3Mbps uploads still fast enough for US homes, Ajit Pai says in final report | Hacker News Reader

3Mbps uploads still fast enough for US homes, Ajit Pai says in final report

arstechnica.com

34 pointselsewhen53 comments

3Mbps symmetrical is the bare minimum for an HD Zoom call. That's good enough for one working person. People in cities have gotten the privilege and luxury of routinely getting network upgrades while rural communities are basically left in the dust. I've been stuck out in the country because I got COVID out here and Zoom calls degrade frustratingly fast.

One reason is because the only method of delivery that they have is point to point connections, which have a good amount of variability in them despite advancements. When I used to live out here I lived in a different area served by CenturyLink. These lines were incredibly old and in disrepair. So, speed isn't the only problem but slow and unstable is not okay in a day when most things require a consistent connection.

So, is it fast enough? I don't think so. I think a policy which stops upgrading cities until a baseline is met in the rest of the country is also appropriate.

> I think a policy which stops upgrading cities until a baseline is met in the rest of the country is also appropriate.

What does one thing have to do with the other? Cities aren't in any way preventing improvements elsewhere.

> Cities aren't in any way preventing improvements elsewhere.

Opportunity cost. Why spend your installation technician's time in an unprofitable rural region when you can better-spend their time in a profitable city?

Well, these rural areas keep electing "free market" republicans, so...

Personally, your argument seems dubious but I'll answer you in earnest.

If your argument were true this map would look much different: https://www.fcc.gov/health/maps

Local politicians all work just about the same way when it comes to approving an internet provider. Attracting them generally has to do with business incentives, which means if the State and/or Federal government do nothing then you'll get what you have today.

Opportunity cost is only an argument when doing one thing prevents you from doing the other. What's the limited resource here that would prevent a cable company from hiring more technicians and laying more cable?

Put another way, how does banning improvements in New York suddenly make improvements in Montana more profitable?

Sure but what are you going to do with those technicians after they've installed the hardware in a sparse region? It takes a lot of man-hours to do the install, especially for long distances that are rural homes, but not so many man-hours to maintain

Plus you have the ability to serve more people, at a lower cost.

I can get 100 urban users up for 1/4 of the price to get 20 users in a rural sub-division. This isn't just capitalism, it's basic utilitarianism.

Cities aren't in any way preventing improvements elsewhere.

Yes they are. No one has unlimited capital for investment, so infrastructure companies will invest where they get the maximum returns. If they're left to their own devices cities would get increasing fast internet, and everywhere else would be left stagnant. That drives more people to live in cities, which in turn makes rural economies fail. Then the people in cities start running out of food.

Slight reductio ad absurdum argument on my part there, but it's just to illustrate the point. For a country to operate having significantly worse infrastructure in one place damages the entire economy. Infrastructure investment has to be managed carefully.

I hear your point but I disagree with the binary nature of it. I do think the challenge of rural internet lies in the incredible amount of infrastructure costs with limited recovery. This is the same challenge the electricity markets have (or any public infrastructure). The other challenge in the US that most european countries don't have (for anyone from Europe) is that the US has incredibly large land mass with sparse populations and challenging weather conditions (tornadoes, hurricanes, flooding, etc.).

This is where something like starlink et. competitors can provide a much better cost per outcome benefit. The challenge is how to bring the cost down for the benefit of rural areas on lower incomes. As it stands right now its pretty expensive for lower-mid income demographic to swallow price wise.

Wouldn't the Skylink latency be too high for video chat?

Internet providers bring in insane amounts of cash and deploy as little of it as possible. If you ban infrastructure improvements in one place I doubt they deploy that money elsewhere. If the other project were profitable they would have done both to begin with.

If a Nation mandate electricity, water pipes and telephone lines ( Not sure if telephone is still the case ) to new housing / flat. Or is that not a case in US?

Why isn't High Speed Internet Access the same? Where High Speed Internet is defined as user able to purchase 1Gbps Internet should they choose to. They could get a 100Mbps if they want it cheaper, but 1Gbps must be an option. This ruled out any DSL tech and most old Cable DOCSIS standards.

And if Phone jack, Which I assume they are all RJ11 used around the world, why cant we do the same for SFP+?

We have tired old arguments in the US about mandate versus incentive. They're two schools of thought and I'm not really sure which is better, when given maximal effort.

Consider this map: https://www.fcc.gov/health/maps

You can mandate that the internet become a utility, much how water can be a private utility, and tell a company that for every segment they get in a popular area, where they'll make money, they must establish presence in an area where they'll lose money in the name of service. There's also price approval activity and probably some subsidizing that happens there. This certainly might work.

The other is for incentives to be drawn and invested in routinely that draw providers to less populated areas. These would have to be monetary, and you might be able to reason their subsidies look much like agricultural subsidies.

Another argument is wireline providers are just not sufficient for scale in the more rural areas. 5G won't work because coverage is abysmal. Starlink and other engineering solutions may be appropriate.

I don't really care how it's done, but being out here has made me realize exactly how awful this problem is.

Not for individuals. No idea what rules there are for developers, but an individual can buy land anywhere and is not required to have any of those things when they build their home. The only things almost universally regulated is sewage. Some states allow composting toilets, but most require a septic tank inspected by the county. Some counties in some states have rules about types of septic systems.

Point to point can also be useful when getting a building connected is expensive. A former employer leased a space in a city, and the space wasn't connected. One of the big providers wanted $10k to connect the building. A local small ISP connected the building using point to point for a fraction of the cost. I think we'll see more point to point adoption until it threatens the market share of entrenched providers.

> I think a policy which stops upgrading cities until a baseline is met in the rest of the country is also appropriate.

The best I can get in San Francisco is 5Mbps symmetric. Would that be good enough for you? :)

The state of US technology has become embarrassing. How much bandwidth does a zoom video need?

Ajit Pai has held back the US and our economy with his action.

Edit: the exact Zoom bandwidth specs was rhetorical. Bandwidth should be symmetrical based on how we use it today and 3Mbps is too slow to support a household with multiple users.

1Mbps for "high quality", 3Mbps for 1080p.

By what reasoning is 3Mbps not good enough for average member of the economy?

Because a household family uses internet for more then a single zoom video. If I have two kids and work from home, 3Mbps isn’t sufficient to support a house.

Or if you have two people at home working and doing web meetings. I've learned that lesson during COVID.

That problem can be solved very easily and very cheaply with better communication and scheduling rather than any sort of tech though.

Have all companies schedule all of their meetings based on their employee's housemates' meeting schedules in a different company? That doesn't sound very easy.

That's like saying we should just work different hours. Boom simple solution, just need to get 20 other people on board to work nights instead of days.

From the Zoom system requirements:

1:1 call

* 600kbps (up/down) for high-quality video

* 1.2Mbps (up/down) for 720p HD video

* Receiving 1080p HD video requires 1.8Mbps (up/down)

* Sending 1080p HD video requires 1.8Mbps (up/down)

Group call

* 800kbps/1.0 Mbps (up/down) for high-quality video

* For gallery view and/or 720p HD video: 1.5Mbps/1.5Mbps (up/down)

* Receiving 1080p HD video requires 2.5Mbps (up/down)

* Sending 1080p HD video requires 3.0Mbps (up/down)

https://support.zoom.us/hc/en-us/articles/201362023-System-r...

yes, but most households have more than 1 person living there and sharing the connection for their own Zoom or YouTube.

Now multiply that by 2.53 (average US household size in 2020) and you'll see why 3 Mbps are probably not enough.

Exactly my point! Consider a group of college students living together or an average family and we start to realize that its not about a single activity but the multiple parallel needs for the same shared pipe that causes congestion.

I find it surprising that so many seem to disagree that this is ridiculously low.

I'm surprised the screen-sharing bandwidth required is so low... I guess it must be quite low framerate? Because some of us have giant monitors...

It's usually pretty easy to compress because you're so often sharing text on a solid background. I'm figuring that there's gotta be a higher bitrate for sharing, say, a movie or video game than sharing my jira board.

Unless you're streaming video games (high FPS, as you noted the inverse, and significantly larger portions of the screen being refreshed, usually the entire one) the 'deltas' are relatively small. Combined with low frame rates, your usual business presentation occupies a small share of the bandwidth.

on the other hand, I'm mostly staring blankly at my webcam with a static background. though admittedly the webcam sensor is a lot noisier :)

but yeah, I suppose the update rate of most screen sharing is pretty low.

I am all for high bandwidth, but it blows my mind that people have the expectation to do video conferencing at 1080p quality. It is bluray quality, which was only available 10years ago in offline / disk format. It is great if we can but I wouldn't call it a requirement.

It's not really blu-ray quality though. You could have a 4K video conference at 1Mbps if you so wished, but it wouldn't look nice. Likewise a 1080p Zoom stream isn't really anything like a blu-ray video.

The quality may not be there but the pixel count is. I agree most webcams are not precise enough to give a real 1080p quality + bandwidth constraints. But you would be better off with a fluid 720p.

His argument is predicated on the antiquated assumption that ISP customers can be provisioned primarily as "internet consumers". The adoption of virtual meetings has exposed this notion to be the fallacy it is.

There’s a lot of country out there in between those two coasts.

The FCC has to set a baseline number for what it determines meets the statutory definition of “advanced telecommunications capability”.

Practically by definition this number is going to be set based on what’s possible to provide over several miles of copper (long distance DSL).

> We find that the current speed benchmark of 25/3Mbps remains an appropriate measure by which to assess whether a fixed service is providing advanced telecommunications capability. We conclude that fixed services with speeds of 25/3Mbps continue to meet the statutory definition of advanced telecommunications capability; that is, such services "enable users to originate and receive high-quality voice, data, graphics, and video telecommunications."

The statutory definition is not whether you can download that iOS update in under 30 seconds. I think it’s hard to argue that 25/3Mbps is not a reasonable minimum for “originate and receive high-quality voice, data, graphics, and video” when the engineering specs for HD streaming are actually within that envelope.

You can move the goalposts and say it needs to support 4 concurrent high quality video streams in a single household, but keep in mind the statutory language was written in 2006, and even as of one year ago no one would have reasonably claimed the minimum viable service level should be set anywhere near that high.

I think it’s reasonable to say that in light of the massive shift to remote work and remote learning it’s worth reassessing the national minimum standards for broadband across our 3.5 million square miles of land. That’s a job for Congress to change the standard.

I can safely say that the ISP situation in the US is simply abysmal. I'm an expat here in France and the market for ISP/mobile is pretty great. Companies are very competitive in their offers, and I can switch at any moment since contracts are less common.

Whats more, fiber is mostly available throughout the country. Our current administration has an objective of providing even more infrastructure to remote locations. Obviously, US is much larger than France, but I still maintain that it can be vastly improved.

Concluding that 3Mbps is _enough_ is not grounded in reality. Lots of people are working from home these days, and I suspect remote work will continue to climb now that companies are becoming more comfortable with that option.

I'm not saying I agree with him, but I'm on a 15Mb down 2Mb up cable modem, and everyone's experience got a lot better once I got an OpenWRT-capable router and turned on Smart Queue Management. I frequently have two Rokus streaming 1080p, two laptops on youtube and an old machine in the basement running torrents, plus my phones connected to a PBX hundreds of miles away, and no one complains that their videos hang or their phone calls stutter (that was the real problem before, poor-quality phone calls).

I don't honestly understand how it works, but it's amazing[0]. I'm okay with the packet loss in that graph, because I think that's how traffic control does its job?

[0]https://imgur.com/a/1vDAtyR

rokus, youtube, and phone calls use very little upload. Try doing a couple zoom calls on 2 mbps upload and you'll have issues. A single HD group zoom session requires 3.0 mbps upload according to zoom https://support.zoom.us/hc/en-us/articles/201362023-System-R...

Same here. I’m on 16/1.5 and I can run a couple of SD video streams outbound with an EdgeRouter and smart queueing. I would certainly love more upstream for uploading pictures, and uploading videos takes overnight, but we can have several people streaming down and at least a couple up.

I wonder if I'm the only one with better upload than download...

$ speedtest

   Speedtest by Ookla
     Server: Windstream - Chicago, IL (id = 17384)
        ISP: Webpass
    Latency:     0.93 ms   (0.10 ms jitter)
   Download:   847.27 Mbps (data used: 642.6 MB)                               
     Upload:   936.79 Mbps (data used: 445.7 MB)

Packet Loss: 0.0%

(maybe this is because of other users in my building...)

Most speedtests aren't designed to test very high speeds. It could just be that the server on the other side was congested.

640K ought to be enough for anybody. And with 3Mbps you can send or receive the contents of your entire RAM in under two seconds.

Lemme see... 3 divided by 8 is 0.375, or 375K... In two seconds that equals 750K of data.

Yep, the math checks out!

I'm going to be the contrarian here and agree with this definition. I think it's more important that the government prioritizes making access to 25/3 available to everyone before making 100/15 (or whatever the new standard is suppose to be) available to the people who already have 25/3.

The internet enabled the dream that we all can be broadcasters, not just big companies.

(In some countries other than the US)

Is this guy going to be replaced by someone new in the Biden administration? He's a disaster.

He was appointed by the Obama/Biden administration, so I'm quite ambivalent on the question. I feel like almost everyone in tech wants him to go but I may be just projecting my own bias.

Surprisingly enough, aside from his grand meme-filled display of repealing Net Neutrality articles, nothing really happened as a result of it.

Appointed by Obama/Biden under Mitch McConnells recommendation. I.e. Obama/Biden wouldn't have been able to get anyone else through with out Mitch McConnells (GOP Senate Leader) green light.

Yes. Was supposed to step down yesterday.

https://www.vox.com/recode/2020/11/30/21726645/ajit-pai-fcc-...

"Biden is most likely to select one of two people: either Edward "Smitty" Smith, a telecom attorney who has worked as a legal adviser to former FCC Chairman Tom Wheeler and at the National Telecommunications and Information Administration; or Jessica Rosenworcel, the most senior Democratic commissioner at the FCC." https://www.spglobal.com/marketintelligence/en/news-insights...

Reply on news.ycombinator.com