> there is cause for celebration in Dallas, Seattle and Austin, after our analysis has shown that these cities are performing extremely well relative to most European capital cities.
Seattle has some actual competition now, and many (most?) of us can get fiber gigabit to the home for a reasonable rate. It's glorious.
I hope other American cities start courting meaningful ISP competition in their cities -- but I'm not holding my breath. Many American city governments see ISPs as companies they can shake down for fees and concessions, not realizing they're only hurting their own citizens by limiting choice and increasing costs.
> Many American city governments see ISPs as companies they can shake down for fees and concessions, not realizing they're only hurting their own citizens by limiting choice and increasing costs.
And yet your typical American on the Internet still blames greedy companies for their slow speeds. As if companies elsewhere with faster Internet were less greedy.
European Internet, according TFA, is for the most part not better. The largest country above us on the list has 17 million people. The big European countries (France, Germany, Italy, Spain, and the UK, which together comprise 70% of the EU) have slower Internet.
> Regulators around the world took notice and strove to emulate the success of the U.S. approach. Denmark in particular styled its own telecommunications regulations on the U.S. model, slashing its chief telecom regulator altogether and assigning small regulatory functions to other departments.
Denmark is basically the situation where you have a “benevolent former monopoly,” like Verizon here in the mid Atlantic US.
Switzerland, right behind Denmark in that chart builds the physical networks either entirely with public money or public/private partnerships, but sets a fixed rate at which the fiber must be leased out to any interested party.
This worked too. A small ISP came along (init7, they're awesome https://www.init7.net/) and started offering symmetric gbit at $65 per month, disrupting the ISP market. The big ISPs were scrambling at first, now there's actual competition, with offers like 10git for ~$40 a month (https://fiber.salt.ch/en). Granted that's marketing 10gbit and not real 10gbit, because it's XG-PON [0], so "up to 10gbit/s" and it's a shitty ISP. But still way better than you used to be able to get.
So it would seem to me that many roads lead to rome. Eitherway I'm pretty happy because I don't hate my ISP anymore, something that was unimaginable 10 years ago.
It depends what your values are. To my values most or all of Europe has better internet. A few things I care about:
- No bandwidth caps.
- ISPs competing to offer the best service in a system similar to how health care is done in Japan.
- Being able to switch ISPs without issue.
- Being able to run my own hardware instead of being required to use their routers.
- Net neutrality. I pay for gigabit duplex but if I'm not uploading to a fast lane I'm limited to 10mbps, or 1% of what I pay for. This is terrible for VPNing for work, and other work related tasks.
Exactly.. I'm on 1Gbit down/1Gbit up at home, there are no caps, and the full speed is always available. It's not throttled, it's equally fast at all times as far as I can tell. It's not expensive, there are no hidden charges. I use my own router (though I'm also using the equipment that came from the provider, for the fiber-to-Ethernet part). All of this I could get in Japan years ago, we're finally catching up in Europe.
You wouldn’t know it by reading their periodicals: cancel culture, right vs wrong thinking, evil multinationals, class struggle against the exploiting "rich", the inhuman capitalism, the wonderful socialist utopias and it goes on and on...
This is quite rant-ish. There is a class struggle, there is a rich powerful few who can use their leverages to affect millions of people that have almost no voice. What is your point?
For someone complaining about divisiness you seem to be quite deep down into the rabbit hole.
The world is a large place. With 7.8 billion people in the world today there is bound to be 1 or 2 who go off the deep end. To say those represent the majority is insulting. It's like you're saying, "You wouldn’t know it about the US right wing by reading their periodicals: antisemitic, pro-fascist, anti humans rights, pro totalitarian, and more."
That wikipedia article is very outdated. For example, the two most common offers in Spain are symmetric 100mbps and 600mbps FTTH.
The most important player indebted itself with ~50k million € and it's basically replacing all the copper coverage with FTTH.
Other HFC players are turning into FTTH too, as HFC won't be able to compete with +600mbps FTTH offerings, and symmetry is more and more valued over time.
Also, I don't have the figures at hand, but the coverage in Spain is extraordinary. FTTH is reaching places that barely had 3mbps ADSL, and it's probably going to cover places that had nothing in the long run.
Public money is paying for capital costs in places that would be impossible to make a profit otherwise, but still the main obstacle are municipalities and nimbys.
I've seen some pretty stupid stuff, like blocking FTTH deployment in historical centers under aesthetic concerns, where they already have ugly copper wires in the facades. In reality what those municipalities expect is ISP to bury the cables and pay for the whole renovation of such streets, which no one in his right mind will. So this people will not have FTTH unless their municipalities change their mind, but they blame ISPs.
And then you have our case here in Sweden [1] with a thought out plan to roll out broadband to the whole country as a strategic need.
The infrastructure is public and leased to private companies, avoiding the tragedy that happens in the US where laying down cables is the most expensive part and which shuns away competition, leaving with oligopolies where each company got their turf and that increases the barrier of entry significantly.
That wikipedia article is somewhat wrong. There is a lot of competition in denmark.
There are 3 separate 4G networks with very high coverage selling mobile broadband rather cheaply (you can fx get 1 TB a month for approx 36 USD a month).
TDC was also forced to give competitors access to their PSTN, coax and fiber network so resellers can compete on selling on their physical network.
The power companies have also been aggressive in recent years rolling out fiber, so several places people can choose either infrastructure from TDC, their power company or one of the mobile providers.
In Denmark,the former state-owned, TDC has to provide equal access to their network for other competitors, meaning they can only charge the cost of maintaining the infrastructure, which has mostly been build through governmental funding
They have on multiple occasions tried to force competitors out of the market by making unfavorable pricing models or straight up blocking third parties from using their infrastructure
They are in no way a "benevolent former monopoly" and there's a lot of regulations in place to avoid them exploiting their market position, since they in the past have tried exactly that
Denmark has cheap, fast internet because of the regulations enforced on the market, which makes it possible for me to get a 1gbit symmetric connection from a small provider for 120 DKK (approximately $20) per month
I moved from Texas (Comcast) and while living in Sweden I assumed internet in the EU was much, much better compared to the US. Then I tried living in Germany for 3 years and it felt like I had Comcast again. I couldn't believe it takes a month for a technician to come to your apartment in Germany to flip a switch so your DSL works but that was reality there.
"Countries who jumped on the Internet earlier can't have fiber because of existing competition"
Fiber is actually MUCH, MUCH CHEAPER to deploy for higher speed networks than any high speed copper, because the later almost always implies to deploying a fibre network too to do everything other than the last mile.
1. You have to build your fibre backbone anyways. That's an even bigger trouble in countries with no culture to share fibre. And you would usually have to get much "brainier," and more expensive fibre equipment in this case.
2. You still need to wire the fibre fairly close to premises.
3. High speed copper network equipment costs huge sums, more than commoditised GPON.
4. Copper often needs to be torn, to be replaced with higher grade copper
5. You still need to drill buildings to install niches for copper equipment.
6. You need to put extra equipment into equipment closets.
7. Copper needs more equipment overall
8. Copper has lower reliability overall, and costs more to maintain
So, in the end, you just do the same FTTH, just with copper as the last few tenths of metres. This is so obvious, but people keep dogmatising over "it's already wired with copper."
GPON by comparison does not require anything in the closet, but a piece of glass passive optical slitter.
You can easily have 1024 end-users on a single powerful GePON OLT, covering few square kilometres of 100% passive net.
They are aware. That's why you get 20Mbit/s, maybe 50 if you're lucky. Cheaper in the short term and it's not like you have a choice.
AFAIU, GP is talking about FTTC (or VDSL), which is moderately faster than typical ADSL via existing phone exchanges or (far away) cabinets, while being much more expensive. FTTC can be cheaper than FTTH, but silly IMO - any increase in bandwidth will be expensive.
Curious why population is even a factor here. If it legitimately should be... why not compare states of the US with states of the EU? (Seems a roughly equivalent thing to do)
Bigger and more diverse countries have a wider variety of political issues and priorities. When was the last time broadband even registered as a political issue in the news, except in the context of some other more important political debate (urban/rural, racial disparity, etc.) And in the US, construction of infrastructure is mostly a municipal level issue, so its both low priority and varies dramatically by state.
It's best to say that both USA, and EU are very, very bureaucratised, and regulated. It's just one is marginally better at not introducing damaging regulations than the other.
Japan too had 100mb/s since nineties for example, and theeeen..... they stuck. Speeds there not been rising by much since because regulations froze in nineties, and so did the market.
Countries which had near zero telecommunication regulations covering the Internet, or had no Internet as such did advance more uniformly.
I can have a 500mb/s over GPON, or even a symmetric gigabit over 1000BASE-LX even in *stan countries, and Africa these days. Fiberlink has 1Gbs for $30 a month in Lahore. PCTL is starting a 10Gb/s GPON trial in Islamabad.
Single mode fibre based tech is much cheaper to deploy, and maintain than any legacy tech, and properly designed GPON net probably being the cheapest option ever.
I had (and still have, when I get back after Corona) fiber to the home in Japan. That fiber was installed some fifteen years ago, it was just a phone call away and then the fiber was hitched on the poles and into the house (everything goes on poles in Japan..). This is a nearly rural area by the way. Japan is not 'stuck' on 100Mbit - 1Gbit was only approximately 2-3 dollars more per month, compared to 100Mbit/s a few years ago.
>Japan too had 100mb/s since nineties for example, and theeeen..... they stuck. Speeds there not been rising by much since because regulations froze in nineties, and so did the market.
The UK internet is worse because of forced competition. British Telecom started plans in the 1970s to roll out fibre optic connections to every household in the country. BT were even going to manufacture the fibre themselves. But the Thatcher government decided that competition was a good thing so shut that down and instead invited cable companies to lay competing infrastructure. A small handful of companies accepted the offer and proceeded to lay infrastructure during the 90s to compete with BT. They installed infrastructure only in highly lucrative towns and inner cities where cable runs are short and easy but subscriber counts are high. No rural infrastructure whatsoever. None of them have laid any infrastructure since the 90s and by now they have all merged into one company which has been rebranded as "Virgin Media". So much for competition.
So what we have now is primarily a copper infrastructure with zero to two fibre backbones to choose from depending where you live. It's a shambles.
The big mistake was not recognising that the physical infrastructure is a natural monopoly and, like any monopoly, should be state owned. Just like sewerage, electricity, water etc. it makes absolutely no sense to have parallel infrastructure. The UK shows that while it is possible, you make great sacrifices in the name of "competition" and "free market".
No, not really. Europe is a little bigger. The countries are more like eastern states. [1] This is somewhat important, considering most of the regulation on utilities are at a state level - and the utilities kind of are as well. Some of the countries are somewhat populous - you know, like states - and some have a pretty sparse population. You know, like states.
Greedy companies are still part of it. I have exactly one choice for high-speed internet service in my building, and that's Comcast - despite FiOS being available right across the street. My strong suspicion is it's because of this: https://www.wired.com/2016/06/the-new-payola-deals-landlords...
(My building's apparently an "Xfinity Community".)
Non-EU Europe, same issue. Only one company has cables in my building.
On the bright side, they're the fastest ISP in the country. On the bad side, they don't offer just the Internet, so I have to pay for cable + Internet, even though I don't use cable TV at all (not plugged into my TV at all).
My point is that the rest of the world also has greedy companies. Yet, some parts of that world manage to build infrastructure just fine.
The US also used to be able to build infrastructure. People weren't any less greedy back then. Other things changed. Things that are very important to look into.
i blame the half-assed infrastructure. finally got a solid fiber line, so i'm good now - but prior it was nothing but hell. every tech out would just say, welp - i don't know why they sold you this, we don't support it...even tho they buried the dang lines. and fwiw i'm in the middle of a decently sized metropolitan area...not out in the sticks
Can you expand on when competition came to Seattle and how? I've been using Condo Internet (now Wave Broadband) for something close to 5-6 years. When was the tipping point in terms of competition?
Centurylink now offers fiber gigabit to the home for about $65 or so a month.
It happened because a former mayor (Ed Murray, who has since left office after a personal scandal) made it easier for ISPs to build out their networks and removed homeowners and neighborhood groups from the process which previously allowed them to block things like communications boxes on sidewalks.
At the time, a bunch of local columnists moaned about this "corporate giveaway"[1] -- but it's pretty clear in retrospect it was the right move.
Don't forget about Google Webpass, they started advertising heavily in Seattle area starting about a year ago.
I switched to them earlier last year due to having periodic issues with WaveG in my new apt building about 8 months into my lease (their technician would come a week or two after I report constant outages, fix something in the server room of my building, and then the issues would start again in a few weeks; haven't had any issues with WaveG at my previous apt though).
Service has been great, pricing is even cheaper than WaveG (I pay about $50/mo for their gig fiber offering), no contracts or any other lock-in (unless you pick the option to pay for the whole year upfront for a small discount). It definitely does feel like there is a good amount of real competition in Seattle now.
Yup! My apartment building was pre-wired for fiber and I can get WaveG, Century Link, or WebPass. I can also get Comcast if I wanted to torture myself.
That’s the most options I have ever had in my entire life for an ISP. We’re with WaveG and have no reason to switch but the fact that I have multiple fiber options if I need to makes me really happy.
In Brooklyn, I couldn’t get FiOS in my building even though the building across the street and next door could do it. Something with the zoning and no matter how hard we (and Verizon) tried, nada. The upshot was Cablevision was an exceptional ISP (let me run a server out of the apartment as part of the written TOS, which was incredible), but it was still slower than FiOS.
In Atlanta, my parents’ house was prewired for fiber and gigabit but they have been stuck with not even the fastest available cable because of ISP tie-ups. It’s painful and they have zero choice.
> The upshot was Cablevision was an exceptional ISP (let me run a server out of the apartment as part of the written TOS, which was incredible), but it was still slower than FiOS.
This is no longer the case since Altice purchased cable vision/optimum. They jacked up the prices and Andy reduced quality of service.
I apologize for my last comment. I didn’t address the character or nature of Ed Murray. I addressed what he did. What he did do. He did do what he was accused of. I’m glad to take the downvotes to vocally say that I believe his victims. The vast majority of sexual assault victims never see a jury, and even if we do we never stand a chance of proof in that court.
It’s not just a legal matter. I’m amazed that as a victim of sexual assault I’m repeatedly being lectured about the burden of proof of sexual assault. I suppose I can just assume y’all don’t believe me either unless I go to court to confront my assaulters and win?
CenturyLink fiber is relatively recent to the Seattle area (2017 or 2018) and serves a broader customer base than Wave née Condo.
Condo was historically only available in a small number of buildings very close to the Westin building downtown (site of the Seattle Internet Exchange). It is slightly broader now, but only slightly.
Here's a big old PDF of Seattle area internet service (2019)[1]. Most of Seattle is CenturyLink + Comcast. The second biggest portion is Comcast-only. Wave covers very little of the city.
Seems like I was hearing my Seattle friends brag about their Centurylink Fiber quite a long time ago -- I want to say 5 or 8 years. Maybe the initial rollout was small and it took awhile to get going?
I'm a Seattle native, down in the bay area now, and JUST got gigabit fiber. Couldn't be happier.
Maybe CenturyLink rolled out to a limited availability sooner than I'm remembering. There was also a very limited Verizon FTTH rollout on the east side some time earlier as well, but I don't know anyone who was in a service area for it.
The speedy providers are all jumping over each other to compete, now that the "gentlemen's agreement" between providers is no longer a thing... but they are cherry picking from people willing to pay and easily providing the service.
Seems like maybe 3 years ago when all the cable companies were able to compete without exclusivity by block?
In my condo building of 200+ units we have 5 providers now:
- Wave broadband cable
- Comcast cable
- Wave G (10Gbps connection to the building)
- Webpass Google Fiber (microwave to the building)
- CenturyLink (DSL maybe?)
The bummer is outside the high-rise towers, there is not much competition in the neighborhoods and single family homes it seems.
I'm going to say "Probably not". You're talking about people who have chosen to live in lower-density housing with lower internet speeds. I'm sure they'd love higher internet speeds, but not so much that they'd give up the 'lower density housing' part :)
Moved to Seattle earlier this year, was sad my internet went up $20/month. Currently paying $80/month for 1G/1G. I do agree though, having fiber to your place with single digit MS ping is quite lovely for all things internet.
If you call CL, you should be able to change the plan to $65/mo (one is pre-paid vs post?). I did that last year and saved $20/mo for exact same service. :-)
Seattle doesn't have competition, unfortunately. It's only in the very core of the city that has access to Centurylink fiber, mostly everywhere else is stuck with Comcast cable or laughably slow DSL.
Total capacity is clearly extremely important. Were it not the case nobody would be running new undersea cables to places that already have undersea cable connections.
Only in the super rural areas that either aren't wanted as serious competition areas anyway or smart enough to run their own local infra properly and lease out the video over IP stuff to a cable company to not deal with the licensing BS for content.
I fully believe Austin can thank Google Fiber for its speed increase. Not necessarily because that many people actually have Google Fiber - it's rollout has been patchy, delayed, and with Fiber rollouts in other cities halted, I'm not holding my breath to get Fiber in my neighborhood.
Averaging can give some really unrealistic numbers especially when people are upgrading their home internet to work from home. Upgrading a single connection from 100Mbps to 1Ggbit is hardly the same as upgrading 100 connections from 1Mbps to 10Mbps.
That is very true, not to mention gigabit cable services probably rarely have anyone achieve those speeds very often. My guess is that most cable companies unlock the DOCSIS channels necessary, but node over subscription means that your only shot is to download a huge amount at 3-4a.m.
In practice, in the US, the fiber connection will actually offer a sustained 100mbps up and down, with a low latency while the cable company will share a 500mbps down 10mbps up connection over 50 houses and sell you 100mbps by throttling everything else for a few minutes and using this burst speed as their advertised bandwidth.
>In practice, in the US, the fiber connection will actually offer a sustained 100mbps up and down
Does this actually matter given that up/down are already prominently advertised?
>with a low latency
Does fiber have lower latency than DSL/cable? Even if there is a speed difference between the two mediums, there probably isn't that big of a difference given the short distance involved (a few miles at most).
>while the cable company will share a 500mbps down 10mbps up connection over 50 houses and sell you 100mbps by throttling everything else for a few minutes and using this burst speed as their advertised bandwidth.
All bandwidth is shared at some point, but in practice, cable is far more oversubscribed. And they don’t advertise any upload bandwidth. And the download bandwidth they advertise are temporary burst speeds they get from degrading others’ speeds.
Cable companies have an incentive to continue allocating their bandwidth for cable tv channels, and offering the bare minimum (see lack of upload) to internet subscribers since there’s usually no competition. And I’m not an engineer or anything, so while I can’t say for sure if coaxial cable is inferior to fiber optic, the fact that every single fiber connection I’ve used is far snappier and higher bandwidth than coaxial cable leads me to believe it is true, at least for all practical matters.
Cable nodes work on a round robin frame structure, so it would make sense that fiber has less latency.
All bandwidth is going to be shared somewhere, but cable is often heavily over subscribed at the node level, which is going to matter more when you are suddenly given more DOCSIS channels but the node is left unchanged.
This was more about the difference in fiber gigabit and DOCSIS gigabit. If you pay for gigabit over cable, you might technically be able to receive gigabit internet, but realistically you share a node with too many people for its capacity to get that speed unless no one is using their internet.
I'm in the first ring suburbs of Minneapolis and going back over 10 years (and for several years prior to that inside Mpls proper) my Comcast connection routinely outperforms my rated speed all hours of the day. My rated speed has also 5x'd (50 to 250) at the same price in that time.
Thankfully, however, fiber _finally_ became available on my block this year.
I wish there was some attention paid to upload speeds too. A 100Mbps connection often only has 1-5Mbps upload which seriously handicaps it for work and hobbyists.
Hence I consider anything that’s not symmetric fiber to not be worthy of mentioning, because I know you’re getting on upload (and probably latency and packet loss too).
50+ is a rather large bin. Netflix recommends 25+ for streaming 4K, so we could assume 50+ is ~2 4K streams, which seems like a normal amount for a family.
There’s a big difference between a house getting 50down and 5up and a house with a symmetric gigabit fiber connection.
This seems more likely than Charter, ATT, et al. investing massive amounts of cash into their networks. I think there would be massive PR campaign if the actually almost doubled available bandwidth in a year. The capacity was always there, they just needed to be paid in order to open it up. I upgraded from 200/10 to 600/35, they just flipped a switch.
Yep. Service providers didn't ramp up their investments at all into telecom equipment providers during covid, which is impressive. They just had good capacity that was never fully used. They're going slowly on the whole 5G thing now as they don't expect customers to be ready for massive price hikes to their plans. Ultimately the capacity of customers (individuals and businesses) to pay is what drive those investments, regardless of how much everyone claim to need ever increasing bandwidths.
It is both. This number is probably driven by people that are upgrading plans. But the telcom companies are spending billions on expanding infrastructure. Look at the capex numbers for any of the big companies. A lot of it is 5G, but wired as well.
No idea. As with all truth, you should check with multiple sources for corroboration first. I know in Hawaii at least there was a massive DCCA (chamber of commerce) initiative started way back in 2012 to outfit all public schools with gigabit fiber and all universities with 10gigabit fiber. Hospitals also got fiber connections though I don't know how fast, and there were a few 10gigabit undersea cables laid between the islands themselves. The state government started offering incentives and subsidies to encourage more cable landings on Oahu. This was also the same program in which they began installing the FirstNet system and the 5G infrastructure. No idea what's going on on the mainland.
Also Zoom classrooms for kids/students. Work and school from home is probably the catalyst for higher inter speeds as employees/parents need faster internet speeds for Zoom. It would also correlate with the rise of Zoom stock price.
I get 300 Mbps down from Xfinity, fine by me. But the paltry 5-6 Mbps I get up with it is BS. Sure, I don’t need 300 up, but for what I’m paying it would be nice to get at least 50.
Sounds kind of like a network designed and configured for selling cable television rather than internet access, right? ;-/
I believe most cable channels are reserved for broadcast TV, with just a few for internet data, and most of those for downstream rather than upstream, since DOCSIS 3.1 is half duplex.
Things will hopefully get much better with DOCSIS 4 though which supports full duplex! Comcast claims that they are going to roll it out DOCSIS 4 and FDX (as it is apparently called) next year, so who knows? You might have 1 Gb/s upload or better.
And it will make it a lot easier to quickly reach your monthly data cap.
And the residential plan TOS still forbids "servers". Specifically:
> use or run dedicated, stand-alone equipment or servers from the Premises that provide network content or any other services to anyone outside of your Premises local area network (“Premises LAN”), also commonly referred to as public services or servers. Examples of prohibited equipment and servers include, but are not limited to, email, web hosting, file sharing, and proxy services and servers;
Funny thing is this seemingly contradictory clause has been added at some point:
> use or run programs from the Premises that provide network content or any other services to anyone outside of your Premises LAN, except for personal and non-commercial residential use;
So "dedicated, stand-alone equipment" is prohibited, but if it's just a "program" you can use it for personal non-commercial purposes. What does that mean for a program running on a RaspberryPi?
Funny, I live in a rural area with a small WISP. I get a solid 20dl/10ul all the time. Hard to imagine that I'm uploading more quickly than city folks.
I, too, live in a rural area, but the local WISP cannot provider service due to the geography of the area, so I have 12Mbps/768kbps via DSL, while a friend up the road is getting a symmetric 80Mbps connection from the WISP.
15 or 20 mbps up is about the theoretical maximum on cable because of how the spectrum is allocated by cable companies. Cable TVs didn't need much upload compared to download.
While no DOCSIS revision has symmetric theoretical caps, DOCSIS 3.1 which is currently in use is capable of up to 2Gbps upstream, while the upcoming (for the most part) DOCSIS 4.0 standard is capable of up to 6Gbps upstream, and was developed largely to response to increasing demands for upstream bandwidth. While these are obviously best-case speeds, gigabit symmetric is quite achievable with DOCSIS3.1 and available in some markets.
However, it is unfortunately true that many cable carriers are constrained in their channel planning by legacy devices with fixed channel requirements. For example, Comcast has at least historically struggled with planning DOCSIS channels around legacy Motorola STBs that require their on-screen guide information on a specific channel. Essentially, in order for Comcast to switch to full DOCSIS3.1 upstream channel allocations they will need to first get every customer with a legacy STB to swap it out for a new one. This is a formidable task and not one they have so far completed, so in my area upstream remains limited to 35Mbps even with 1Gbps downstream.
No, that’s just cable companies allocating significantly more to download spectrum than upload. There isn’t anything fundamental in docsis preventing you from giving more spectrum to upload to have way less contention on the upload slots.
Nah, there are cable companies that offer 50 mbps upstream, and the DOCSIS 3.1 standard that's currently in use supports much more. They're just stingy.
There's absolutely no way that can reflect reality. I don't even think it's possible for people to downgrade their speeds that much even if they wanted to.
Anecdotal, but I just checked and the slowest speed I can get for my Stockholm apartment is 250/250, and the slowest speed my dad can get for his apartment is 100/100.
This has got to be some data anomaly. If they're basing their numbers off of tests that people voluntarily do, then maybe in 2019 a lot of people tested their mobile internet, while no-one bothered testing their home internet because it's fast enough?
I've been looking at that graph several times and I have no idea how they got those numbers for those European cities.. it's like it's from years ago. The numbers are way too low.
The title makes it seem like broadband infrastructure in the country significantly improved over the last year, but I'm guessing people just paid more for faster speeds at home.
Something interesting I recently found out is that New Zealand has just started offering 4gb/4gb connections.
There's always a lot of discussion around cable companies and monopolies etc whenever internet is brought up so I'll explain how NZ has done so well,
In the 2010s the govt decided they we're going to 'fibre to the home' most of the country.
They did so by public private partnerships allocated by regions. Most of the regions went to one company but that hasn't been too much of a problem. The 2nd largest city went to a competitor etc.
These companies now own part of the fibre network but don't necessarily sell to consumers. ISP's buy time off the fibre due to something that was done in the early 2000s with the advent of DSL called "Unbundling the local loop".
This means anyone can setup an ISP and sell to you using the "governments" fibre. This has resulted in plenty of competition and meant that NZ has gone from lagging behind many countries to now far exceeding.
Datacaps were the bane of many 2000s internet consumers 25gb for $x with 256kbs download etc. when USA connections were unlimited 256kbs for half the cost.
Another interesting thing that's happened with the rollout of fibre across the land, business connections are now very similar in price to consumer connections. You can also now get business connections to your home and without any additional infrastructure.
Just a few years ago a business connection might cost you 2000NZD a month and require an expensive install. Now that the only difference is a SLA, you're looking around 150NZD in comparison to a 100NZD "home" connection.
Something I think NZ has done well and could be a good model for other countries.
There needs to be a bipartisan political will for it which is hard to capture.
Australia tried to follow the same model only for a more conservative government to get in, claim that it was too expensive, butcher the plans (fibre to the node instead of to the premise) to "save costs" the country was left with a more expensive half-baked product that still incorporated a lot of copper.
What's worse is that because there was a "government fibre" model it meant that private industry that wanted to lay its own fibre started facing various endless legal issues[0], which, funny enough, is entirely supported by the originally less conservative political groups that pushed for the rollout in the first place. So now we're left with a bad network and legal structures in place to stop it from being privately superseded.
Just to bring some data to the conversation, the price for 1Gps symmetric fiber is $78/mo here in the rural parts of Denmark. That also includes a bunch of other incentives such as streaming services, mesh WiFi hardware that is kept up to date as new technologies arrive.
https://www.waoo.dk/fibernet/
We had a large push from the power companies back in the early 2000’s bringing fiber to homes that was outside the metropolitan areas. This has kept competition going across the country between cable, xDSL, 4-5G and fiber.
Seems higher than average to me. US$45 is what I pay for 1/1Gbps in Denmark. You could likely switch and get better service on the same network for less.
I hate that across the highway near by (Houston) is ATT Fiber with 1G up and 1G down for around a 100$/month. I moved across the highway, and now stuck with Xfinity coaxial 1GB down and around 50MB up if I am lucky for 150$/month including no data cap fee.
I don't understand why a company like Xfinity with such negative public image wouldn't try to give everyone a speed bump to improve their image.
While it's great that the average speed is in the mid-30s, it's still just the average. If it was closer to 100 I'd feel more elated by the news, since that means the lower ends would be acceptable no matter where in the USA a user is.
Also notable is that a lot of countries in the chart they showed had a 50% or more increase over 2019.
Good point. I imagine the average is going up because the already fast speeds in cities are just getting faster with fiber and faster cable connections.
What I've concluded from measuring things on my home network is that my cheapo old router is actually only giving maybe a third of the download speed my ISP is providing me.
I'd guess that is the case for a non-trivial subset of the population.
And I mean I get ~15Mbps to my laptop when my connection gives me 50Mpbs, nowhere near something like gigabit which I'd guess is just an excuse for ISPs to fleece money out of people in all but a small minority of cases where people buy it.
But also I have zero incentive to upgrade my router because the internet is "fast enough" for everything I need, even with only using a small fraction of the available bandwidth.
Edit: Also to the headline of the article, my ISP doubled the base internet rate from 25Mpbs to 50Mpbs this year. So I guess the report lines up with my reality quite well.
Wow, I can't imagine being in your shoes at all. I would immediately go to the store and get a cheap semi-modern router for $50 to get my full ISP speeds, especially when missing 2/3 of it. Even a 10 year old 802.11n AP ought to be able to get you 50mbps at reasonable range.
I probably should get a new router. It's sort of become low priority.
The main signal issue is that our router is in the garage next to where our internet comes in.
For a few weeks in the spring I was looking at setting up a wired network since all the phone jacks in our house are actually Cat5e cable that could be upgraded to ethernet jacks pretty easily.
But then I also lost ambition for that idea.
I did spend $20 a couple months ago to get a latest-generation wifi chip for my laptop, which I'd recommend to anyone since it's a trivial thing to upgrade and I never even realized the wifi chips were swapable. That improved my network speed by a factor of two I think.
For most non-techies, buying a router means hours of fiddling with an inscrutable setup ending in failure, then either giving up, or if you can afford it paying the Geek Squad to set things up for you.
The advice I give most people is to use whatever the ISP provides, so that when things go south the ISP is on the hook for support.
+1 here. Routers are absurdly more complicated, hard to fix, and full of brand new terminology to deal with than basically any other piece of normal house tech. Most should absolutely stick with their ISP's, as crappy as most of those devices are.
It may also be the cable modem. It's not uncommon at all for folks who own their own modems to have something that's so old spec that when they upgrade their service they don't get the speeds they are paying for.
I had this happen... Bumped to 500/50 with Wide Open West and immediately went "...oh" then grabbed a new modem.
I'm also completely satisfied by my mediocre internet speed. Video streaming, conferencing works. I don't torrent anymore. I just don't know what I need more bandwidth for. Curious to hear from those that do use the data bandwidth provided at 50mbps.
My measured speed at home is typically over 600 Mbs. Most (not all) of the house and devices are networked over ethernet. I'm super happy that we have all the bandswidth because on most days:
- We have 2 adults working full time from home. We're in the entertainment/technology industry. We're streaming video calls and transferring massive files/repos _constantly_. One uses a remote desktop solution to get to their media workstation in the office regularly, requiring a low latency connection.
- We have 2 children going to school from home (schools haven't reopened here), spending the day on either Zoom or on Netflix/video games.
And then there's all the smartphones and other smart devices around the house regularly eating up bandwidth.
I don't always need all that bandwidth, but when I do, it's crucially important.
I often get better than 50Mbps, depending on the location in the house it can be 200. It's not about "What can you do", but "what don't you have to wait for". If I'm downloading a 100MB file, I don't look away or browse somewhere else - The 10 second it takes to download isn't enough to interrupt my concentration, while if you're on a "mere" 50 megabit connection you're waiting 40 seconds and your mind has wandered and you've walked away.
Snappy is important. The more that's snappy, the better you're going to feel.
I have 100mbps, and I sometimes wish it were faster. Mostly for downloading video games, honestly. Non-indie games these days are often in the tens of gigabytes, so, even on 100mbps they can take a while to download. And patches can be big, too. And then if it's both me and my son downloading a game to play together...
I suppose work video calls with a dozen+ people in them probably take up a decent chunk of bandwidth as well, but I haven't checked closely.
According to Speedtest the average US speed is 165 down 62 up, or to put it another way it's enough for the average household to stream SIX simultaneous 4K-quality streams while also uploading photos and video chatting with friends.
So I mean I went here https://www.speedtest.net/global-index ... and it is basically average broadband speed... I mean... that excludes pretty much everyone without access to broadband. And doesn't say anything about caps, which, let's face it, reduces your monthly throughput to an instantaneous average well well below average, and approaching dialup speeds. All of these issues... access, caps, average vs. median vs. minimum vs. maximum speeds... are all things that would be handled by regulation.
> American internet users have had a very good 2020: according to research performed by FairInternetReport, median US internet speeds in 2020 doubled to 33.16mbps, up from 17.34mbps in 2019. Covering the five years of 2016, 2017, 2018, 2019, and 2020, this is the largest speed increase seen in the US, with speeds staying essentially the same in 2016 and 2017 (8.91mbps and 9.08mbps respectively), and 2018 recording a median speed of 12.83mbps.
It just seems faster when measured in millibits per second.
I know multiple founders who have 1gbit/s fiber lines to their homes in Berlin. Plus all 3 major providers (1&1, Telekom, O2) offer up to 250mbit/s there.
Also, you expect me to believe that the internet speeds in France and Germany magically went down from 2017 to 2018 and 2019?
BTW, my guess would be that most people stay at around 30mbit because it's good enough for 4K YouTube streaming and still affordable.
Recently my upload speed dropped from 50mbps to 10mbps. Trying to get the ISP to address the issue is like taking to a brick wall. The tech they sent out couldn't find anything wrong, but hinted that it was configured that way and there was nothing I could do to get it changed since the ISP wouldn't guarantee the upstream speed. We really need a functioning FCC to get this now critical infrastructure fixed.
I wonder how much of this is an artifact of people increasing their bandwidth subscriptions vs. the investment of various DOCSIS operators in Node+1 architecture, which at least in this area rapidly accelerated over the last year - I don't think that has anything to do with COVID per se but just the incumbent ISP really getting to the main stages of that project right now.
I recently switched from Comcast to AT&T Fiber, and yes, my speeds are up according to speed tests. But on certain routes I only get 30Kbps (sometimes to Azure!). Wikipedia and Twitter only loads half the time. Sites that don't use TLS sometimes load a gibberish.
So yes, the speed is faster, but I kinda want to switch back to Comcast. And don't get me started about the idiotic, mandatory $10/month shitty router they require you to use.
For a big chunk of the stay-at-home order, didn’t a bunch of providers remove caps/limits? Seems like that alone would cause a huge increase in home speeds.
In South Africa most fibre providers doubled the speed, aka 50/50 became 100/100 for 3 months without increasing the price. It has since been reverted.
I am surprised France is so low. They are rolling out fibre to the premise with gigabit connections in every big city and even in many smaller cities and villages.
It's probably down to the dataset. I went looking for their source and it's not anything I've ever heard about. Most Europeans would probably be more familiar with Speedtest.net [1], which seems like a closer match to reality (with France having descent broadband coverage and trash mobile).
I wonder if their method of excluding cellular connections treated tethered computers as cellular. When I saw the headline my first thought was that speed tests run over LTE could account for some of the difference (but not reflect people's primary internet connections).
> Speed test data for cellular connections was excluded from our analysis by using MaxMind's Connection Type database
I wish more attention was paid to upload speed and latency to the rest of the internet too.
People are backing up more personal videos than ever (including cloud connected security cameras, etc.). Latency is also very important to lower actual transfer times (since it will allow TCP to take full advantage of the bandwidth available). Latency also will be helpful for new experiences like Stadia and Geforce Now.
The latency is not a problem as long as you avoid packet losses. If you have losses then you pay dearly if there's latency. The latency itself doesn't present a problem, the TCP window size increases and takes care of that. The problem is packet losses, because (even with 'intelligent' retransmission) there's a lot of (the buffered) window to retransmit and then the performance dies. When the packet losses stay low, say, 0.001%, the latency is not an issue.
(At work I did months of testing on TCP performance for a 900ms rtt high-speed link)
But yes, upload speed is more important than ever. For my work-from-home setup I need to transfer lots of data both ways. And for my "cloud" backup etc. I need to push large amounts of data out of the house. And with fiber I get that, fortunately.
The power of WFM and Zoom classrooms. If there is one thing people will pay extra for, it's their employment and especially their kids' schooling. When a school tells a parent that they need high speed internet for their kids's Zoom classroom, they'll pay for it.
Anyone older people with kids know if schools subsidized parents for high internet connections?
2) Those numbers for EU are incorrect. Look at Stockholm for one. It is at a minimum 2x too low. Even Denmark which is in top is too low. I live in Denmark and even if I pick cable the lowest I can even order is 300Mbit/s. But that would just be stupidity since I can just click on Fiber and get 1/1Gbit/s for US$45.
Gigabit level speed for commercial users wasn’t even imaginable until Google came along, and all of sudden, everybody and their mother seemed to be offering it as if it had somehow always been a part of their rollout plans.
I should move back to Switzerland. Shortly before I left in 2016, we had a small ISP, init7 providing symmetrical 1Gbps connections for 70 USD.
In 2020 they get 10Gbit for 50 USD...
Meanwhile I pay 60USD for a 50Mbps asymmetrical connection in Bangkok.
It’s interesting that larger countries have lower median speeds. It suggests that ISP’s should not be allowed to serve more than about a million people.
I wonder what other industries have inverse scaling properties like this.
I wonder how long that'll keep up? I already have gigabit fiber, and by that graph he will too. But then what? Faster is better, but 10Gb wouldn't be much of an improvement in practice over what I have today. Game update servers don't saturate my bandwidth, and we can have multiple people streaming 4K video at once without blinking an eye. What would be the next thing pushing faster speeds?
US cellular services have absolutely miserable data caps compared to wired ISPs (yes, even on "unlimited" plans). The increased speeds of 5G are basically irrelevant if you get throttled after downloading 50 GB in a month. This is a problem that cell providers could solve pretty easily -- but they've had that option for years, and it seems like they'd rather not.
Starlink is density-limited. There's a finite amount of satellite capacity available to any amount of ground, and they can't target upgrades to specific geographic regions. They're a great upgrade from existing satellite Internet services, but they aren't going to replace terrestrial services for most people who have that option.
There's no reason why high density cities in the US could not have the same infrastructure as "South Korea"—which is also made of high and low density areas like America is.
This idea that some countries have better Internet infrastructure because they are "high density" is silly because the places where people live in the US also have the density to support the exact same infrastructure.
The farm land / country side of South Korea is just as low density as the rural areas of the US, and America's cities and suburbs are just as dense as South Korea's.
The U.S. actually has similar weighted population density to most of Europe. Simply dividing the total population by total area is a very poor metric, especially for a country like the U.S. which just has an enormous amount of unpopulated land area.
An example of a weighted population density metric is the local density of the median household. Unfortunately, weighted metrics are rather new and last I looked I couldn't find population numbers using the same metric. The U.S. Census Bureau uses a metric based on census tract as the local area; EU uses square kilometer. For that and other reasons the published data isn't directly comparable (at least not easily). But to a first approximation the U.S. and Europe seem very similar--the U.S. is far more dense than commonly believed, and Europe less dense than commonly believed.
Of course, even with a better metric I presume South Korea is in another league entirely.
And how much of Sweden's population is concentrated in the southern quarter of the country? 80%? Dumb average density doesn't reflect the reality of population's distribution.
And how much of the US’s population is concentrated in and around cities?
Most people in the US doesn’t have fiber internet because the rent seeking monopolies have no reason to make capital investments to provide better internet, not because of the population density.
If anything, you’re more likely to find fiber in more suburban less dense areas with new developments. They will spend to put fiber down on new construction, but for existing urban areas, no one wants run fiber through existing buildings and pavement.
This is the sad part. I live in a suburb of Houston, and while it’s certainly nothing like NY, it’s still near a major city. Comcast is the only game in my apt building.
Who cares? With everything capped at 1 or 1.2TB of transfer per month, there's no point to having internet any faster.
For anybody who needs the speed, they are transferring large amounts of data. If you can't transfer large amounts of data, what good does the speed do?
The other big problem is, a huge part of the country still has basically no internet. Rural communities stuck with 1.5MBps DSL or even slower can't even have 2 kids in school learning virtually.
Imagine if you had 1 remote parent working from home, and 2 kids trying to do school. Most rural internet just won't cut it. I feel terrible for people having to deal with our shit connectivity in this country.
> Who cares? With everything capped at 1 or 1.2TB of transfer per month, there's no point to having internet any faster.
What? I definitely care how fast my internet is regardless of my bandwidth cap. For most people, internet consistency and speed is a more noticeable metric than data caps. Speed effects video/audio call quality, page load times, etc.
Look, i get annoyed at data caps. I've taken multiple comcast employees to task over it. But the idea that speed is irrelevant when caps are present strikes me as pearl clutching.
I have mobile data connection where transfer fluctuates during the day.
In the morning 15Mbit during the day 5Mbit and in the evening might go as low as 2Mbit.
It mostly works fine but at some point of the day it might just stop working, like 2x a day.
I have 50Gig data cap, I am not annoyed by it because I can control it.
I might watch yt on lower resolution or skip netflix binge in the weekend, because I need transfer for work.
What I cannot control is transfer speed and glitches. It totally depends on what other people in mobile network are doing. If they all start streaming at the same time or having meetings, bad luck, have to wait it out. It is not making me angry because I understand it, but it is uncomfortable. Where data cap does not bother me that much.
> Who cares? With everything capped at 1 or 1.2TB of transfer per month, there's no point to having internet any faster.
My ISP doesn't cap at 1-1.2 TB (CenturyLink, Seattle metro area, "gigabit" service plan). Comcast is often a monopoly, but it isn't the only ISP in the US. Your point is certainly valid for Comcast customers, and in other areas where all ISPs cap at some relatively small amount of 'maxBW x hours'.
But to reiterate what has been stated every where else over and over: for most regions in the US, there are at most 1 provider with usable speeds available, and they get to do whatever they want with their customers.
There are many regions with hardly any competition, and the lucky few areas with 2 or more usually comprise of another high speed provider that changes in lock step with the local #1, and a bunch of smaller providers on DSL or some other old technology.
Wireless is even more capped, high latency, and, thanks to the FCC, can be more privacy invasive or restricted than land line internet.
I live in Seattle and we had a Comcast monopoly for something like 15 years until 2017 or 2018. I do not mean to suggest that real competition exists in most places or that anything about the situation in general is rosy.
I really wonder how Starlink is going to shake things up. It’s currently expensive, but a continuously improving service could represent a huge risk to otherwise stagnant ISP’s.
Starlink has a constant density they can't vary, so likely they'll have lots of unused capacity in the rural areas while having lots of used capacity in the urban areas. Thus, they'll likely want to increase the price in urban areas to reduce the number of customers while keeping it cheap in rural areas. Otherwise they have to come up with some other system that keeps customer usage low.
Maybe they'll offer "mobile internet" at a price point somewhere between rural and urban pricing that you can e.g. use in your car. Maybe folks will then hack their parked car and use it as a home uplink, while driving around to keep the "mobile" pricing level.
As for the price in the urban area where density is too big for them to serve everyone, likely it'll be modelled after the price the incumbent ISP has, and be a bit cheaper, but not too cheap as they can't serve everyone in that area. TLDR: nothing much will change.
So in urban areas they likely won't cause any change, but in lightly populated places there will be a big revolution, and people there will be very happy :). At least if Starlink choses to make internet cheap. Maybe they don't want to, and instead want to orient themselves after the local ISP, being significantly cheaper so that almost everyone switches to them, but not being as cheap as they can be. Entirely possible, brings them more money, but is obviously not very nice for customers :).
We'll see a real revolution when we have multiple competing companies in space that try to outbid each other in a fight for market share.
From what I have read they can deploy nearly arbitrarily amounts of satellites with nearly linear cost increases. They can’t really adjust density for specific areas very well, but total bandwidth is highly variable over time. Publicly they have stated 10% of local Internet traffic, in high-density cities as their benchmark, that’s likely an economic not technical limitation based on the percentage of time a satellite spends over dense cities vs rural areas etc.
Anyway, it’s really not a question of total bandwidth so much as any meaningful competition stopping the most egregious abuses. Assuming they can reach even 1/2 that it’s still plenty to provide stiff competition in midsized cities, suburbs, and rural areas which are generally the most problematic areas.
You can put a lot of satellites into the sky, but spectrum puts an upper limit onto bandwidth. Eventually you'll probably adapt directed motorized antennae, but those are way more complicated. Internet traffic is increasing dramatically around the globe, sooner or later we'll hit the limit. In the long term, only fiber can serve the bandwidth needs.
At most their aimed to maximize the amount of viable sky not individual satellites. It’s not going to scale to infinite satellites, but 10x their current target should be viable.
Yep, we got WebPass, WaveG, CenturyLink, and Comcast/Xfinity in Seattle area, and only Comcast/Xfinity has any sort of caps. I personally tested both WaveG and Webpass, and neither of them have any caps, just like advertised (been watching Netflix and downloading/uploading tons of stuff, on the order of over 1TB a day, no slowdowns/hidden caps either)
Thanks. Edited my original comment to clarify that the cap/no-cap status of CenturyLink may be local to my metro, or related to my particular internet plan.
> For anybody who needs the speed, they are transferring large amounts of data
Uh, no? Regular web browsing, email, SAAS, video calls, streaming, gaming – all greatly benefit from faster speeds. Data caps are terrible, sure, but saying better internet is otherwise pointless is stupid.
> Who cares? With everything capped at 1 or 1.2TB of transfer per month, there's no point to having internet any faster.
Yeah, that's ridiculous. At least some well-known providers allow you to add an "unlimited" option to your internet plan.
> The other big problem is, a huge part of the country still has basically no internet.
That's a difficult problem to solve. There has to be incentives for corporations to bother laying infrastructure. When municipalities try to pitch in and do it themselves, they get shot down by the same players.
I guess the only viable option near term is Starlink.
I worked for a small time WISP for a while, partially servicing a town of ~10,000 (but mostly not because of interference) and large swathes of the rural surrounding areas. He was heavily harrassed by police for various nitpick infractions regarding the equipment on his truck and where things were placed and whatnot, things that the local power company and AT&T blatantly violated with regularity without repercussion. He was also met with resistance from the city board/mayor, who were trying to require some regulatory fees that they (come to find out) didn't even have on the books. I think the only reason his business survived this stupid pushback is because he is former law enforcement.
I care; caps do suck but given the choice I'd rather have double the speed at half the transfer cap. To be clear, by double the speed I mean half the latency between requesting a resource and page load, which is a significant amount of my usage pattern.
If something (like uploading a backup or streaming a movie) is going to run for a long time, I don't care how fast the average speed is as long as it's enough to stream 1080p.
But if my next task demands some info from a 20MB datasheet that I don't have locally, I want that to load as fast as possible. And I'm not going to do that 25,000 times a month.
> For anybody who needs the speed, they are transferring large amounts of data. If you can't transfer large amounts of data, what good does the speed do?
It saves a lot of time. Having a new update/game or sharing large files in minutes instead of hours matters. Why not just start download earlier? Foresight on a long timescale takes a lot of brainpower and is pretty stressful because cost of miss is high. I will start this download and go get water is a whole lot better than I will queue this update and and it will be done in the morning.
Also there's not much foresight you can apply when an online game gets a required updated 5 min before you want to play it. And these days it's only going to get more common.
4K movie seeking on consumer devices can be dramatically faster when you increase from 50mbit/s to 400mbit/s, even though you'll still download just as much movie as you would on either speed. We're talking the difference between 5-10 seconds of buffering and beginning playback immediately. Video game updates also accelerate dramatically, allowing more efficient use of free time.
Incidentally, here's anecdotal data from a family of two adults and two children from today's Comcast thread (where I assume you're participating, given the comment), that shows that they have stayed below the 1.2TB Comcast threshold for over a year, even through the pandemic, while not making any effort to stay under that threshold since they aren't on Comcast: https://news.ycombinator.com/item?id=25200882
Just because there are data caps on some providers is no reason to tolerate unnecessarily slow internet connections. The speed of the connection is an effective and measurable stand-in for the modernity and capacity of the infrastructure delivering internet to that area.
Sadly, latency isn't a popular measure for Internet connectivity, and as my year of testing uncovered, 50mbit down and 5mbit up is more than sufficient for downloading + 4k streaming + zoom all at the same time, as long as your router has smart shaping/queueing capabilities and your connection has stable latency in that scenario.
Agreed, average speed is generally a pointless metric with it comes to actual usage for many people. We are one of those rural families and we have to have both a 3Mbit DSL connection and a 4Mbit Satellite connection just to make sure that at least one service works well enough so we can work. We can't wait for Starlink to be available and will happily pay nearly any amount to have it if it means that I can kill 2 different really bad services.
Starlink will cost "over $10B" to build and will cover essentially the whole world. The FCC is getting ready to award $16B in rural broadband subsidies just for 10M people.
When web pages are 10s of MBs today, faster Internet speeds are absolutely not all about being able to transfer more than a terabyte in a month.
Also, the US having “shit Internet” is contradicted by TFA. It shows we have faster Internet than the big European countries. Only a handful of small, rich European countries—which are the size of states like Maryland—are faster.
> The other big problem is, a huge part of the country still has basically no internet.
Infrastructure isn't free and while it's hilariously embarrassing to have slow speeds in "rural cities" (i.e. actual cities that just don't happen to be on the coast) simply due to capitalism doing it's worst - living in the middle of no where is an expense on society.
It falls into the same camp as people who live in hurricane areas - the government bailing you out once so you can move is a fine idea (and the government needs to be providing sufficient funds for these people to actually relocate) but living in remote areas and living coastally in the south east come with lots of benefits to you... and costs for everyone else.
I am A-OK with folks who insist on living in the middle of nowhere paying for it.
Oh I'm going to add a big old caveat for people who were forcefully relocated - if you're living on Rosebud Indian Reservation then relocation really isn't an option.
You do realize that roughly 100% of your food comes from those rural areas that you so casually throw away as an expense on society?
Add to that, telecom companies have been collecting additional "fees" from every user for decades that was supposed to go for expanding infrastructure and never has been.
Seattle has some actual competition now, and many (most?) of us can get fiber gigabit to the home for a reasonable rate. It's glorious.
I hope other American cities start courting meaningful ISP competition in their cities -- but I'm not holding my breath. Many American city governments see ISPs as companies they can shake down for fees and concessions, not realizing they're only hurting their own citizens by limiting choice and increasing costs.