Common Sense Skeptic: DEBUNKING STARLINK - Part 1
Part 1 - Starlink vs the Competition
Disclaimer
I really don’t think this should be necessary, but here we are. I am not an employee of any of Elon Musk’s companies. I do not have a financial interest in any of Elon Musk’s companies. I do not worship Elon Musk. I do not think he is infallible. He says and does things that I find weird, unlikely to happen, or plain immoral. This is not a defense of Elon Musk. This is a critique of a poor debunking attempt by Common Sense Skeptic.
Introduction
On July 4, 2021, Common Sense Skeptic released this video attempting to debunk Starlink. For those unfamiliar, here is a little bit of background.
Starlink is a satellite internet constellation being developed and deployed by SpaceX, currently in beta at the time of this writing. It is not meant to be a replacement for traditional broadband, but rather to provide broadband service to those who don’t have any other available options.
Common Sense Skeptic is a YouTube group aiming to “provide viewers with the ability to break down outrageous claims using simple mathematics, examples, and logic.” The vast majority of their videos center around Elon Musk and his companies, with a heavy emphasis on SpaceX.
At 0:42 in their video, Common Sense Skeptic lays out the main arguments against Starlink: the economic model, benefits over competitors, and hazards of operation. In order to better understand the arguments presented by Common Sense Skeptic, I believe it is important to first provide a clear understanding as to what Starlink is and how it compares to the competition, so we will be starting there.
Part 1 - Starlink vs the Competition
Part 4 - Conclusion and Score
The Realities
Besides Starlink, there are two main providers of satellite internet in North America.
HughesNet, with ~1.2 million subscribers in the US. (pg 31)
Viasat, with ~590,000 subscribers in the US. (pg 41)
Both companies operate satellites in geostationary orbit(GEO), 35,786 km (22,236 mi) above the Earth’s surface. Starlink, on the other hand, will operate satellites in Low Earth Orbit(LEO) and Very Low Earth Orbit(VLEO), 335-570 km (208-354 mi) above the Earth’s surface.
There are benefits and drawbacks to each approach, but what we’re concerned with at the moment is the comparative performance of each network.
- Latency
Latency is the amount of time it takes a signal to reach its destination. The theoretical minimal time it would take a signal to reach a satellite in GEO and return is around 0.25 seconds. For LEO, it’s 0.002 seconds. For a variety of reasons, real-world performance doesn’t actually achieve those times. The median real-world performance for the three companies is:
HughesNet: 0.724s
Viasat: 0.630s
Starlink: 0.045s
Because Starlink satellites are so much closer to Earth, they introduce very little extra delay compared to a ground-based system, and due to the difference in the speed of light in a vacuum vs in a fiber-optic cable, can actually beat land-based performance (more on this later).
While this might not seem like the biggest deal, it’s important to keep in mind that this latency will apply to every packet of data sent and received. Now, I’m not a network engineer, but simply watching some Youtube videos of HughesNet/Viasat speed tests will show the problem. Nearly a second of a delay here, there, here, there, etc… can easily turn into a 10-20+ second delay before the thing you want to do starts actually happening.
- Data Caps
Because any system is limited in how much bandwidth it can handle at any given time, and because adding additional capability to a satellite in GEO is so much more difficult than doing so on the ground, satellite internet providers have established data caps on their plans. This prevents ‘power users’ from single-handedly(or as a group) hogging the majority of the available bandwidth, but can seriously constrain how a customer has to ration their internet usage throughout the month. The current data caps for each of the three companies are:
HughesNet: 10-50 GB
Viasat: 10-150 GB
Starlink: No data caps*
If you’re unsure of just how much data this is, consider this. Watching 4K programming on Netflix uses up to 7 GB of data per hour, per device. A single user could use up their entire monthly data cap in just 1.4-7.1 hours with HughesNet, or 1.4-21.4 hours with Viasat.
When you use up your monthly data allotment, these services will throttle your speeds. HughesNet discloses speeds of 1-3 Mbps (which may or may not be able to stream Standard Definition content on Netflix), while Viasat doesn’t disclose specific speeds.
*Starlink currently has no data caps, and no plans to implement them at this time, although it is likely that some measures will have to be taken to prevent abuse. If Starlink does implement data caps in the future, they will have to be at least 2TB due to RDOF funding requirements(more on this later).
- Speed
Starlink and HughesNet only provide one tier of speed, Viasat has different options.
HughesNet: 25 Mbps download, 3 Mbps upload
Viasat: 12-100 Mbps download, 3 Mbps upload
Starlink: 50-150 Mbps download, upload speed not advertised.
These are merely marketing claims, however. Real-world performance gives us the following averages:
HughesNet: 19.73 Mbps download, 2.43 Mbps upload
Viasat: 18.13 Mbps download, 3.38 Mbps upload
Starlink: 97.23 Mbps download, 13.89 Mbps upload
- Pricing
Because of the different tiers of service offered by HughesNet and Viasat, it is difficult to envision exactly what a customer will get for the same money. I have created this table showcasing the most comparable plans, accounting for everything from promotions to equipment fees. Averaging out the costs over a 5-year period gives us fairly comparable plans. Note that Starlink requires the customer to purchase the equipment, Viasat requires the customer lease the equipment for $12.99/month, and HughesNet offers a choice between purchase($349.98) and lease($9.99).
So, the tl;dr is that Starlink currently provides 4-8x higher download speeds, ~7x higher upload speeds, 1/15th - 1/17th the latency, and no data caps for comparable pricing. This is important to keep in mind, as Common Sense Skeptic repeatedly calls the three services comparable.
The Claims
3:16 - CSS begins to talk about the dish subscribers use to access the service, nicknamed Dishy, and makes two interesting claims. The first is that no service personnel are assigned to install it for the customer, while showing an image of a traditional satellite being installed. The second is the claim that “most people are just putting it on the ground”, while showing a collage of four pictures where Dishy is sitting on the ground.
The first problem - installation. Dishy is designed specifically not to need professional installation. Traditional satellites have to be carefully aimed to be able to connect with their satellites in GEO, while Dishy aims itself and uses phased-array antennas to steer its connection to Starlink’s satellites in LEO. Additionally, while the service is still in beta and subject to change, SpaceX offers several different mounting options for the dish, including a ridgeline roof mount that requires no drilling.
The second problem is the completely unsubstantiated claim that “most people are just putting [Dishy] on the ground.” When browsing the Starlink subreddit, it can be seen that many people do indeed test Starlink on the ground when they first get it, but the assumption that everybody then leaves the dish in their yard is not only a bit silly, but also contrary to what many people have posted. I can just as easily make my own collage ‘proving’ CSS wrong. Sadly, this is typical of the quality of argument they make across their videos. There are no statistics, no polling, no numbers, just four pictures attempting to prove a point.
3:44 - CSS talks about how the dish is a sealed, unserviceable unit. This is true, as is the fact that Starlink is currently in beta and different versions of the dish are being developed. Nobody knows what the final product will look like. CSS then argues that if the dish or cable are damaged and need replacing, the customer might need to buy an entirely new dish. Interestingly enough, they say this while showing a Reddit post about cable damage, without including the comment by the OP that explains that SpaceX sent him a new dish, no charge.
5:22 - CSS starts to talk about the satellites themselves, and an important observation needs to be pointed out. They consistently post screenshots of article headlines and summaries, and use those as the basis of their arguments. The problem is that headlines are designed to be catchy, and often unnuanced or outright incorrect. In this case, they show multiple articles talking about SpaceX’s plan to launch “up to 42,000” satellites, but failing to understand this figure causes compounding errors in future calculations, so it’s time for more background information.
In November 2016, SpaceX filed with the FCC for a non-geostationary orbit satellite system using Ku- and Ka- frequency bands, consisting of 4,425 satellites at altitudes ranging from 1,110 to 1,325 km. Since then, SpaceX has filed many admendments to their original application, and today the constellation stands as thus:
1,584 satellites at 550 km.
1,584 satellites at 540 km.
720 satellites at 570 km.
348 satellites at 560 km.
172 satellites also at 560 km.
This is a total of 4,408, so where does the 42,000 figure come from? Two places.
SpaceX has also filed for 7,518 satellites operating in Very Low Earth Orbit(VLEO), from 335 to 345 km, using V-band frequencies, bringing the total to just under 12,000.
SpaceX has also filed for a second-generation constellation(Gen 2) of 30,000 satellites, using Ku-, Ka-, and E-band frequencies.. It is important to keep a distinction between the first 12,000 satellites(Gen 1), and Gen 2, as they will not coexist.
CSS introduces an error that compounds through their calculations by imposing the costs of not only the current generation constellation, but also its replacement on SpaceX at the same time.
10:16 - CSS now moves onto the second section of the video: advantage over competitors. They start by saying they won’t compare Starlink to fiber or cable, because that is not its intention to compete with them, but they do compare later on in the video anyway(@11:58). They then move onto the other satellite internet providers. “Since there are already satellite service providers covering the entire country, Starlink will merely be an option to one of those services.”
11:24 - CSS makes an egregious argument here. They screenshot this article from August 2020, which was before the public beta even began, and say “Starlink manages a max of 61 mbps transfer rate…” They continue to use this 61mbps figure for the rest of the video. As I have shown previously, Starlink currently has an average speed higher than what CSS claims the maximum speed is. They could be forgiven because they had to use older data than I am using now, however:
The same source CSS used has this article from October 2020 showing an average speed of 79.5 mbps down, higher than what CSS claims through this entire video is the maximum speed Starlink is capable of.
I asked CSS why they used such old data on Reddit, YouTube, and Twitter. They declined to engage with me on Reddit(as in, they posted “We had been entertaining your suggestion of a debate, but that seems rather pointless.”), didn’t respond on YouTube, and posted this exchange on Twitter before blocking me. Notice the article they link isn’t the TeslaNorth article. Also notice that the article they link lists median speeds as high as 93.1 mbps for an entire county, which again is higher than what CSS claims the maximum speed is throughout their video.
When people point out that they get higher speeds than CSS said is the maximum, they respond rather poorly.
12:19 - CSS says Starlink is suited for video calls or gaming, for a single user. They point out that using multiple devices simultaneously decreases the bandwidth available for each user, which is correct, but they fail to point the same thing out for Starlink’s competitors. They specifically make a point that having a device streaming Netflix at 25mbps, “your experience will vary”. I find this very interesting considering the topic of data caps, which I have already addressed.
They then mention that more users per satellite could degrade performance. Again, this is true, but also for Starlink’s competitors. I will go into the math behind this in the economics section.
13:15 - “Now the question is, how does that data transfer rate compare against other satellite internet providers?” This is the real question, and one CSS doesn’t do justice. They bring up Viasat and HughesNet to make their comparison, but unlike what I did earlier in this post, they seem to insinuate that the cheapest plans Viasat and HughesNet offer are comparable to Starlink. Remember the slight problem of running out of data after 1.4 hours of 4K streaming?
“Neither of these services require the customer to buy the equipment that is being installed for free, up front.” While this is technically true, they never mention the monthly equipment lease fees, or HughesNet’s option to purchase the equipment up-front. CSS provides this comparison:
As previously stated, the equipment rows are incorrect. Viasat charges $12.99/month. HughesNet offers either $9.99/month or a $349 purchase.
The speeds for Starlink are incorrect as discussed earlier. It should be listed as 50 to 150(during beta). Due to RDOF funding, the minimum speed is required to be 100 mbps within 6 years.
The “Basic Price” is misleading for Viasat and HughesNet. These basic plans are simply not comparable to Starlink:
Viasat offers the Liberty 12 plan for $30/month, increasing to $50/month after 3 months. This gives you up to 12 mbps download speed, and has a data cap of 12 GB. Using CSS’ Netflix example from earlier, your monthly allowance would run out after ~1.5 movies.
HughesNet offers a plan for $49.99/month, increasing to $59.99/month after 6 months. This gives you up to 25 mbps download speed, with a 10 GB data cap.
Once again, plans at a similar price point, equalized for equipment costs over a five year period, are shown below. The suggestion that these plans are in any way comparable is laughable.
CSS then claims that there is one reason, and one reason alone, that Starlink has their satellites in LEO instead of GEO like their competitors: latency. Again with Starlink, that ping is, on average, 42 ms. For Viasat, it is 643 ms, and HughesNet, 728 ms. CSS claims this only matters for gamers, but it also matters for people who use their connection for work, who need to participate in Zoom meetings or make video calls, and for general web browsing. When this is brought to their attention, CSS again responds poorly.
27:44 - CSS argues that people seem to have overlooked that Starlink will have to tap into existing networking infrastructure, without even attempting to validate the claim. Of course Starlink has to tie into the existing net, but CSS doesn’t stop there. They then claim that Starlink will never outperform “the ground-based nexus”, because “it still has to tie into that same system.”
So, some background is needed. A Starlink user will send a signal through their dish up to a satellite. From there, the signal will either be redirected to another satellite using laser interlinks(which are currently being rolled out), or to a ground station within range of the satellite. From that ground station, the data will continue to its destination just as it would with any other ISP. What CSS fails to mention is that this process of bouncing up to a satellite and back down to earth bypasses the otherwise ground-based path that a signal would have to travel.
Now, this can get very complicated, but let’s just look at a simple example. Here is a screenshot taken from a live satellite tracker, showing the range of a single satellite outlined in red. Additionally, I have added an image showing approximately the distance that circle covers, approximately 590 miles. Adding in the altitude of a satellite at 345 km(we’re giving Starlink the best it can provide because CSS’ claim is that Starlink can never beat wired internet), the total distance for the satellite signal to travel is ~728 miles.
The speed of light in fiber optic cables is approximately 124,188 miles per second, while the speed of light in a vacuum is approximately 186,282 miles per second. In air, that only drops down to approximately 186,226 miles per second, and very little of the satellite signal travels through thick air, so we’ll simply use the vacuum speed.
So, assuming perfectly straight paths(easier to do with satellite signals than ground-based cables), the travel time for a signal from one side of the circle to the other would be:
Fiber optics: 4.751ms
Satellite signal: 3.909ms
So, even assuming perfectly straight fiber-optic cables, Starlink can provide better latency without even using satellite interlinks. Mark Handley has created a simulation showing theoretical minimums for latency with Starlink and fiber, as well as comparing to existing latencies which account for realities such as ground relays and fiber lines that don’t always go directly from Point A to Point B. For example, he found that a connection from London to New York could be as low as 46ms with Starlink, while fiber provides a theoretical minimum of 55ms, and a real-life latency of 76ms.
CSS is wrong because they don’t account for the ground path a signal would have to take, which Starlink skips with a faster speed. Instead, they call it all “that same system.”
Conclusion
Two things should be clear at this point.
Viasat and HughesNet are not competitive with Starlink.
Common Sense Skeptic either doesn’t understand just how much worse Viasat and HughesNet’s plans are, or they are intentionally being misleading with the way they present the comparison.
Notice the way they present the information. 13:30 - “Viasat is the faster network with packages starting at 30 dollars per month, data rates up to 100 megabits per second…” It sounds as if a user could get 100 Mbps for $30/month, or close to it, when that $30 plan only give 12 Mbps down with a 12 GB data cap, and increases to $50/month after three months.
Now that we’ve gotten all of this out of the way, we can take a look at the economics, keeping in mind the difference between Starlink and its competition in terms of performance. This is extremely important when we begin discussing how many customers Starlink is likely to get.
If the services were comparable, as Common Sense Skeptic claims, then it would be much more difficult for Starlink to capture a significant market share. However, by offering 4-8x higher download speeds, ~7x higher upload speeds, 1/15th - 1/17th the latency, and no data caps for comparable pricing, well, who wouldn’t jump at that?
Part 1 - Starlink vs the Competition
Part 4 - Conclusion and Score
Very good job. Common Sense skeptic is a charlatan who has an axe to grind with Elon musk personally and therefore is extremely biased in all of the videos
You completely overlooked one important fact! Starlink has no contract or cancellation fees. Viasat and HughesNet have both. I had HughesNet, it's not even in the same ballpark as Starlink.