Hacker News new | past | comments | ask | show | jobs | submit login
Microsoft says its data shows FCC reports overstate broadband adoption (techcrunch.com)
341 points by sethbannon 15 days ago | hide | past | web | favorite | 102 comments

The FCC says that I have seven possible ISPs at my home. Never mind that three of those are just different levels of service from AT&T. And one of those is Sonic that just resells AT&T for more money at lower speeds (not their fault, they have to ride on AT&Ts lines because AT&T won't let Sonic into their cabinets).

And two of those are wireless broadband providers who don't actually service my house (although they do service within a mile of me). So really I have two choices -- AT&T and Comcast. And I'm one of the lucky ones that has two choices. And even luckier that one of those is Gigabit.

So even for my house, with Gigabit service, the FCC report is still grossly overstated. I can't imagine how bad their data in rural areas is.

Same for me in Manhattan: the FCC reports that I have 6 options, including Gigabit fiber.

In reality, I have two: Charter (a/k/a Spectrum, a/k/a TWC) at 200/20, and Verizon ADSL at 5/1. Confusingly, the 5/1 hookup is actually more expensive than 200/20.

And that ADSL doesn't even count as broadband anymore!

> So even for my house, with Gigabit service, the FCC report is still grossly overstated. I can't imagine how bad their data in rural areas is.

Is that corruption, incompetence or apathy?

I'm willing to bet that their reports are entirely based on annual reports submitted by the telecom industry. The FCC probably doesn't have the manpower to do the surveying on it's own (or the funding to create such a task force), so they just ask ISPs to tell them what services are offered and where on an annual basis.

Unfortunately, they probably also lacks the funding to provide an auditor for those reports, and I bet there's no regulation requiring the industry to have an FCC auditor for them.

In other words, Congress doesn't care how accurate the report actually is.

In America corruption is not labeled "corruption" . It is labeled "lobbyism" . It is a legal mean to purchase your politicians. Now it happens to be the telecom industry that historically has been the biggest spenders across industries until Google came along. If you want to fix the sad state of the American political system this is very you should start. Make all contributions over, let say, 100 usd illegal and all contributions should go on public record. I believe it really would be that simple.

Contributions are on public record


All contributions already are on the public record. Voters just don't take the time or interest to study it. Limiting the top amount is already in place in a number of states, and results in complicated legal bundling schemes where people host fundraisers where everyone in attendance signs a contribution card and is basically given the cash to turn back over signed under their name.

There isn't an easy answer.

Pretty sure the answer to all of those is yes.

I would go with corruption. I don’t see how it’s in anyone’s personal interest who works at the FCC to have reduced access to internet, and it’s super convenient for everyone, so there must be some side benefit to go out of their way to juice the numbers.

Regulatory capture. The FCC is the agency responsible for oversight of ISPs. They spend less in buying the politicians then implementing regulations.

ATT took a formal complaint to the FCC to get DSL hooked up...DSL. Charter is shown in our rural area. I live 20 yards from a two lane paved "Highway"...literally it's name...in rural Georgia (~45 miles as the crow flies from downtown Atlanta). Comcast actually has the rights. A Verizon hotspot (excellent 4G speeds) off my phone is actually the fastest, most reliable download speed we have. I have a large state university 15 minutes from me.

...this is the state of broadband for semi-rural America.

[EDIT] Oh and the ATT easement stops next to my driveway. There is an old T1? (as thick as my wrist) of copper entirely unterminated next to my driveway as well.

Well as an additional datapoint, it lists my two broadband ISPs (AT&T Fiber, Charter Cable) correctly here in suburbia Cary, NC. It also lists three satellite providers which I have no idea if that’s correct or not.

Pretty much anyone with a view of the sky can use the satellite providers. But they are terribly limited, expensive, and laggy.

These look ups mostly use Zip. A better test is to call and ask if you can get said Premium service at your street address and then see report back.

Same in Daly City. It's a complete mindfuck that I live immediately south of what claims to be the technology capital of the world and yet my only two viable options are "pay through the nose for Comcast" or "pay for AT&T DSL or some reseller thereof" (even the wireless broadband providers don't serve Daly City, or only the northernmost parts of it if they do; nothing around Colma where I'm at).

I ended up going with Sonic.

I get excited every time I hear the name Sonic now. A month ago they announced expanding gigabit fiber into my neighborhood. I'm all pre-ordered and ready to go. Saw AT&T running the lines a few weeks back. Refreshing the ATT service checker to see when it goes live. Multiple times a day. Looks like, at least for Gigabit FTTH, Sonic is vastly cheaper, even riding on AT&T's lines, and no caps.

In fcc's defense that's how govt counting works formally. Whether or not they are doing anything informally is a different issue.

Sonic is more expensive for just internet? Or are you bundling with TV etc.?

Yeah just bare internet is more with Sonic. I love Sonic (had them in the reddit office in SF) so I talked with the rep for a while. Eventually we figured out that for my house, they would literally just be renting a line from AT&T. So I'd be paying extra each month for access to Sonic's support instead of AT&T. I almost did it anyway, but then I found out that AT&T had faster options available to me that they won't let Sonic resell.

FYI, if you get service from a reseller on top of att, you’ll still have att level wait times for service outages since sonic will have to call att. In some areas your wait times will actually be longer because att deprioritizes support for resellers. It’s a mad world.

Sonic is a fair bit more expensive for me, and I can only get DSL so I have terrible upload speeds.

Worth it. Calling support has been fast and pleasant, and I'd prefer to not give that money to comcast.

I had Sonic DSL & was happy. Had to ditch it when until I needed faster upload for work purposes.

Yeah - I get by with overnight uploads or coffee shops, but if I needed it regularly I'd have to jump ship :| It's enough to do video meetings or a single high quality movie stream, and that's really all I need for 99% of the time.

Faster would be nice occasionally, but adblock has a bigger impact on page load time than bandwidth at this point.

MS's findings largely lineup with my parent's experience in rural Montana. Their only non dialup (yes, dialup) was an experimental microwave connection that was expensive, line of sight and peeked at about 5 Mbps down, and far less up. It was also very tempermental, subject to weather outages in snow or rain.

The local cable internet provider wouldn't run cable, as it wasnt cost effective. They'd have had to run around 5 miles of cable to serve less than 50 households.

The cable TV provider did run some cable, and made it obvious they didnt care or plan to really service it. Personally, dont know why they bother other than to possibly meet a regulatory requirement as they didnt run it until 2000-2001 and anyone that wanted "cable" already had Directv or Dish by that point.

After every rain storm or snow melt, a different section of cable became exposed (they ran the cable in the ditch alongside a dirt road on the side of a mountain with less than a foot of top soil on average before you hit bedrock, so cant complain too much they didnt go deeper).

Running five miles of cable would cost hundreds of thousands of dollars: https://www.reddit.com/r/networking/comments/7mp9gz/comment/....

Then, if you’ve got 50 houses, how many will subscribe? Hint: not many, since folks who buy houses in places without broadband probably don’t prioritize broadband. Say you get 20 houses though, with $200,000 invested. That’s $10,000 per house. That’s an optimistic figure. Will you ever make that money back? No. Charter has a market cap of $5,000 per subscriber. If you blow double that on capital costs alone per subscriber, you’ll never come close to making it back. (Especially in a scenario like the above with high maintenance costs.)

In rural areas at least it is not going to cost 40k a mile to run coax. Those kinds of areas are almost always serviced by overhead utility lines that already exist, and just running a new line on those poles just requires township permit. It would be probably an order of magnitude cheaper than the buried lines or new lines building required in other circumstances.

Typical MBA-type way of calculating costs.

Let's take an average over country then apply it in my back-on-the-envelope. Doesn't matter if areas alike Manhattan would probably have 100x the cost of running coax in the rural area (which obviously drives the average).

Then justify the existence of monopolies when in fact purely local providers are actually better able to reduce the costs associated with broadband installation. Those monopolies subsidise central areas by providing way worse service for rural.

I've actually got this issue in London (I didn't check before moving, that our house is literally in the middle of about 5 exchanges). Went as far as reaching providers relating to Gigabit Voucher Scheme or even thinking of asking a friend living about a mile away to provide a wireless link from his house.

While I don’t know what providers have to do to actually access it, the Universal Access Fee is supposed to be there for exactly this kind of reason: providing internet to rural areas that wouldn’t otherwise get it, due to low population levels.

Local cable company wanted $10,000 to run a half mile of cable to my brother-in-law's house. Their other option is some local wireless provider that gives about 1.5mb/s, or cell phones which come in with 1-2 bars of signal.

They offered to split the cost with the other neighbors on the street. One are elderly who aren't interested in cable, the other are very religous and don't own a television or use the internet. IMO it'd be worth it for just the resale value of the houses.

Your going to have to run fibre to the Cab and probably Upgrade the Cab as well

Oh, I am well aware of the costs involved. Hence why I stated they likely only ran the cable to fulfill a regulatory check box and appeared to have done it as cheap as possible where you could see dozens of yards of cable exposed in a ditch, or even exposed to vehicles driving over it. But that's also why Congress has passed acts to compel providers to provide service.

Hell, I currently live in an affluent Chicago suburb, and I have exposed wires from both Comcast and AT&T just inches from the sidewalk out my back door. My dog literally takes dumps on the internet/TV cables to the townhouses I live in.

Haven't had AT&T service since they sold AT&T BI to comcast, but Comcast has the balls to tell me when I complain about an outage I should upgrade to business class if I want fewer outages (lower speeds, more money, supposedly service guarantees).

> But that's also why Congress has passed acts to compel providers to provide service.

There are no federal level universal service mandates for broadband. In any event, mandates aren’t free money. Consumers end up paying for it, the cost is just hidden and unaccountable. The mandate obscures the political question of: should we give a $5,000+ subsidy to people out of other peoples’ pockets, or are there are needier people who are more deserving?

Thee whole purpose of monopolies an the actual subsidies that they take with glee is to cover these costs. I completely disagree, Congress should compel these people to provide what they stated. If Romania can provide fiber for their rural populations for cheaper than we have there's no excuse for not being able to provide internet in the United States for all but the most remote locations.

Congress banned monopoly cable franchises in 1992. These companies have no legal protection from competition that would justify a mandate. They also don’t generally receive subsidies. The universal service fund only recently started offering subsidies for broadband. (Obviously, if a company elects to take those subsidies, it should build the infrastructure they say they will build.) And at least for larger providers, the “subsidy” really isn’t a subsidy. The money just comes from a special tax on the industry, so for say AT&T, the “subsidy” is really coming out of their own pocket.

As to Romania—that country has vastly lower broadband penetration in rural areas (where more than half the population lives) than the US: https://www.broadbandtvnews.com/2018/05/22/romanian-broadban.... Also, costs are not easily comparable across countries. The vast bulk of costs of building fiber in the US is labor costs, both for initial build and for ongoing maintenance. Building infrastructure in a blank slate with developing country labor costs is a very different proposition than doing the same thing in the US. (It’s also why China can build tons of high speed rail while the California HSR project ran out of money before getting out of Fresno.)

You’re totally right that it’s much harder to run fiber across the US vs a much smaller country. However I don’t think you’re accurate on a few other points:

> the “subsidy” is really coming out of their own pocket

ISPs lobbied really hard during the 1996 rewrite of the USF so they could add the “Universal Access Fund” line item on your bill separately. So, by and large, it’s not coming out of their pocket.

> The universal service fund only recently started offering subsidies for broadband

Not sure what you’re referencing when you talk about subsidies, but the USAC website says that they offered $200m in 2000, and the number quickly goes into the billions from there.


The USF is charged to providers. Providers can seek to recover some of that through a USF line item on the bill, but that's just optical cover. The actual burden of the tax is split between the provider and customer according to the relative elasticity of supply versus demand: https://www.investopedia.com/terms/t/tax_incidence.asp. (While I'm not aware of a specific study on the issue, telecom service subject to the USF has low elasticity both on the supply side and demand side, so it's probably a reasonable assumption that AT&T is bearing a big chunk of the tax incidence.)

As to broadband subsidies: prior to 2011, the universal service fund only covered subsidies for broadband to schools and libraries, through the e-rate program. It wasn't until 2011 that the USF started subsidizing broadband generally through the Connect America Fund.

You could charge $60/month and assuming there are no marginal costs to providing service and everybody subscribes (two big assumptions) that would be >7% yearly yield. Not bad at all, actually, in ideal circumstances.

Those two assumptions are wrong. Charter’s (who I use as an example because they’re a most pure cable company) shows an EBITDA margin of about 36%. That means that just under two thirds of revenues get eaten up by operating costs. And that’s on average—maintenance costs will be higher in some respects in rural areas where you have to roll a truck every time a tree takes out a power line. You have to charge a lot more than $60 even if everyone subscribes.

Which brings us to assumption 2. At > $100 per month, most people won’t subscribe. (Especially given that, as a matter of local law, in many places you’re required to offer a basic cable tier for $20-$30.) Atlantic Broadband, which covers rural areas around PA and MD primarily, has about 250,000 subscribers, but is available to almost 2 million people. (At least on a census-block basis.) My 40% assumption was quite generous.

> Running five miles of cable would cost hundreds of thousands of dollars

There are more modern laying techniques for rural areas that don't require you to dig up that much. A five person team can lay 3 kilometers of fiber a day in rural areas with modern microtrenching solutions. (looks like this for instance: https://www.youtube.com/watch?v=4bfbh3ILLD0)

The investment on 50 loss making houses can be subsidized by the investment on 50 profit making houses.

Estimates aside (I imagine these vary a lot by area anyway.. city vs rural, etc.), aren't these the same considerations as phone lines?

> Will you ever make that money back? No.

They almost certainly would make it back over a long enough time period. Broadband is going to get more important over time. Even if just a few people sign up now, future residents may prioritize it higher.

Isn't this the EXACT scenario government grants and subsidies are supposed to resolve?

Don't they have access to any satellite services?

I work remotely on a satellite connection. Been doing it for a couple of years.

The minimum connection latency is 800-1000 ms, and the maximum download speed is 1-3 Mbps. That's with a 10 GB bandwidth cap though, so my actual download speed is 10-100 kbps for 90% of the month.

Any live video/audio/screen sharing is not possible. I have to schedule transfers for large files (1-5 GB) overnight, which are not bandwidth constrained during early morning hours.

I can get away with it because all of my work communication is through text channels.

It's not for everyone, but it's possible. For me, getting out of the city and living in the mountains with nature is well worth the trade-off of giving up some modern conveniences.

Yes, but people who can't imagine living without Netflix don't consider that an option although I do know people who work remotely in rural areas for tech companies who make do with just satellite.

In all fairness, it's not a great option. My dad is at the very end of a DSL line in Maine and Internet is pretty bad. Neighbors who are a bit further out use a combination of satellite and wireless hot spots (on marginal cellular) and no one is very happy.

Satellite for internet is pretty terrible. Last I was on that, the ping was over 2 seconds (sometimes 5 seconds). Practically impossible for video conferencing and real-time gaming is impossible. However! You can download well enough.

Asking this as a foreigner - how is the mobile internet connectivity in such areas? (3G, 4G). I wonder if 5G will help such areas where it is too costly to provide cable internet.

My experience visiting two years ago, it was spotty. If you're in town or along an interstate, it's usually good coverage, but maybe not great speeds. E.g. my phone would show 5 bars coverages, but the speeds were noticibly far slower than what I get in Chicago.

I was able to get serviceable, if albeit slow 4G LTE coverage at my parents' home about 7 miles out of town and about 1000 feet on the side of a mountain, but sitting in town at a brewery with a cell tower in sight out of the window, I couldn't get any signal at all. I presume the tower I could see was from an incompatible network.

In Montana, I've been able to get surprisingly strong signals in places I'd least expect it. E.g. I was hiking in the Bitterroot National Forest up Lolo peak, and I had near perfect coverage for nearly my entire hike as far back as 2006. But, if you head out into north eastern Montana, you may well find yourself in a dead zone for quite a while. I dont know if it is still the case, but used to be cells were useless anywhere near an ICBM silo (of which there are quite a few along route 200 east of Great Falls, MT - I attended a boy scout camp a few times out that way in the late 90s).

Regarding 5G, not sure it would help much in my parent's situation as I understand it is even more reliant upon line of site and more susceptible to interference from weather. There are decently sized mountains blocking them between the closest decently sized city and them. There would probably need to be a series of relay towers to overcome that.

Sounds terrible.

I've so much used to have almost unlimited 4G data for 5bucks a month...

I'm in Montana (pretty close to the middle of nowhere, but am close to a highway and a small city). The 4G is dramatically faster than my home internet, but my data is capped for tethering. After I hit the cap, the speed is barely able to load HN in a reasonable time. My internet is DSL and I get about half the advertised rate of 25mb down, 5mb up. Cable internet is available literally across the street, and so I've paid a copay of about $5k to get it run under the highway and be able to be run to my house. I ordered it last October and they estimate they will start in the next few months, weather allowing. The long wait aside, I'm so excited. Supposed to get 400mb down and I think 20mb up.

It is getting better, but a big part of the problem is that it is the only thing getting better. They run fiber lines to cell towers, but the people around still can't even get DSL. More people are relying on mobile data and a few others have gotten mobile data to a home router for internet since the only other option is dialup or slow and expensive as fuck satellite. It is enough for a single person to watch a video at 480 or maybe 720, but that is about it and slightly temperamental.

5g is shorter ranged so you would need more masts and still need backhaul

Microsoft is right. 477 data is notoriously inaccurate and there are few things stopping companies from not exaggerating their coverage.

Basically companies need to reasonably be able to service any address within a census block without a major effort to be able to report that block as covered.

Then if a block is considered to have enough competition providers can’t get funding from various government sources to support their build outs.

To further exasperate the problem it appears the FCC is essentially doing very little validation on the data they are given and subsequently release.

We encountered all of these factors over the past ~5 years building BroadbandNow.com and lobbying the FCC to start collecting better address level data... or to at least minimally validate what they release.

On iur end, with each 477 data release it takes a team of 5 almost 2 weeks to process and merge the data properly. In that process we do a ton of geospatial and some statistical validation to make sure the data actually makes sense. (A CA provider also covers 400 people in Maine??)

We also compare it to historical data sets and a few propritary ones to make our data as accurate as possible. (We’ve also got some other tricks up our sleeve but can’t disclose those).

The only real long term solution is for the FCC to collect and release address level availability data from providers. The providers will rant and rave that this is a regulatory burden but the reality is that most of them already have this data and make it publicaly available if you go through the right channels... this is why we built businessinternet.com. We could easily get our hands on “lit building lists” and even fiber routes for business customers but when you ask ANY major ISP if they have address level coverage for residential customers they’ll tell you their systems aren’t advanced enough to have that data.

It’s crazy.

When we started BroadbandNow in 2014 we did so because we realized there were millions of Americans who we considered “underserved” meaning that they didn’t have more than 2 wires providers that offered 10/4 mbps.

Today we measure that metric at 25/3 but it is sad to say the number of underserved Americans has barely improved. As a team we believe better data improves competition. Hopefully Microsoft bringing attention to the issues around 477 data will create enough political leverage to help fix the issues.

Here's a map from the FCC that will let you look up a location and see what service the FCC thinks is available there [1].

[1] https://broadbandmap.fcc.gov/#/

Echoing the top comment in this thread, FCC lists 6 providers for my neighborhood; in reality there's only two and Comcast is the only one who provides cable.

This means if I have an issue with them, there's no competitor I can join instead and they take full advantage of this. I got it resolved recently after many years of issues, but it turns out that Comcast installed my cables wrong. They

1) Used television grade cables that weren't properly shielded so every time my neighbor ran their microwave my ping would spike to 800-2,000 and

2) Didn't hook up the cables right so my other neighbor and I have been splitting connections for the past five years

Every time I bugged them about this they would send out a tech that would poke around in the attic for a bit, but ultimately just shrug their shoulders and be on their way. I was finally able to get it resolved, but only because I happen to know a local Comcast executive that was able to escalate the issue for me.

If I provided such a negligent service and charged my customers full-price as an individual I would be arrested and charged with theft and fraud.

Same here. As if the issues aren't enough, Comcast is also the "preferred" provider at my building and they are citing "safety reasons" to not let another provider who is willing to extend to my place.

Apparently some company called Pencor has gigabit symmetrical fiber where I live, but I can't find any reference they even provide ISP services, and the mentioned subsidiary only advertises DSL. Considering I live in a rural town of 3k I doubt it.

These criteria are nuts. Since when are satellite, DSL, and "fixed wireless" broadband?

Per Wikipedia [1]: "In the context of Internet access, broadband is used to mean any high-speed Internet access that is always on and faster than dial-up access over traditional analog or ISDN PSTN services."

[1]: https://en.wikipedia.org/wiki/Broadband

For my address, they list two cable providers and of course one of them is not available... Are there any areas in the US where two different cable providers are allowed to service customers in the same area?

bwahahahaha. 6?! I have exactly three. One is satellite, another is point-to-point who do not take any new customers, and the other is DSL with max speed of 25mb down (of which I get half). It says I have 100mb down service available.

In Italian we say “chiedere all’oste se il vino è buono” which translates to “ask the inn-keeper if the wine is good”. The meaning is clear

Here's the link to the actual Microsoft report: https://news.microsoft.com/rural-broadband/

I have a Fios modem in my apartment but with Verizon's extortionate pricing I couldn't ever consider them an option.

I'm paying about $60 for 100/100 FiOS in Pittsburgh. I'm actually more pissed about not having IPv6, despite claims of it coming for years.

This is comparing broadband "availability" (whether you could have >=25Mbps if you paid for it) with broadband "use" (whether your downloads actually go at >=25Mbps). The latter is a better metric, but there isn't any mystery why these numbers are different.

It's 2019. For a first world country even the BASIC package, within an actual city, should fulfill the "minimum broadband" definition of 25Mbit/sec down and 3-5Mbit/sec up; even during peek hours.

If it isn't being selected, then clearly there's insufficient competition to drive better prices and better packages.

Akamai's speed test reports are getting a bit dated, but in Q1 2017, no country had more than 40% at 25 mbps+: https://www.akamai.com/us/en/multimedia/documents/state-of-t.... The U.S. was #10 at 21%. The only European countries to make the list were Finland, Denmark, Switzerland, Sweden, and Norway. Putatively first world countries like France and Italy were at 18% and 12% with 15mbps+.

Even today, I'd wager that for the vast majority of the EU (by population, so the U.K., Germany, France, Italy, and Spain), 25 mbps is closer to the median than the floor.

That's quite interesting. I live in rural Japan and have fiber, so I get well above those speeds. I had a quick look at the document and I think things certainly have changed in a few years -- at least in Japan. These days there is virtually nowhere that you can't get 4G as far as I know. The vast majority of people here use the internet on their phone so I'm willing to bet that percentage has gone up a lot. Feature phones used to be popular, but now if you don't have a smart phone it's pretty darn unusual. There is very little free wifi here as well, so almost everybody who has a phone has a 4G data plan.

In terms of availability, even fibre is ubiquitous now. At least in the Tokai region, I would be shocked if there was a single town that didn't have fibre now. NTT rolled it out aggressively about 10 years ago. When I got hooked up in my apartment 4 years ago they didn't even bother to check availability.

It would be super interesting to see what the numbers are like now. Of course it's apples and oranges. The US is much harder to serve than Japan (although I guess the many earthquakes keep the fibre repair guys busy). But I think you'll see SE Asian countries having nearly ubiquitous coverage in the next handful of years.

I always wonder how much of this is availability vs what the market wants. I'm constantly surprised how many people just don't pay for faster internet even today even though they have the option.

>If it isn't being selected, then clearly there's insufficient competition to drive better prices and better packages.

No, not everyone is going to have your use patterns or knowledge. Most ordinary folks are just price sensitive. "What's the lowest price thing I can get that's not dialup."

I agree there need's to be more competition, but so long as there exits cheaper packages bellow the "minimum broadband" definition there will be many normal people who buy them.

Good point. I live in an area that has broadband availability, and I choose to be on Comcast’s cheapest tier (18/3) because I don’t need faster speed and it would cost $250/yr to be at a higher tier.

If there were more competition (Comcast has a monopoly in western SV), prices would likely be lower and I’d probably be at/above 25mbps for the same price I’m paying now.

Hmm. I am paying more than I should for fiber through a rural coop, but I have fiber. It’s been way more stable than any broadband (including a local line of site WiFi) and my prior abode in the Bay Area.

While 4B people don't have access globally, it's astounding to me that affordable broadband access is still an issue even in the United States.

Glad that new solutions like satellites [1] are coming online to solve this problem.

1: https://www.forbes.com/sites/alexknapp/2019/01/16/twenty-per...

Disclaimer: I'm an investor in Astranis, mentioned in the Forbes article.

> it's astounding to me that affordable broadband access is still an issue even in the United States

Not sure how to respond to this, other than to point out that by most metrics, the US is largely a 3rd world country:






etc., etc.

Except when it comes to broadband! The US is top 10 for average connection speeds in both Akamai’s and Ookla’s global surveys. This is the mind boggling thing to me. By objective metrics, our infrastructure is shit compared to other first world countries, except when it comes to broadband. Roads, schools, transit, etc. Broadband is like the single infrastructure thing we do better than say Germany. But for some reason broadband is the thing folks on HN complain about.

The US as a whole might have statistically great broadband, but that's cold comfort to folks paying $200/month for 10 megabits. Roads are better in some counties than others, but generally not two orders of magnitude worse on the other side of town.

Every hotel or airbnb I've been to in Europe has had pure shit internet. 1 Gigabit to every hovel or cabin in Europe seems to be a meme, not reality.

Yes, every (almost) hotel I've been to in Europe has had questionable internet. Homes are a completely different thing. Hotels provide enough capacity to be able to charge you for access, homes get enough to do the things they want to do.

Satellites will never fix the broadband penetration problem. There are fundamental laws of physics in the way. If you can change those, let me know.

If you want to solve broadband adoption, you have to get Congress to stop allowing broadband companies to lie through their teeth.

For anyone else curious... geostationary orbits run at an altitude of 22,236 miles or 120ms. Double that for round-trip and bump it by a bit more for the angle and I'd expect pings >= 300ms on top of the delays in the electronics.

current satellite internet is also in geostationary orbit (22k miles):

    IP Gateway REDACTED
    Packet Loss 0%
    Average Delay 621 ms
    Minimum Delay 570 ms
    Maximum Delay 650 ms
that's to satellite. here's a traceroute to show you actual times:

    traceroute to (, 64 hops max, 52 byte packets
     1  1.941 ms  0.933 ms  0.930 ms
     2  2.535 ms  1.232 ms  1.313 ms
     3  redacted  2.668 ms  1.366 ms  2.584 ms
     4  601.534 ms  569.407 ms  600.134 ms
     5  1483.616 ms  615.166 ms  620.347 ms
i managed to get this low of traceroute times as the clouds happened to be gone for the moment, but given that i live in the nw, that is a fluke :)

an article on LEO options from SpaceX notes the same latency: https://arstechnica.com/information-technology/2016/11/space...

i'd love something as low as 300ms (or the 25ms that LEO is promising), but it's just not going to happen.

Why don't you think the lower latency LEO is going to happen?

ah, my note was that i'd love something as low as the 300ms that math seems to think we'd have, and the 25ms that LEO should provide - that's where my skepticism lies.

as for whether an LEO solution will occur for rural americans? the monetary math for that doesn't make much more sense to me, but lots of things get deployed that don't make monetary sense so that's at least possible.

That's a lot better than existing sat based solutions. if the throughput could get close to 4g speeds I'd see it as a legitimate contender, but there are still some applications it will never really be viable for unless you can build some sort of network of the things and circumvent most of the terrestrial network entirely until you get really close to the destination, But even with that I doubt it could get much better.

What about low orbit satellites?

Local Loop Unbundling is the way to go

> Glad that new solutions like satellites [1] are coming online to solve this problem.

i wouldn't call satellite service in geostationary orbits (which was the subject of the article, though they did mention LEO options as competition) something that really "solves" this problem. you end up with fairly low bandwidth with current providers (for instance, hughesnet maxes out at 25mbit, which is theoretical since busy hours mean much less), with enough latency to make anything interactive extremely painful ... connecting to a remote shell without something like mosh is worse than 300baud dialup due to the latency - first hop to the satellite typically runs 800-900ms, and the next hop to a terrestrial gateway is 1200-1300ms.

streaming isn't exactly feasible either, given the latency and that a typical plan is measured in gigabytes per month (before you are throttled to an average speed of 1.5mbit).

but it works where there are no other providers and cell service doesn't reach.

The only solution is fiber optics. To the home everywhere its not cost prohibitive, and then via 5G everywhere it is (ie, anywhere there is insufficient development to congest wireless networks).

To make that happen will require substantial federal legislation that will need to overcome the lobbying and influence of a 500 billion dollars worth of big telecom parasites and near total change of leadership at the FCC.

It should be publicly funded, publicly owned, and done akin to the Eisenhower Interstate System of the 50s, because it is about as important as those highways were to the future prosperity of the country.

But Google will roll out Fiber nationwide in the US and via balloons in Africa!

I'm not really aware of how FCC collected this data and how the analysis were done. I'm also not aware of the policies that are set by the US government to provide rural connectivity. But I have a few general thoughts on the issue.

Estimating the broadband adoption is a tricky issue because of number of stakeholders involved. First there are the companies that provide the physical connectivity. Their job is done as soon as the lines are laid to the base station or the satellite is sent into the space, etc [1] According to them anywhere the physical layer has reached there is internet connectivity (which is true...to an extent)

But this does not actually get the internet in to your home. Local distribution is a very big challenge. A similar thing happened in India with Bharat net [2]. While the government successfully laid down the fiber optic cables the last mile connectivity was overlooked. So of course the broadband is inadequate when the capacity is not being used.

I agree with the article that internet is connected to a higher economic growth and overall prosperity. The same sentiment is also echoed by the WEF report [3]. However these reports miss one very crucial point. Internet is not a service provider. Internet is a transport. You need to have people provide service on the internet if it is to be used. Local internet economy is almost non existent.

Many of the faults in FCC's report can be explained if you look at the internet provider business as a marketplace.

- Big companies own the platform and assert that the platform can reach anywhere. Which might very well be true.

- Local internet service providers use this platform to provide services to the customers.

- Customers pay for the internet service according to usage.

Why should a customer pay heavy charge for the broadband usage? Why should the local internet service provider invest into last mile connectivity?

On a global scale internet rapidly reached new levels of adoption when the application layer succeeded. Something similar needs to happen on a local level, I think. And this is what microsoft is also trying to do with their airband platform [4].

[1] If you do a bit of digging into physical layer distribution there are many exotic ways to provide network connectivity including sending weather balloons up in the air! Very interesting area of research if you're into this sort of thing.

[2] https://www.prsindia.org/parliamenttrack-policy/report-summa...

[3] https://www.weforum.org/projects/internet-for-all

[4] https://www.microsoft.com/en-us/airband

Except that for most of the country, there aren't actually local ISPs. There are enormous regional/national ISPs—Charter, Comcast, AT&T, etc. If there really were local providers, they might have more of an incentive to invest in last-mile connectivity.

(Of course, the really effective way to do that would be local loop unbundling and requiring a certain level of physical connectivity for everyone.)


Don't post low quality zero content garbage to HN.

None of this is gonna matter, worldwide, when Musk launches his constellation. And he will launch it eventually. I, for one, will sign up even if I don't need it, provided there's a plan for under $50/mo, just to show my outstretched middle finger to Comcast.

Microsoft is basing the data on users of their services, which is largely a self selected group. Would it be surprising if people using Bing instead of Google or Duck Duck Go were the kind of people who use Windows, default to using Bing, and don't see the value of a high speed internet connection? I have older family members who fit this profile.

Bing isn’t the only service Microsoft has. Update, office, Xbox, Windows itself, etc

Last I checked Windows was still the second most used OS. I am sure Microsoft collects a lot of telemetry through Windows even when the user is not using Bing (or Edge.)

It's the primary domestic OS isn't it?

I strongly suspect that the data comes from Microsoft's MSN page. Video downloads is really the only usage scenario broad enough to generate such a data set. MSN serves loads of them and is the default page on Windows. Since their contents target the audience favored by advertisers (i.e. young urban folks with money), the data could have a strong selection bias. Rural Americans simply aren't viewing the pop-culture-centric materials on MSN.

How about Windows Update, Xbox, OneDrive . . . There are plenty of services Microsoft runs that could help provide this information.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact