In both setups, there are multiple levels of bottlenecks. The bottom line is that a properly configured network makes sure they can cover their usual peak usage easily, and has enough spare capacity that special events don't impact much of the customer base, and those it impact it doesn't do so much.
Consider that ISPs that offer gigabit speeds can't easily put in individual (or even paired) networking equipment that can serve thousands of people at that speed, because such equipment doesn't exist or is far too expensive to be using all over the place in the network. What you really see is that 100 people on gigabit circuits during peak usage (likely Netflix prime time in the evening) will use less than even a fiftieth or h undredth of their allocated capacity (unless you're steaming 4k all the time, and then you'll probably use about a fortieth of a gigabit sustained). As usage goes up, ISPs allocate more resources and shift traffic around to make sure they can supply for the demand. If done well, you don't ever notice.
You can also see this in highways and freeways. No highway is build to serve all the possible people that could use it at the same time, it's built to service what the number of people expected to use it in a specific time period is. This is also a good idea of what it's like when planning fails, as these works require a long lead time and a lot of accurate forecasting that's hard, so you often see vastly under-provisioned systems, and even by the time they upgrade them, they are sometimes already under-provisioned at the new size (or the time until it's under-provisioned is relatively short).
I'm haven't looked into the economics of running fiber everywhere, as compared to launching an extensive constellation of satellites but I will note that SpaceX has the competitive advantage of having at-cost access to reusable rockets, compared to an outside party paying SpaceX market rates for transit to space.
Really anything is better than what’s currently available in the rural US. Having grown up there I’m kind of mad that we give so much spectrum to cell companies who block people from setting up terrestrial radio networks in places like this.
Couldn’t you pay for more than one account and combine them?
Or are you saying the cap for an entire satellite is 20GB per month?
https://www.satelliteinternet.com/resources/unlimited-satell...
We can argue about whether latency matters for most people, but it's not going to go anywhere.
Any houses for sale nearby?
Sometimes I'm not sure that's necessarily a bad thing?
But if the only way to find out what your homework is is a website, and that homework is posted as a multimegabyte PDF (this is how it's working for my kids, not a hypothetical), then lack of an available internet connection immediately affects your ability to effectively attend school, for example.
Now we can have a discussion about whether school is necessary for a fulfilling childhood, of course. :)
On the other hand, it's far worse than "20G per satellite." Bandwidth is limited by available RF spectrum, and that's not magically gonna increase with satellite count. Best thing they can try is limit the coverage area of each radio as much as possible, but there's limits to that and now you have an onslaught of handovers since the satellites are moving quite fast.
I honestly am too lazy to do the math, but my gut says this may work great for larger areas with lower population density (let's say the Amazon, U.S. Midwest, Australia, Sibiria, etc.) but it's not gonna replace high-density broadband access.
Also, just for comparison, current optical network technology (in the market, not research) puts 50G on a single wavelength on an optical fiber. There's 80 wavelengths in a medium-density DWDM setup, so that's 4T ... on a single strand. A low-density cable has 12 cores (that's what you deploy on the outskirts of your network), so, uh, 48 terabit/s with current technology. Infrastructure cables are commonly 144 cores, that's 576 terabit/s...
(These are obviously "maximum" numbers, you rarely see all 80 wavelengths or all 144 strands in use, and there's gonna be slower links like 10G occupying some wavelengths too.)
That's not true. More satellites allows you to put them in lower orbits, increasing spatial multiplexing. Yes, even for the same spectrum.
This is not like mobile where the phones have fairly omnidirectional antenna (improving to a handful of elements under 5G). The user terminals will have dozens if not hundreds or thousands of phased array elements, allowing connections to multiple satellites simultaneously and very narrow beamwidths. The satellites also have phased arrays with lots of elements.
I agree that dense cities and optical fiber will still be able to beat out Starlink (and similar), but there is PLENTY of room for using the same amount of RF spectrum for vastly greater bandwidth using spatial multiplexing and phased arrays on both ends and multiple, cooperating satellites in view simultaneously.
On the other hand, we finally got FTTH and it's great. So nice to have good speeds.
One way to look at it is up to 42 million 4k streams at the same time. But, a significant fraction of the constellation will be over the ocean or other minimally inhabited regions so something like 25% of total bandwidth is likely utilized at any one time. However, that’s spread around the globe with some areas like Hawaii having much better than average available bandwidth.
Some countries attempted to claim sovereignty over geostationary orbit, but failed http://djilp.org/the-bogota-declaration-and-the-curious-case...
SpaceX got approval to launch these satellites from the FCC.
They could sell a certain data speed to areas with low subscriber base though and provide more speed than advertised/agreed upon though.
Think how much easier it gets to stand up a tower when all you need is electricity on-site.
I offered $1/yr lease to the providers, all of them, to put a tower on my farm and not one of them is even a little interested.
I know that I feel ripped off every single day by what we pay for mobile data compared to the rest of the world. If Starlink suddenly shows up there is an alternative to Bell Rogers Telus. They should be shaking in their boots. It should be easy for Starlink to undercut their business because they made a business out of customer gouging.
Obviously you meant “bars” but I found myself wondering what kind of throughput a single Bard has, assuming they sing at maximum speed...
Assuming a single (very talented) bard can sing five words per second at maximum speed, and assuming each word is a few bytes, plus some kind of checksum occasionally, we are looking at very roughly 200 bps.
So “two bards” is about 400 bps, or far, far slower than my first modem :)
If it's not sensible, I'm sure the speed will also be correspondingly slower.
The big ISPs make things like municipal broadband very hard to bring about. This leaves connectivity of these communities in the hands of an executive many miles away, and they will always place more attention to the massive profits of bringing 5G to people in rich cities (themselves included) than servicing low or even no-profit markets.
At this point some level of internet access could and maybe should be considered a public utility.
I routinely drive through a tunnel that’s 4.5km long, sometimes at 30kph due to traffic.
Aside from that they already have plans to send up 30.000 more satellites [1] and of course future satellites might also have a significantly higher throughput.
[1]: https://spacenews.com/spacex-submits-paperwork-for-30000-mor...
It's ridiculous to assume that everyone will be using 100MB/s at all times.
An average constant bitrate per person in the evenings is perhaps around 10 Mb/s. So 25 million people could use this assuming 100% satellite use.
I certainly find Starlink interesting, but I think its impact will be a lot smaller than the author does.
https://twitter.com/chmn_victor/status/1219721896075366403
Most likely this is mainly to be used by bankers and computer stock traders to buy and sell stocks even faster over the ocean.
For an extreme example, an area with extremely low population (like the middle of the ocean), the bandwidth (at least surface to satellite bandwidth) will be wasted if no one is using it. The price could be much lower.
A poor community in a rich and population dense area probably would be priced out.
Will be interesting to see how it shakes out in practice. I could see it go both ways honestly.
For instance, my typical monthly mobile data consumption is around 4 GB, which is .0015 MB/s -- a far cry from 100!
Also the target audience of this, the people that will be able to afford it, are going to be highly paid people (assumption: it's going be charged higher than 30£/month, or even £60/mo) who will be more likely to do those sorts of activities due to having more free time.
The claimed that most people aren't going to use 100 MB/s 24/7/365 and gave themselves as an example.
Their data point is useful, as the chance of the typical user downloading 263,974 GB a month when they download 4 GB is incredibly low.
Even if they were an outlier, it's far more reasonable to think that they are an outlier by using far more data, not far less data, then typical. Their data point remains useful even if they are an outlier.
To reiterate - they never claimed to be average or an outlier, in fact there is no claim like that at all.
One data point is better than none.
- Verizon’s Beyond Unlimited plan is capped at 22GB of high-speed 4G LTE data per month, while the Above Unlimited is curbed at 75GB.
- AT&T caps both unlimited plans at 22GB of data per billing cycle, after which it may slow your speeds down
- Sprint will throttle your speeds if you exceed more than 50GB of data in a month but, as you can see, your speeds are throttled from the get-go for most things anyway.
- T-Mobile One includes 50GB of high-speed 4G LTE data. Once you go above that, as with other carriers, T-Mobile can drop this to slower speeds.
[0] https://www.cnbc.com/2018/07/13/unlimited-data-plan-caps-ver...
I don't think it is. People think they're going to get unlimited data, gigabit connection, etc etc and I think they're all highly delusional.
Starlink is not meant to replace wired internet, it's meant to give HughesNet customers something that is actually usable. It's meant to give people in remote areas with no wired connection available something that actually allows them to meaningfully interact with the internet.
Realistically I see it like this:
I think the bulk of the customers will always be companies working in remote areas such as ships, oil rigs, mining operations, gas stations primarily serving semi trucks in rural areas along highways etc. Also, first responder agencies as it will be handy in grid-down natural disasters for command centers.
Then you'll have RV parks, camp sites, rest stops, EV charging locations in remote areas, etc reselling access by subscription or in time increments.
Then you'll have folks that just live in rural areas that want something better than HughesNet and possibly THOW/vanlife folks that work remotely and move frequently.
Then you'll have low-profit/served-at-cost customers that will be schools in developing nations, especially in rural areas.
It’s not 100% of it, but it is absolutely an essential prerequisite. Slow or no internet makes places a nonstarter. Having a fallback option to a single national telco lets lots of places start the long process of building an information economy that they cannot begin at all without it.
If you were to do napkin math on which cities have the best internet speeds and the lowest cost of doing business, and the best regulations then probably we'd all have moved to Estonia by now, but people are still more likely to go to Israel, a place in the middle of a warzone.
So for that reason, I find hyping technology up as the great equalizer very cynical. It's mostly still people and culture at the end of the day.
If people can fall in love online, they can certainly collaborate without having to smell each other.
If you’re remote, they shouldn’t care where you are. It’s just a sales issue, then, and can be overcome like any other sales objection.
It depends on whether you are an employee or contractor. For a contractor, it mostly shouldn't matter. (Some countries, especially countries with which the HQ country has difficult relations, may represent unacceptable risk even for a contractor.)
With an employee, the employer has to be registered to pay tax in the country; they need to familiarise themselves with the labour laws of the country to ensure they comply with them; they may need to make various arrangements with respect to benefits (insurance, retirement, etc); usually to do all this they need to have a legal entity set up in the country, which in turn also needs things like local bank accounts, directors appointed, accounting and auditing contracts, etc. For the first employee in a country, that can be a lot of overhead. If some manager wants to hire some random person in country X where there are currently no employees, it probably isn't worth it unless they plan to hire more people in country X, or that person is someone very special.
A workaround some places use is contract with some sort of agency who acts as the employee's legal employer and handles all those local legal/tax/benefits/etc issues. However, that adds overhead costs, and if a firm doesn't already have a relationship with such an agency, they probably aren't going to start one just because some manager wants to hire somebody in a country in which there is no legal entity.
And of course, people don't use 100MB/s constantly. The ratio is not uncommonly 1:100 in max vs average throughput, and this ratio is higher the greater the max bandwidth (average used throughput increases sub-linearly with max throughput). So say about 1Mbps average per user with 40Pbps constellation-wide, and you're looking at on the order of billions of connections possible at the largest scale.
Still about 12Pbps possible at 12000 satellites. Enough for billions of connections still.
Even at 20Gbps and 12,000 satellites and 1Mbps average usage per connection, that's still on the order of a hundred million connections.
More likely, they'll have fewer-than-billions of connections with more available average bandwidth than 1Mbps. But it definitely is a gamechanger in rural internet capability.
Because they can reallocate satellite time to a new region quite cheaply, they could have approximately zero customers while forcing ISPs around the globe to behave sensibly.
I wonder if tracking down some malicious attacker will be much harder without the need for wires, although probably not too different
this reads like a marketing piece rather than actual analysis. Throughput of a single satellite is 20Gbps. The finished constellation if I recall correctly is supposed to be about 12k satellites. So about ~250.000Gbps. So assuming you offer everyone a 100MB/s connection, that's... one city? And that doesn't account for transmission between the satellites or satellites not in range and so on, and the fact that those satellites are short-lived and need to be replaced.
How is this supposed to be economical or scale compared to regular terrestrial networks?
And one point on the social impact. We have heard this idea that you just need to 'connect people' over the internet to somehow give every disadvantaged place economic opportunity for decades now. This isn't how the real world works. Everyone still runs to California/their national equivalent You can count the areas where the VC money goes on one hand. Putting a satellite in the sky over Siberia isn't going to upend the social realities of geography.