United States Broadband speed ranks 12th according to SpeedTest Global Index | Hacker News Reader

United States Broadband speed ranks 12th according to SpeedTest Global Index

speedtest.net

66 pointspidgin123113 comments

That’s…not terrible?

Getting broadband in rural America is rough, and the permissive atmosphere for monopolistic behavior by ISPs is maddening. Pricing for internet access is way worse than in East Asia or western and central Europe. Some folks can’t get real broadband at all, and they’re disproportionally lower down on the socioeconomic spectrum. That’s not fair. These issues are real.

But overall, 12th sounds…pretty reasonable? Those top ten spots will always be filled by nations with highly concentrated, relatively wealthy populations. 5-10 years ago, the US was in a much worse place comparatively.

We moved to a small rural community where a T1 isn't even available. I can get a 3Mb DSL connection or satellite. I have both because I need to redundancy.

I did call to see what it would cost to have 'real' internet brought out here. They were happy to bring out a fiber line from the nearest town to serve our area. But it was going to cost me around $5k/mo with a 3 year contract for a 10Mb dedicated line. It's cheaper to do point to point relays from that far.

The real issue is that the telcos have been collecting fees for years specifically aimed at rolling out broadband into rural communities and they have done nothing but pocket the money.

It's cheaper to do point to point relays from that far.

I would encourage you to look into this more, if you haven't already. I set up such a connection for my in-laws who live in rural Oklahoma, and it's been a resounding success. It still has issues, but it's worlds better than the antiquated satellite options they had before.

They'll likely transition to using Starlink when it's available as the point to point connection (3 hops for them) is difficult to troubleshoot since I'm a thousand miles away, but if Starlink isn't an option for you for some reason, a point to point connection can be a great choice.

I prepaid for Starlink 4 or 5 months ago and just waiting for my turn to get it.

We did talk about setting up a community point to point system but the 'old guys' that run the area are more interested in getting (forcing) more people on to 'city' water than anything actually useful because city water fees are the only thing keeping this town alive.

If good broadband is so important to you why did to move to an area where it's not available?

It's common for folks to have tightly constrained moving choices.

We recently moved into our one and only choice. We beat lottery odds to get it and it was one more choice than most here have (doz-100s applicants for each listing).

Because I also have a wife that wanted to live in the country? Where people live is rarely about 1 single thing, it's a compromise between all the needs and wants. For us, living in the country on a small farm checked an awful lot of boxes. And we were "told" that there was decent internet here - 25Mb was normal. After moving we found out that 2 or 3 was normal and TDS just flat lies about their offers.

Ah bummer! But I hope Starlink works out really well for you.

If you don’t mind, could you provide any more info on how to find providers for point to point relays? I didn’t even know this was an option and I’m having a tough time finding any information that isn’t specific to large businesses.

In my case (very small town), we shared a connection with a local business (feed store) that had a tall structure we could mount a pair of radios on (grain elevator), so we're not really using a "provider".

Sorry to not have better info for you, but hopefully someone else will see this and chime in!

Not the OP, but the people I've seen this do it DIY have all done it with Ubiquiti products. A single link can get a mile of range with clear LoS. Depending on how far out you are, that lets you connect to a "real" internet connection for cheap.

Mikrotik products are very popular among people running WISPs, and their product range reflects this, with a whole range designed specifically for that purpose.

Even their standard wall/ceiling-mount AP is weatherproof and comes with pole mount attachments.

Yeah, this is definitely a common experience & a real problem. Rural broadband _access_ can be spotty towards terrible. But it's probably not captured here--in fact, that might skew 'average speed' stats _higher_, since there's fewer folks willing or able to fork over money for those slower lines.

> and they have done nothing but pocket the money.

Oh what slander. Verizon, for example, has been paying that cash out as dividends.

> a T1 isn't even available. I can get a 3Mb DSL connection

3Mbps DSL is faster than a T1. That speed also suggests you're 3-4 miles from a CO, so it's not that rural. 3 miles might even make a fiber run possible if you're willing to pay.

Right, but it doesn't go up and down like a jackhammer when the weather is rainy, windy, sunny, hot, cold, or dry. And yes, we are 3.1 miles from a CO and the nearest fiber drop is 14 miles from here. The service guys are at the CO resetting equipment to get me back online so often I know them all by name and they know what the problem most likely is when my address comes up on a service ticket.

If a complicated and highly capable internet connection was important to you, why did you move to place where you almost certainly wouldn't be able to get it?

Starlink might be a decent option for you. You said you have satellite, but that has me thinking HughesNet.

It's shockingly good. I'm actually very proud of this statistic. Most won't see material difference between 200mbps vs #1 spot of 250mbps. It's negligible.

However.

I think there's a rather huge sampling bias, because the data is gathered from people who come to Speedtest.com to run the test.

I'd also be more interested to the the first or fifth percentile rankings. I'd bet that we do pretty poorly, especially in terms of the 10mbps benchmark.

Edit: these stats also exclude people who have no internet at all (which is a subset of speed test users, but an important subset).

Seconded on sampling bias. It's a bit difficult to get a directly comparable figure from the FCC's Measuring Broadband America report, which tries to be representative. https://www.fcc.gov/reports-research/reports/measuring-broad...

I guess the closest would be "In September-October 2019, the weighted average advertised download speed was 146.1 Mbps among the measured ISPs, which represents a 100% increase from 2017 and a 8% increase compared to the average in September-October 2018 which was 135.7 Mbps."

I'd wager most people's WiFi routinely stays well below 200Mbps anyway.

>I think there's a rather huge sampling bias, because the data is gathered from people who come to Speedtest.com to run the test.

Any sampling bias would apply to all countries, not just the US.

I used to live in rural and the best to offer was a wimax connection. The problem was there wasn't enough people locally to justify building a second tower to handle more people, but there were enough people to slow down the single tower they all shared.

Ultimately it would have still made money for the company, just not enough, gotta get those earnings up!

I also tried satellite providers but they price gouged via bandwidth caps. After about 50gb used for the month you could upgrade to their ultra high bandwidth package, or accept dialup speeds for the rest of the month.

I still have plenty of friends who live in these places, and it really has an impact on life. People still use phone books and have regular cable tv. Only recently have some started to cut the cord.

Wired broadband in rural America has historically never been socially efficient and is mostly a patronage-based subsidy scheme. I say this as a nominal beneficiary of subsidized rural broadband.

That said, with the advent of starlink, reasonably high-speed internet in rural areas is now socially efficient, or getting there rapidly. I’m currently using starlink at my rural house and it’s great. It’s obviously still subsidized by venture capital, but there’s a very plausible route for it to continue with no subsidies whatsoever (unlike rural wired broadband).

I do see good things in the future for internet in rural areas. Between Starlink (maybe even other competitive satellite in the future?), 5G, and WISPs getting better and better, effective internet access in traditionally underserved areas should really start becoming both available and affordable.

With smaller towns/rural areas having a much lower cost of living and the expanding ability to work from anywhere with a decent internet connection, I hope we begin to see a resurgence of small town America. Not everyone wants to live in giant megacities.

>Those top ten spots will always be filled by nations with highly concentrated, relatively wealthy populations.

Romania, Thailand and Chile is top 10...

Chile is not a counter example. Almost all the population lives in the middle third of the country, and even then the Santiago area is drastically more dense.

I had similar feelings. At least in populated cities, internet speed has always been pretty great for me. I'm more concerned with the absurd notion of bandwidth caps and the exorbitant cost. Access is tricky: on the one hand, it's pretty unreasonable to expect internet companies to get cables out all the way to every mountain cabin and desert house; on the other hand, they charge so excessively and get so much government assistance that they should have more than a few dollars to spare.

On reflection, I'm curious how much of this is directly caused by US demographics shifting towards greater urbanization.

1 Singapore 262.20

2 Hong Kong (SAR) 254.40

3 Monaco 242.89

4 Switzerland 222.00

5 Thailand 221.00

6 Romania 217.91

7 South Korea 216.67

8 Denmark 216.13

9 Chile 209.83

10 France 201.61

11 Hungary 201.55

12 United States 199.00

First of all, Singapore, Hong Kong and Monaco really can't be compared to the US. Two of them are wealthy city-states and Monaco is very small (and wealthy).

Once you get to #4 Switzerland, the difference is negligible. The difference between a 200mbps connection and 222 is barely noticeable. It's more than enough to stream movies and use the internet at home, and it's way too slow if you are running a large tech heavy operation. But if you are in the latter group, chances are you aren't doing it at home and much faster commercial broadband of 1GB+ is readily available in most major cities. And if you are like most places, you probably do most of your heavy lifting in data centers belonging to Msft/google/amazon, where the internet speeds are significantly faster.

It seems to me that the important metric is not average speed, but % of people with, say, reliable connections of 50mbps or faster. That's probably how much you need for reasonable home usage with with a few TVs streaming. I don't think most people will notice the difference beyond that.

> but % of people with, say, reliable connections

Agreed. It's hard to really know how useful these stats are. First, who are the people using speedtest.net? I'd assume there is some kind of bias in the person using speedtest.net. But also, if you can't get any broadband, does that count against the stats?

Second, what's the distribution? Is it some people with 10Gbps+ and then some with 1Mbps?

Third, how is upload factored in? With e.g. video conferencing/home office/remote learning, symmetric internet is becoming more important, and latency plays a part.

And finally, how expensive is it? Personally, I think it's not great claiming e.g. Monaco has amazing speeds if a lot of people can't afford it.

If these are averages, then they are most likely skewed by outliers. I want to know the median values, and various interesting percentiles, like 5th, 25th, 75th, 95th, 98th, 99th, etc….

12 United States 199.00 down / 72.7 up

I'm fairly skeptical of this figure, with most sq mi of the US unable to get anything like that.

Further, the typical upload speed for a 200Mbit cable connection is ~10Mbit. I can't see enough symmetric providers to skew the avg upload numbers that far.

note that we don’t need 50mbps (and certainly not 200) for most homes/families either. 25mbps comfortably supports 3 1080p streams with room left over for web, email, and other apps.

Even ideal, uninterrupted deliveries don't happen in a steady stream. Data arrives in bursts. I've seen 25Mbit connections choke on far less.

God help anyone with a 25Mbit/down cable connection who tries sync their media to the cloud. That <5MBit upload will not only take forever, the saturated outbound pipe impedes ACK packets for incoming traffic - which forces inbound video and other connections to stutter while waiting for their endlessly retransmitted ACKs to arrive.

Real life also commonly includes gaming. Steam updates alone max'd out our 500Mb connection.

i've been sharing 25mbps with another person and the occasional guest for the past couple years, often including multiple video streams at a time. there have been some relatively rare upstream buffering issues (as most issues are with wifi), but overall, it's been just fine.

gaming doesn't seem like a good case for making a general bandwidth argument. sure, if you want to spend your money on bandwidth for it, go to town, but the general populace doesn't 'need' it. video at least is used for communication, education, and other non-recreational purposes.

It's not about the 50 mbps download bandwidth here. It's a combination of bandwidth, latency, and variance in latency on both the upload and download links. Usually residential 50 mbps is a peak download bandwidth of 50 mbps with pretty high variance latency. 50 mbps symmetric with low latency would be fantastic among 3 individuals, but when realistic speeds are 30/5 mbps down/up with bursts of high latency, you're in for a frustrating time.

I was on 25mbps during lockdown with my parents for a few months and it was very painful. You probably don’t need the 500 I get now but 25 is not good.

wait, the US is $12 with 199mbs, trailing Singaport at 262?

This is a useless comparison and in fact shows that the US has deployed near globally maximum speeds to a huge populace.

I often see people commenting on the poor quality of America’s internet compared to European countries, but how much of that is due to the physical size of America? I imagine it’s much more difficult to build out a larger network especially when many parts of America are not very populated.

Looking at the countries above the USA... They all mostly have one thing in common, their populace lives almost entirely within a few primary concentrated areas. It is much easier to build out the infrastructure to only support a handful of major cities which certainly helps raise the average.

edit: I want to make it clear that I think the fact that I can get gigabit internet in a town of 300k smack in the center of "flyover country" is absolutely amazing. I think that the USA competes with these other countries despite the size disadvantage is nothing short of astounding.

This is the usual argument against doing large infrastructure in the USA, and I think it’s kind of missing the point: the USA also has much larger financial and physical resources than those countries. We also have counterarguments: we have a universal electrical network. We built a large railroad network across the entire country. We have an interstate highway system reaching every continental state (and with “interstates” in Alaska and Hawaii as well, if I’m not mistaken). Also, universal telephone service.

I live in the heart of Silicon Valley, and I can’t get cell reception at my house. I have basic broadband, but I can’t get fiber despite being on a street where other houses have it.

I would be very interested in a comparison of what the citizens of each country have paid to fund internet service. US taxpayers have given the telecom mono/duopolies a ton of free money and right-of-way and the providers have broken every promise to provide universal service in exchange.

Railroads, highways, and power/cable lines all go hand in hand. It's an amazing network and they were all built a long time ago with little improvement. The problem is shoveling money to other "universal" slush funds while ignoring actual infrastructure for decades. ISPs should have been a utility and they should treat it like Texas treats its grid. Have a company that manages the actual lines, and have multiple providers that can compete and sell data on it.

Something like this was actually tried in the '90s for local phone circuits ("local loop unbundling"). As far as I know, it basically worked and is technically still policy, but the type of access mandated is no longer competitive for most purposes.

> they should treat it like Texas treats its grid

like garbage that falls apart when people actually need to use it? or like a regressive money funnel to line the pockets of an oligarchical captured regulatory body?

texas' power grid is not a goal any serious advocate can advance as doing anything right.

Texas' grid is the reason why 100% of the dollars I spend on electrical energy goes towards renewable energy sources at $0.084c/kWh for energy+transmission. What do you pay and what mix is renewable?

Texas' grid also makes it incredibly easy for you to get your local solar/wind connected to the grid to sell your electricity back, further incentivizing local and renewable energy production.

Finally, when Texas' grid has issues with big events which knocks out power, it captures national headlines for months on end. When California has blackouts, its just California being California.

https://www.dailysignal.com/2021/06/25/blackouts-loom-in-cal...

Maybe there's more states than just Texas having to deal with the effects of climate change.

>Finally, when Texas' grid has issues with big events which knocks out power, it captures national headlines for months on end. When California has blackouts, its just California being California

Meanwhile, the other 48 states, including a tiny island and a giant tundra wasteland larger than Texas have no problem keeping their power on in adverse weather.

I have family in Michigan, Ohio, Kentucky, and other states. They've definitely lost power in major snowstorms before, sometimes for multiple days. I've lost power from hurricanes and other kinds of weather events elsewhere. Losing power from big weather events isn't an incredibly uncommon thing.

https://www.nbcnews.com/news/weather/over-1-million-lose-pow...

https://www.msn.com/en-us/weather/topstories/death-toll-rise...

https://weather.com/storms/hurricane/news/2020-08-06-isaias-...

Were you really trying to suggest the other 48 states never lose power in extreme weather situations?

The differentiating factor is that most states at least attempt to prepare and design for the understood risks, unlike Texas. I'm not saying other states do not lose power in extreme weather events, but the weather that hit Texas was closer to a ten year event than a hundred year event

You literally stated:

> have no problem keeping their power on in adverse weather.

It sure reads like other states have no problem keeping their power on and not other states sometimes do lose power as well but hey at least they're trying. And besides, hurricanes hit the East Coast pretty dang often and yet still hundreds of thousands of people will lose power. Sure sounds like they're not trying hard enough if they should have no problem keeping the power on.

Also, as someone who has lived in Texas for 30 years, this storm was absolutely not a ten year event. It may become one with the changing climate, but I have never experienced an ice storm quite as extreme. The major winter storm that happened ten years before was still not close to the one this year. For the one ten years prior, most pools didn't freeze over. Even with the pumps running on my pool, I had nearly an inch of ice form. I've never seen it get that cold in Texas.

If you're referring to the 100 year winter freeze, yes maybe it should have been winterized, and they are now, but that's not what I'm talking about.

Maybe all California homes should have A/C for these heat waves. Sometimes it doesn't make sense economically.

I'm talking about having a transmission provider and having ISPs compete on the same lines.

It keeps prices down (unless you tricked into buying variable wholesale via startups like Griddy) and it keeps lines in good shape.

Uh, no. The Texas grid is not winterized now.

They’ve pledged to start work thinking about doing something, sometime soon, maybe.

That’s about as far as they’ve gotten.

> USA also has much larger financial and physical resources than those countries

Certainly not on a per-capita basis, as far as Norway and Luxembourg are concerned.

The top 3 - Singapore, Hong Kong (SAR), and Monaco - are smaller than Rhode Island, the smallest state - even when you add them together.

My gigabit internet in HK was similarly priced to my gigabit internet in suburban USA (like US$80/mo). That said, in HK areas with robust competition between ISPs, some of the offerings were shockingly cheap. A friend of mine got 10Gbps for less than US$100/mo after negotiating. I don’t remember if HK lines were symmetric or not - my US line certainly isn’t.

Note that symmetrical gigabit connections in the US are sometimes available. Its highly dependent on the type of network deployed. If its a PON deployment there's a good chance you're going to get a symmetrical connection (AT&T, Verizon, Frontier Fiber). If you have a coax connection (Spectrum, Comcast, Cox), there's little chance its symmetrical.

For example, at my home on a residential PON connection a recent speedtest today showed ~990Mbit down, ~978Mbit up. At my office's new business coax connection just a few miles away, I usually get ~980Mbit down, 30Mbit up.

LOL they are smaller than the Dallas/Ft. Worth area alone!

Yup, but I wonder why there were bandwidth issues in Europe when the pandemic first started (video streaming services started throttling bitrate very aggressively in Europe) but there were no issues in the United States.

Perhaps the last mile average is slow but the overall infrastructure and backbone is more robust in the US (this would make sense historically as well)

Wireless internet is more common in European and Asian countries because you can put up a few towers and be done, it gets throttled when you have congestion.

America has fiber and mostly copper (from our phone line laying days) strung all throughout America, the bandwidth is much higher in total.

The job left here is replacing copper w/ fiber (or at least DOCSIS 3.1 via copper), but it's expensive and some ISPs it isn't worth it for the customers they have, so you have some areas bringing the national average down.

Starlink will be nice for really remote areas where it make no sense to run a line or a tower for a couple customers. Existing sat internet is unusable. There's also been a big push to modernize slightly bigger towns with fiber by smaller ISPs.

As far as I can remember, the stress was mostly put on exchanges more so than overall capacity. I'm not aware of any real degradation beyond the panic in the news about « running out of Internet ».

There was no issue in Europe either, it was a political decision, not a technical one.

Now take those same countries and look at GDP, number of publicly traded software companies, and capital investment provided by governments. There is some valid critique about how poor the USA is in terms of bandwidth. Not to mention cultural appropriation by most Americans of "being the best" at everything. Thus there's a reason (maybe not 100% valid) why people complain the USA isn't #1 on this list.

While density may be part of it, my (admittedly: anecdotal) experience in splitting my time between Los Angeles and the nearby mountain communities over the last year is that LA broadband speed is consistently 5 times slower.

I live in Baltimore - could certainly be considered a city - internet speed is appalling even with "business service".

The thing about rural America is nobody lives there. More than 80% of Americans live in urbanized areas, the same fraction as France, Norway, Spain, and the world generally.

The reason America doesn't have broadband or trains is not because of its extent, it's because our war budget is $2000 per capita per year.

"Urban" in the US can still be quite different from "urban" of Europe or other places around the world.

I live in a really close suburb of one of the biggest cities in the US. I technically live in an "urban" area, and places further out from the actual city that are less dense are still considered "urban". The population density here is ~4,000 people/mi^2. A similar comparison to Paris location-wise would be something like Le Blanc-Mesnil, which has a density of 18,000 people/mi^2. Both of these are considered "urban", but they are extremely different kinds of cities density-wise.

"Our urbanized places are garbage" is a completely different statement from "our country is large". One of them is circumstance and the other is self-inflicted.

Whether or not "our urbanized places are garbage" is very much subjective. I don't consider the place I live to be garbage, personally I prefer it to living in a place as dense as Paris.

Whether you think US "urban" areas are garbage or not, your statement doesn't change the fact that "urban" in the US versus "urban" elsewhere can be pretty radically different. Saying "more than 80% of Americans live in urbanized areas, the same fraction as France, Norway, Spain, and the world generally" is kind of comparing apples and oranges. Yeah sure 80% of the population lives in locations labeled "urban", but the average densities of those urban areas in Europe are much more dense than the US. I'd imagine there are plenty of places in the US classified as "urban" which would easily compare to the densities of "rural" areas in Europe.

I do agree with the idea that a partial part of the problem is a lack of focus on infrastructure spending compared to other budgets such as the military. Hopefully that will change soon.

As far as I can tell it's not that America as a whole has bad internet, it's that we mostly put modern cable/conduit/equipment in neighborhoods that are either new or rich. There are places where brand new fiber and rusting DSL are separated by 100 feet.

I would like to see some metrics that take into consideration landmass, population distribution, grid scale, or the like. Singapore's road network is roughly 3500km vs the United State's 6,850,000km.

https://en.wikipedia.org/wiki/List_of_countries_by_road_netw...

I live in a semi-rural area, on a dirt road off the highway with 15-20 other neighbors on our half-mile stretch of street. Until Starlink hit, the best internet we could get was 500kbps unmetered, and maybe 4mbps metered if you were lucky. It was genuinely that bad.

I think you're highlighting the huge challenge of American broadband. Every dirt road off every highway that has a dozen or two homes needs broadband -- that is an immense amount of infrastructure that's needed.

Do they have electricity? I'm betting "yes". This means there are either above ground poles that fiber can be hung from, or if it was constructed in the last 30 years, the ground was opened for trenching, and a fiber should have been buried along side the power (wou'da, cou'da, shou'da). Dig once policies and pole sharing policies have been fought tooth and nail by the entrenched utilities.

I can attest you can find bad internet in small cities just the same as in rural places.

Sure, size is a factor, but it's not the only factor, nor is it the main factor. The main problem is monopolies.

While it's a point that intuitively makes sense, I'm not too sure of its true relevance. The US is just too rich for size to be an excuse. At least it can't be the only reason. It may be anecdotal, but all the complaints I've ever seen about broadband/fibre in the US have been in relation to local monopolies and the general fragmentation of the nation-wide network. In most EU countries, we have laws that force network owners to mutualise (they have to allow third parties to deploy the "last mile" so that people can pick any ISP within three months of a new local). Infrastructure and commercial operations are legally separate. I think the US would benefit from something like this on a federal level, at the risk of being called communism.

I think the issue is that the problem is being generalized into a 'one size fits all' sort of problem. When the issue is a bunch of items. Rural is an issue. But so is duopolies. So is poor competition. In the year 2000 I could pick from 20-30 different ISPs. Now I can pick from basically 3. Woe unto you if those 3 decide it is going to cost you 20k just to do a hook up or just say 'not available' even though all of your neighbors have it.

That's because the government forced Bell/AT&T phone lines to be leasable by other companies. Cable and fiber lines are _not_ required to have that kind of competition, which is why Comcast often delivers the only high-speed connection in a suburban area and can price gouge you.

Cable had that for a small time. But I think maybe 2 other companies actually did it. It is a mess and a lot of it I suspect is as you suspect. I think DSL lines had the same thing for awhile too but that got removed as well.

> The US is just too rich for size to be an excuse.

We have access to cable internet and shortly fiber (from Metronet). We have some friends in an established (and not poor mind you) neighborhood not 2 miles away that doesn't even have cable run. They can't get ANY internet if it's not wireless.

It baffles my mind.

I travel internationally for work. My experience has been that while pockets of Asia (the large cities) are great, once you get outside of those cities it collapses. In europe, Internet access seems to be almost universally "meh". Australia is pretty good (although I haven;t gone rural there, and Australia, while land-mass rich, is still pretty concentrated in a more limited number of rural areas).

The big knock I have is the cost per byte in the USA, especially once you consider caps. That's a huge huge issue. Like everything right now in the US, its haves versus have notes.

What I find interesting about this is how embarrassingly low the United Kingdom is considering its history as a telecom power house. In my eyes, and maybe I'm reading this data wrong, it really showcases the devastating impact the privatization of our services has had on our country. We have basically stagnated on speeds since the early 2000's.

I still have the same speeds I had 10 years ago (100Mb/S with no opportunity to improve it). My family who live in the countryside (keep in mind in the UK, countryside can mean less than 10 miles from the nearest city) have been stuck with the same provider with the same near unusable broadband for the last 7 or 8 years. It's a sorry state of affairs.

What provider are you with that is stuck at 100Mb? and what's the countryside speed/provider?

I have had gigabit in the rural homes and the flat in London that covers the last 6 years, before that is was Virgin media.

We actually have ~40%(1) of the country now on gigabit capable infrastructure many of those can choose between gigabit from multiple different first mile infrastructure providers. The government project 85%(2) gigabit coverage by 2025, I think we will get higher than that from all the plans I have seen. The biggest problem we in the Altnet world (non BT/Virgin media FTTH) is take up. Most people don't actually care and are happy to just stay on the cheapest 60Mb Sky/BT/NowTV/etc package for £22 a month. We see about 10% takeup/interest in areas with Virgin and FTTC, about 15% in areas with FTTC.

https://www.ofcom.org.uk/__data/assets/pdf_file/0028/218881/...

https://commonslibrary.parliament.uk/research-briefings/cbp-...

You’re entirely right about the cause. BT even had a plan to roll out fibre quite early, but Thatcher’s neoliberal government cancelled it in favour of privatisation.

I live in rural Michigan. My nearest "big city" has a estimated population of 2800. No big ISP has run anything our direction. We are just now getting municipal gas run in our area, our house has been on propane since it was built in the 90's.

My current internet connection is provided by wifi access points strung across neighbors roofs that terminates into a T-Mobile data connection somewhere. I get on average 8Mbps down and 2Mbps up. On a good day those might be 12 and 3. It goes down frequently enough that I have a wifi power switch hooked up so I can toggle the PoE and reset the router that never seems to recover on its own.

I work remotely of course, and video is highly pixelated and voice frequently cuts out and stutters. Thankfully I don't work on _huge_ git repos.

Netflix and Youtube adaptive streaming is good enough that their content works well. Streaming from Plex is terrible though, I need to set it to the lowest streaming setting to view anything.

My next best option is "Hughesnet" which is satellite and gives you 10gbp transfer for $40/m the most you can get is 50gbp transfer at $130. No unlimited option. I'll eat 10gb in a day.

I'm dreaming of fiber, and have signed up for Starlink which.. hopefully is a little bit faster and stable than what I've got if it ever arrives.

I also have a place in the rural Upper Midwest (not MI). Our community-owned electric utility has been very aggressively putting fiber out to all the farms and villages in the counties who participate in this cooperative. I thought it would take years to get to my farm, but I was surprised and incredibly happy when put fiber onto my property within weeks of me filling out a form declaring my interest. Now I just have to pay a modest installation fee to get it put into the house.

>I live in rural Michigan ... I work remotely of course

Why? Wouldn't reliable broadband access be a primary consideration for someone working from home?

Consider that I moved here from Alameda, CA where I also only had DSL. I have yet to see "reliable broadband access" anywhere I've lived in the US (5 states).

Even though we have "choice" its generally: Comcast Cable or DSL (some other "subpar" option).

I detest Comcast and wish our country had actual ISP competition. And Comcast isn't available at my address either.

It’s hard to tell if these results hold any meaning or not. Where did the numbers come from? Are these just averages from their users? If so, I’d imagine a huge selection bias pushing these numbers up. Are these households only or do they include data centers? I’m betting the US has a higher density of servers per capita than most countries.

Using the median or including standard deviation seems like the only way to approach an even slightly useful picture.

Yeah, it's just averages of whoever happens to go to speedtest.net and do a speed test in a given month, with a minimum of 300 users for a country to be listed.

For some of the countries, if you click through they give medians for sub-entities (states/provinces or cities). As you'd expect, they're significantly lower than the means on the front page. Not sure why they don't give country-level medians.

A few examples for major cities (median downloads in Mbps): Austin 182, NYC 169, Milan 153, San Francisco 144, Calgary 130, DC 122, Beijing 121, Paris 119, Miami 118, Shanghai 117, Toronto 99, Minneapolis 96, Rome 90, Montreal 65.

Important to note, most broadband providers realized they were getting hammered by speed test in the ratings, and so launched their own (fake) speed tests that show them over performing what they actually can do.

So customers suffering from bad internet are more likely to be directed by their ISP to other services rather than speedtest.

I happened to recently stumble upon a report[1] describing FTTH/B deployments in Europe. Personally I find that much more interesting/meaningful than simple Speedtest average (as others have noted its issues). I think Spains good result (87% coverage) there is fairly impressive, especially considering that it as a country has plenty of low-density regions.

For comparison, the best data I found[2] on US FTTH situation is that there are about 40M homes passed and 14M subscribers for FTTH out of 125M households, giving 32% coverage and 11% penetration.

[1] https://www.ropa.de/wp-content/uploads/2021/05/FTTH-Council-...

[2] https://www.fiercetelecom.com/telecom/ftth-will-pass-60-mill...

What is wrong with being 12th?

The numbers all look pretty massive anyway and my 70Mb/s is fast enough for me to work from home and stream video.

>What is wrong with being 12th?

It's more that people are outraged by the US being 12th, because they have it engrained into their heads that US's broadband is in 100th place or something. They thus come up with nonsense like "selection bias" as excuses for why their presuppositions must be right and Speedtest is wrong.

It seems that for a subset of people not being considered best is an insult.

One thing though that is often brought up is that the US situation would be improved by having more competition. Unfortunately it seems that the regulators are turning a blind eye on this problem.

This number isn't very useful for the purpose a lot of people here use it for, ,to determine the quality of broadband in a given country.

I'm using Singapore as an example, where 1 Gb/s fibre is available to almost every household (the government ensured that two fibre pairs were installed to each household, and the resident can buy any service they want from different ISP's).

The number shown, at less than 100 Mb/s must mix in tests over mobile as well, which begs the question as to how useful it is.

This might be the 'tail wagging the dog' but one of my criteria for a retirement location is the presence of fast, municipal broadband. My parents live maybe a quarter mile outside of the city limits, and all that's available is 5mb dsl.

ATT has made it plain that they would be more than happy if everyone on their street canceled their internet. They've stated the network my parents are on is legacy, won't be maintained and you can't sign up for it today.

Whenever it comes to comparisons between the United States and other countries of the world, it is best to keep in mind the difference in mass between the two. I put together this map to illustrate what I mean: https://thetruesize.com/#?borders=1~!MTY2MDk5MTk.NzY3NTM1NA*...

As you can see, ALL of the countries that rank higher are significantly smaller than the United States. In fact, some of them are even smaller than some of our smallest states. And to truly capture the point, you can see all countries can fit within the United States with plenty of room to spare. The fact that America can rank so highly with citizens so spread out, a situation of which has its own scalability challenges, is rather impressive.

Size doesn't matter though, I think the measure you're looking for is population density and perhaps terrain (thinking of Switzerland's mountains here).

Population density and terrain are important factors as well. Including those two other factors on top of size actually supports my argument further. The United States is significantly larger, less dense, and has a wider range of terrain/elemental conditions (ie. mountains, wetlands, deserts, forests, etc.). The fact that the United States ranks this high with these un-optimal conditions is again rather good.

And to address your point that size does not matter is incorrect. Size certainly does matter. It is more cost effective to service a small region of land where high amounts of people are concentrated than it is to spread infrastructure thin across great swaths of land to service the few.

> your point that size does not matter is incorrect. Size certainly does matter. It is more cost effective to service a small region of land where high amounts of people are concentrated

Size as a standalone statistic does not matter. You're correct in this quote but it contradicts your earlier point: you earlier put it forth as a separate metric with that land area tool to compare different countries without looking at population at all. I would explain it further but I think we're on the same page, just not sure what makes you think what I wrote is incorrect.

Ah, yes. I see what you mean. Thanks for clarifying. Yes, I think we are on the same page.

Do these stats represent anything useful? First of all it's going to be biased towards people who test their connection speed. I could imagine that skewing the numbers upward. But even as a relative metric it doesn't say much because we don't know what the distribution is in different countries. Some of those averages could be bumped upwards by a small number of extremely fast connections.

You should consider population density when evaluating broadband penetration:

    Country      Mbps    People per km2     
    ---
    Singapore    262.20  8,041
    Hong Kong    254.40  6,677
    Monaco       242.89  19,341
    Switzerland  222.00  207
    Thailand     221.00  135
    Romania      217.91  82
    South Korea  216.67  511
    Denmark      216.13  133
    Chile        209.83  25
    France       201.61  118
    Hungary      201.55  104
    US           199.00  34

So most of these countries have population density ranging from twice as high to several orders of magnitude higher compared to the US, with the exception of Chile, which has 1/20th of the US population.

What the ranking obviously isn't showing is variability in the bell curve within a country. In the US, you can get fiber in big motropolitan areas, but at the same time, there are plenty of areas in the US where you just don't get any connectivity at all.

By default speedtest only reliably measures last mile speed, since it defaults to the nearest server it can find. That data has its uses, but it doesn't tell the whole story, and is often misleading. Backbone, stability and practical latency tend to be much bigger issues, which are not represented here at all (last mile latency is represented but not even accounted for in the ranking, nor is upload bandwidth).

You could make a much better index using speedtest by taking several measurements for each ISP using near and far nodes, and using weighted averages accounting for upload, download and latency.

I work from home in the U.K. with 50Mbps downstream connection. Slow by these standards and I don’t think I have ever found the broadband speed a limitation.

For me ‘always on’ and low latency are important.

As long as congress continues to willfully ignore cable monopolies with the help of lobbyist friends, it’s a cesspool. Why does this https://en.wikipedia.org/wiki/Comcast even exist in its current form.

I live in a city with an excellent home grown independent fiber ISP which really should be used as a case study for such things around the country.

https://usinternet.com/fiber/coverage-map/

That's extremely good for a country the size of the US that can hardly compete with the likes of Singapore or Hongkong.

Rural broadband vs. Urban broadband? The US has a much larger geographical area of rural communities than the countries ranked higher. Hong Kong for instance is 1/3 of the size of Rhode Island with 7x the population. Much easier to give good broadband speeds with high population density in a small area. Hell Monaco is one of those countries and it covers less than 500 acres...

It’ll be curious to see how this progresses with Starlink.

I live in the country, work in a fairly data-centric job (infosec) and have worked from home a majority of the time for the past 15 years. Until this Feb my only connection to the internet was a 15M/768k DSL connection.

In Feb I got involved with a statewide project to bring Starlink in to see if it would support rural broadband initiatives. I've been running on it since. Average download and upload speeds are 10x and 30x respectively.

Up until the last month or so, however, the intermittent outages in Starlink had been a much bigger problem than the low bandwidth of the DSL connection, and so I worked over DSL and the rest of the house ran on Starlink. It's slowly getting to the point of being acceptible and I just recently switched work over to Starlink, but there are days where i switch back because the unplanned interruptions are extreeemely irritating.

All of that said, I'm super stoked about Starlink and think it's going to do a lot for rural/mobile living around the world.

https://www.speedtest.net/insights/blog/starlink-hughesnet-v...

Reply on news.ycombinator.com