and service.
However, there’s a limitation to this type of fixed cellular service – if the wireless company doesn’t have towers close to you, or if it does, but they don’t have spare capacity, then of course it is not helpful to you. I know in Washington, Idaho and Montana, there are large areas of territory with no cellular service at all, or, if present, only with weak signals and slow 3G type service – okay to use your cell phone on for primarily voice calls, but useless to power your home or office internet with.
Happily, the “sweet spot” for this type of service seems to be in that middle grey zone between high and low density – areas where there are enough existing phone users to justify adding a tower, but not too many as to mean the tower is fully utilized by regular phone users.
There’s another potential type of internet provider – drones that fly through the air at relatively low altitudes, getting the power to stay aloft via solar cells mounted on their wings. Instead of being 300 miles above the ground, they might be “only” 70,000 ft (13 miles) up (making for very short latency). That’s twice the usual height of regular passenger planes, and definitely well above the weather so their solar cells can get close to full uninterrupted solar power during each day.
A British startup plans to build drones that could stay up for a year at a time. And at 70,000 ft, each drone could have very small compact service areas on the ground, making it easier to provide better coverage to more people per square mile. This has tremendous appeal, also because you can vary the density of coverage, and with drones at low altitude, it is easier to have smaller service cells on the ground, meaning more people can be serviced per square mile.
We also suspect that a drone, using sunlight to fly, and operating many repeated one year flights, would be cheaper in terms of capital and operating costs than a Low Earth Orbit satellite with a 5 year lifespan prior to then being destroyed.
These days, fast internet service is almost an essential element of life and necessary for us to be able to fit in to the expectations of modern society. Imagine not being able to order on Amazon, video call to friends or business colleagues, or to stream video. Imagine if it was not possible to update your Windows operating system with its latest 100MB+ system update, because your internet speed was too slow and service too unreliable. Scary thoughts!
Unfortunately, while our need for faster and faster internet grows every year, the cost of providing this also grows, too. The solution, in areas where few people live, is for new technologies and new ways to serve the needs of people in such areas.
Starlink and, eventually its competitors too, promise a wonderful new type of service for some lucky people, many of whom have been suffering from little or no internet access up until now. We expect this will unlock a lot of the currently constrained potential value of rural land and encourage more people out of the cities and into the “great outdoors”.
This was the third part of our three part series on the evolution and deployment of universal fast internet service.
If you’ve not already read them, here are links to the first part, talking about our growing need for faster internet, wherever we are, and the second part, looking at Elon Musk’s new Starlink internet service.