SpaceX’s Starlink satellite broadband project seems to have thrown down a pretty definitive gauntlet; one that should cause furrowed brows from rural ISPs, as well as approval from the FCC’s Rural Digital Opportunity Fund (RDOF). However, the $500 upfront equipment fee does leave a bit of a sour taste in the mouth.
But for many consumers, getting even a tenth of that top-end 120Mbps throughput is impossible, not for love nor money (although one early user claimed a 174Mbps connection). In the rural areas that Starlink is targeting, not even dial up is guaranteed, and often, what these consumers can get access to ends up being multiple times the cost of the equivalent broadband package in the nearest city.
Not that all rural ISPs are crooks, but collectively, that industry has not done a good job keeping pace with urban services. The major MNOs are slowly migrating towards using 5G networks as fixed-wireless too, and if Starlink is successful, there’s a pretty good chance that the bulk of rural America never heads down the copper-to-fiber upgrade path – and jumps straight to wireless. This is a lesson that rings true in many other parts of the world.
Such a move would force major consolidation within the ISP market, with most rural fixed-line providers being starved out in short order. From a competition and monopoly perspective, there should be serious concerns about this consolidation happening apace, with a serious risk that the MNOs can carve up cities and counties in a similar fashion that cablecos divvied up regions into little fiefdoms.
Those concerns are perhaps a little premature, and so we turn our focus to the latest development in the SpaceX story. A wonderful Reddit user decided to take his SpaceX beta hardware for a stroll, hooked up to a battery pack, and carried out some tests for our enjoyment.
We have previously reported on the average speeds that these beta testers were getting, but there’s something more impactful about the images of the plucky little satellite dish hooked up to a laptop and a battery. Hiking out to the woods of a national park in Idaho, this tester hit downloads of 120 Mbps, uploads of 12 Mbps, and a latency of 37ms.
At home, the speeds reached 135 Mbps, 15 Mbps, and 41ms, respectively, in conditions described as “limited obstruction.” The tester moved the setup to a location with “significant obstruction,” consisting of bad weather, treetops, fences, and houses, and reported 46 Mbps, 15 Mbps, and 41ms.
Out in the woods, the tester placed the dish under a tree canopy, to see how badly it would affect performance, and found that it could only sustain a connection for around 5 seconds at a time. They also suspect that the dish would not be able to connect outside the cell that it was registered to, meaning that while you can certainly take the dish out to the woods not far from your home, you probably aren’t going to have much luck if you try to take it for a long drive.
As for set-up, you apparently only need to plug the satellite dish and router into the power supply, and let the dish automatically work out its best orientation. The only user inputs are the WiFi network name and password. Besides having to put the dish on a roof, that really does sound exceedingly simple – but then again, that’s a pretty astute way of describing the technological capabilities of the average consumer.
The motorized dish’s design is notable for its lack of a LNB (low-noise block downconverter) and boom arm, and also seems to have less of a curve to it. The in-home WiFi router is an angular design, and looks a lot smaller than the home gateway you’d expect from a cableco. The power-over-ethernet cables supplied have also been given a visit from the design department, and SpaceX is definitely trying to chase the premium feel.
This is wise, after all, because it is asking for some $600 upfront, and there is still a chance that most users won’t enjoy speeds like this Redditor. But if you don’t have a fixed-line broadband option, which FCC figures suggest some 40% of the US does not, then satellite is the only way to go. CEO Elon Musk has noted that reducing the cost of the dish is SpaceX’s biggest broadband challenge, so we don’t expect much to change on that front.
The question is whether SpaceX can achieve the same performance once it starts adding significant numbers of users per cell. There is a finite amount of satellites it can hold in orbit over a location, meaning that there is a fixed upper threshold for the number of users it can reasonably support, before performance begins to deteriorate.
The question within this question is whether it would be worth bringing the top-line speed down to add more subscribers to the balance sheet. Would getting 4x the customers at a quarter of the speed make more sense for the bottom line, or are these triple-digit throughputs available on fat-enough profit margins? Can the incumbent rural ISPs hope to lower their prices accordingly?
There have been some attempts at aggregation already. PCMag put together some Ookla data points, from an earlier point in the beta program, and found an average download rate of 79.5 Mbps, with uploads of 13.8 Mbps. It then compared these results to Viasat’s Exede and Hughes’ HughesNet, which hit 24.75 Mbps and 19.84 Mbps, and 3.25 Mbps and 2.64 Mbps, on download and upload, respectively.
But the most impressive distinction between the incumbent satellite providers and Starlink is the latency. Of course, physics is on the side of SpaceX, as its constellation is operating in a Low Earth Orbit (LEO), and so its signals have a much shorter distance to traverse. The geostationary orbit that Viasat and Hughes are dealing with result in latencies of 643ms and 728ms respectively. What’s more, SpaceX reckons it can reduce its latency from 20ms to 40ms now, down to 16ms to 19ms in 2021.
The current roadmap for Starlink is expanding the beta program, still focused on the northern part of the US and southern part of Canada. In January, the plan is to open the service up to the rest of the US, with European operations commencing in February and March, with India due by mid-year. We would not be surprised if the timelines are revised.