Hacker News new | ask | show | jobs
by lprubin 2034 days ago
> Both Greenland and Bermuda can blame their expensive packages on geography and uptake, but what about the United States? At $50.00 on average and coming in in 119th place worldwide, one would expect American packages to be considerably cheaper. But while broadband in the United States is widely available and uptake is high, lack of competition in the marketplace means Americans pay far more than they should compared to much of the rest of the world.

Does the HN community agree with the assessment that the lack of competition is why the prices are high? If so, does anybody have some explanations/links for why there is a lack of competition and some proposals to improve the problem? I know community broadband projects are starting to popup, I'm guessing partially due to this issue.

Also, is there a tradeoff? Is the "widely available and uptake is high" benefit in part because of the low competition?

7 comments

Part of it is the odd way Americans respond to try to increase competition.

Essentially, entrenched players are given the pens to write their own rules under a banner that reads "deregulation" effectively locking themselves in more.

Then people blame "the government" for this problem and hand the pens over to the companies again to write even more rules that favor themselves.

Pointing out this brazenly obvious pattern is shouted down as socialist or leftist, which makes it a Forbidden Thought in America so it continues unabated out in the open. It's pretty remarkable.

The majority of people in the US live in a place with only one useful broadband option, which is usually the cable provider along with DSL which I do not consider broadband.

If you recall the drama of Google Fiber and other relatively recent attempts to rollout fiber, the existing companies have often gotten local and state laws passed making it extremely difficult to add competitive services at a more reasonable cost.

Well managed DSL isn't terrible. I've got bonded VDSL2 at a combined 80 Mbps/10 Mbps. That's good enough for most things, although it's tier limited (my lines could do more, but the sync rate is capped), and I'd love more upstream, and a bunch of other stuff... But it's not bad enough for me to foot the bill to get muni fiber extended to my house (and then tear up my driveway to get it to the house).
From what I understand this isn’t what’s available to most. Anecdotally the DSL I’ve found available as an alternate at a few places in the country has been less than 30 Mbps at best.
Might make sense if you view it as a capital improvement. I wouldn’t consider buying a house without fiber, or if I had the option, I would value a house with fiber much more.
The guestimate from the utility district was $50,000. And that probably doesn't include onsite work, because I'm quite a ways back from the street and have underground power and telephone, so there's no poles to go from the steet with aerial fiber.

I'm going to hold tight and wait for either a neighborhood group to form (unlikely) or someone else on the route to extend the network so I don't have to extend it all the way.

Lack of competition is a big factor in US pricing, although cost of labor, and regulatory burdens probably raises the floor price.

Competiton is rare because of several factors. The most important one being overbuilding a new network on an area with existing wiring is expensive, time consuming, and unlikely to result in enough customers to make a good return on investment; especially since the incumbent networks are likely to respond by upgrading their network, usually faster than the overbuild. For a historical example, look at Google's fiber build out announcement for 10 or so cities, followed by AT&Ts announcement for most of the same cities, and then Google's retraction after AT&T actually built the network, and Google hadn't even got planning approval for their Fiber boxes.

The solution is local loop unbundling, as congress provided in the telecom act of 1996, and the FCC dismantled over the ensuing decade. Ideally, more fully realised as in other countries, where the local loop provider(s) are never the service providers. It doesn't make sense to run a large number of networks to the same houses in the same city; instead, that part should be leasable, with providers servicing the connections from central locations.

this supply-side analysis is relevent and well-spoken, yet one wonders what has been done with the already-in-place guranteed income from managed-monopoly markets around the US for the last 25 years ?
It’s all about competition — these services have 60-90% margins depending on the capital expenses needed in a region.

The costs of all of the service delivery components drop over time — automation eliminated many of the expensive positions and the costs of switching and other gear has been in free fall for years.

My brother in law lives about 1000 yards from my house, in a suburban town. I’m in the city. My spectrum internet is $70 for 75 down, he pays $32 for 150. Spectrum is not losing money.

The difference? The town has FIOS. Verizon isn’t moving into the city because the margins on future 5G connectivity is much higher than FIOS.

It's pretty much directly a result of the traditional regulatory/structural preferences of the US. Works well for some things, but not so great for telcom/infrastructure.
There is no competition; let's exclude the probably sub-10% of the US where you are in a major city, live in a MDU, and can get genuine fiber (at least 1G/1G, last mile metro-e/eth/fiber, for the purposes of this), or live in an area that has at least one fiber provider.

Your choices are: cable (either Comcast or Spectrum, pretty much everyone else has been acquired into one of these), ADSL1 (3 Mbps or less), LTE ($10-15/GB used, no carrier actually has an unlimited plan and they will terminate you for using a phone plan SIM in your hotspot, all heavily DPI and modify your traffic). In rare areas, AT&T or similar offer "home internet" (250GB cap $60/m on contract) over LTE.

Comcast, Spectrum, et al have long since advertised fiber (they do coax/cable). They brag about their fast fiber network in an attempt to mislead people in the US that don't know better. Last mile is cable, and upstream is generally sub-10Mbps + because the last mile is cable, your latency for things such as video games are significantly worse (15ms vs <1ms to something in the same city).

In fact, Comcast does not even list upload speeds on their site anymore the last time I checked. No matter how hard you drill down, click Details, read the fine print, etc. This is incredibly good for "shut up and consume media" and incredibly bad for "learn, upload, create". On top of that, they repeatedly use the word gigabit for plans offering sub-ADSL upload speeds. They do not mention upload. It appears that currently the highest tier plans on Comcast offer up to 35 Mbps upload based on HN comments and DSLReports forum posts.

Comcast and their cable friends also repeatedly attempt to use their billions to sue small towns into the ground when they try to start a community broadband project[0][1][2][..etc]. Trying to start muni fiber is a great way to drown in lawsuits from every incumbent provider that _does not even serve your area or have plans to serve your area_, they just want to set a precedent. Like they literally do not care for your city at all and do not plan to ever offer service there because it's a waste of money to them, but they will happily spend many millions trying to kill this.

Contracts are usually mandatory, signed for several years, along with other anticompetitive things like requiring the use of their remotely-administered equipment that you are locked out of, or charging you large fees to use your own. For example, ludicrously low bandwidth caps on Comcast, where they charge you the cost of an unmetered gigabit plan in most countries just to remove the cap. But you can get a discount if you use their remotely-controlled hardware that also requires you to sign in to your account to view your wifi password and runs a public hotspot off it.

[EDIT BASED ON COMMENT REPLIES: OK, contracts are no longer mandatory these days, but in reality you are going to pay at minimum 1.5x to 2x the cost per month without the contract discount, at least on Comcast. For the vast majority of the US, they will often take this, eat the hard credit inquiry hit to save on the money, though.]

There are a few incumbent-esque fiber providers like Frontier that offer bizarre speed choices and mandatory contracts (2 years is often the case), which cap out at something like 200 Mbps for "FTTH" which is odd, and has useful support roughly equivalent to Comcast on a good day. These providers also have spent money lobbying on the above instead of, I don't know, updating their networks.

Generally in any location in the US (with exceptions like in SF, Seattle, Oakland, etc. where you are lucky enough to live in a MDU with multiple fiber providers or a municipal broadband area), you do not actually have a free choice of provider, it's either cable or ADSL1. Almost no provider of DSL does VDSL or ADSL2+ or similar; if you are in an area that demands DSL it's likely you are sitting at 3-6Mbps down and sub-1Mbps up. Almost no provider will bond multiple DSL lines.

In most other countries, DSL is generally significantly more tolerable to use (and I love the fritzbox compared to the abomination of ISP-provided hardware I've seen in the US), which is probably one of the reasons why people think the US has choices when it comes to DSL. They have VDSL2 - the US has ADSL1 that caps at 3 Mbps on a good day and goes out when it rains; DSL is the "I have no other options" choice, not the first choice. For example, in Germany, 1&1's $25 DSL plan actually beats Comcast 'gigabit' on upload.

[0] https://broadbandnow.com/report/municipal-broadband-roadbloc... (potentially biased source; affiliate advertising for home internet)

[1] https://www.theverge.com/2015/5/1/8530403/chattanooga-comcas...

[2] https://muninetworks.org/content/mediacom-lawyers-slow-compe...

>Generally in any location in the US (with exceptions like in SF, Seattle, Oakland, etc. where you are lucky enough to live in a MDU with multiple fiber providers or a municipal broadband area),

I've never seen a residence in the US with multiple fiber ISP options. And when purchasing a house in the past year, I noticed fiber was available if the entire subdivision was newly built in the past 5, maybe 10 years, otherwise fiber is unavailable. So the telecom companies are laying fiber, but only if the entire area is new and they're laying new everything. If you build a new house in an area with existing homes, you won't get new fiber. And I didn't find any pre-2010 houses with fiber period.

In the bay area there's some overlap. I've seen places that have Webpass in addition to something else, but it's always a MDU and not a single family home. I'll edit it a bit to clarify.

From what I've seen fiber providers are split into these:

- $$$$$, business, anywhere, pay per mile to pull fiber

- old-ish home incumbents like Frontier, often strange plans like 100 Mbps and 200 Mbps or 250 Mbps - I have no idea what their equipment looks like to provide this

- planned neighborhood telcos like https://smartcitytelecom.com/residential/

- municipal fiber, local utility co (EPBCo, Longmont)

- MDU-only urban fiber (Google Fiber Webpass)

Ting brought fiber into our 40 year-old neighborhood a couple of years ago. Symmetric gigabit with no caps.

https://ting.com/internet/town/centennial

I’m in a neighborhood developed between 60 and 75 years ago. We have fiber.
They probably had to replace aging infrastructure. Telcos aren't laying copper anymore. Verizon caught flak for neglecting aging copper networks and wanting to upgrade everything to fiber.

A drawback of fiber is that it generally doesn't work during power outages where as copper does.

The VDSL2 remote terminal that services my house cuts off the second utility power does. So don't count on copper working. Although when the utility power outage was due to a downed line between me and the remote terminal, the DSL didn't drop.
Not really. They mostly just threw it on the poles.
Bell South ran fiber all over metro Atlanta in the late 90s but never into neighborhoods and they never lit it up as far as I know.
Sonic Fiber is in many places in SF. I have it. Best ISP ever.
Exactly. They don't lay copper anymore, just fiber and then use an ONT to break it up into Phone, Data, Security, and Video.
I don't disagree with your sentiment but you have some factual errors that others will use to invalidate your argument.

Contracts aren't required and for a long time there was backlash against them so they weren't pushed hard.

Upload speeds aren't that bad, they're usually around 1/4 of download speeds. I had 125mbps Comcast for years and the upload speed was around 30mbps.

In every market I have live in, the US broadband market generally has 2 providers. One offers a cheap bottom tier and then higher tier ridiculous plans no one would buy. The other provider is the opposite, offering a comparably worse low tier plan but better higher tier.

There's no competition, rather a gentleman's agreement that one provider would own the low tier and another the high to give the impression that there isn't a monopoly.

Who is the high or low tier provider varies by market but there's never any real competition.

I've updated my reply a bit to take these into account. The last time I used Comcast it was something like 400 Mbps down, advertised 10-15 Mbps up (reality sub-5). The Comcast site no longer shows upload speeds anywhere - I remember they used to at least do that. I don't think Comcast has any upload tier above 35 Mbps in 2020, though.

When I was in Europe these were considered DSL speeds.

The atrocious thing I remember while ISP-searching is that virtually all cheap plans would demand you provide proof of assistance, like if you just wanted to sign up for a 10 Mbps plan for backup, there was no such thing, Comcast would demand documentation[0] of low income.

[0] https://internetessentials.com/

Additionally, the major cable providers (e.g. Comcast, Spectrum) all support BYOD and actively maintain lists of certified devices you can purchase.

I was on Comcast for over 15 years and never once used their equipment.

The equipment thing is in reference to their latest cap rollout universally (it's now applied everywhere, instead of just the areas they don't really have competition in) - $30/m addon to remove the cap or $10 (13?)/m if you rent their gateway.

Anecdotally, I used to get an obnoxious amount of robocalls and emails asking to rent their modem back in the day ("We've noticed that you have an older modem that isn't giving you the best service possible.") when I was on a low plan as a backup service. I wonder if they still do that.

> Upload speeds aren't that bad, they're usually around 1/4 of download speeds. I had 125mbps Comcast for years and the upload speed was around 30mbps.

I don’t believe this. I have never been able to find any upload specs from any cable coaxial internet provider. In addition, the multitude of homes with Comcast I’ve been in crap out after a couple simultaneous FaceTime calls.

If Comcast is capable of giving you 30mbps upload consistently throughout the day, you must be super lucky. If they actually budgeted for this for everyone, they would advertise it. But since they don’t, the only conclusion is they massively oversubscribe the upload.

I know their network varies greatly from market to market but, in my experience across 3 markets and 15 years, they always delivered what they advertised. That's not to say the prices were fair or the speed was good, just that it was what was sold.

At one point I was running 4 Nest Cameras that were uploading to the cloud at a constant 27mbps. I was running a Unifi Gateway and religious compared my bandwidth consumption with what Comcast reported and they always erred on under reporting.

Their data caps a complete bullshit though designed to extract revenue from anyone who wasn't a cable tv subscriber.

Interesting. I wonder why they wouldn’t advertise their upload speeds then. I’ve called Comcast about numerous locations and they would never promise me any upload performance, which I take to mean that they are not willing to upgrade their network in the event people saturate it, or more houses are added to the neighbors.
I'm guessing that, up until streaming started to take off, it was mostly irrelevant because it was, and still is for many, sufficient.

Having asymmetrical channel bonding allows cable providers to use cheaper equipment and still offer gigabit download speeds.

> Upload speeds aren't that bad, they're usually around 1/4 of download speeds. I had 125mbps Comcast for years and the upload speed was around 30mbps.

You've been very fortunate. Upload speeds are atrocious. I've lived all over the US, and been stuck with cable Internet almost everywhere. I've never had more than 10M up with cable. Right now, I have 230M down and 5M up with Comcast. Once I was lucky enough to live in a FIOS area, and had 100M/100M.

At Comcast speeds, it's impossible for me to do offsite backups (except for selected files) and forget about streaming any video. I started building a Plex server for the family but gave up when I realized it wasn't possible for even one outgoing video stream.

There are valid technical reasons why cable has such terrible upstream bandwidth. I just wish FIOS would roll out faster. And StarLink.

No contracts here for years. Have choice between super reliable and low latency cable and super reliable and low latency fiber. Fiber has 5X higher speeds for given spend, but either way is objectively sufficient for household needs. Neither have monthly limits. 5G is being deployed and will someday be a third option.
If this cable provider isn’t advertising upload bandwidth, then it isn’t the same product as the fiber. If you intend to have security cameras/video calls, screen sharing simultaneously between multiple people in a house, I haven’t seen a cable connection be able to handle it.
One factor that doesn’t seem to be considered is cost of labor. American technical labors aren’t cheap.
It also isn't cheap in most comparable countries. I think that a quick look at the interactive map show some correlation between labor cost & broadband price (that's probably why Norway stands out, for example), but it's also obvious that it doesn't dictate the price .