SoylentNews is people
SoylentNews
Jon Brodkin over at Ars Technica is reporting on a filing submitted to the FCC by Netflix last week asking thc FCC to "declare that home Internet data caps are unreasonable and that they limit customers' ability to watch online video."
From the article:
Netflix submitted a filing last week for the FCC's annual investigation of broadband deployment, a review that is mandated by Congress in Section 706 of the Telecommunications Act. Specifically, Congress requires the FCC to determine whether advanced telecommunications capability is being deployed to all Americans in a reasonable and timely fashion and "take immediate action" to accelerate deployment if it's not happening to the commission's satisfaction.
The commission's assessment generally focuses on availability and speed, but Netflix wants the commission to add data caps to the mix. "Data caps (especially low data caps) and usage-based pricing ('UBP') discourage a consumer's consumption of broadband, and may impede the ability of some households to watch Internet television in a manner and amount that they would like," Netflix wrote. "For this reason, the Commission should hold that data caps on fixed-line networks and low data caps on mobile networks may unreasonably limit Internet television viewing and are inconsistent with Section 706."
[...] Netflix argued that a 300GB-per-month allotment "is required just to meet the Internet television needs of an average American," without accounting for other things consumers want to do on the Internet, like Web browsing and downloading games and applications. "The Commission should recognize that data caps and UBP on fixed line networks are an unnecessary constraint on advanced telecommunications capability," Netflix said.
Comcast, the nation's largest home Internet provider, recently raised its caps from 300GB to 1TB, making it easier for customers to watch online video instead of Comcast's own cable TV service. But consumers' data needs are increasing quickly enough that "today's 'above-average' Internet consumer is tomorrow's average Internet consumer," Netflix said.
Data caps also aren't necessary for network management, Netflix argued. The online video provider pointed to a government survey from 2014 in which ISPs told regulators that congestion wasn't a problem on their networks. ISPs have alternatively described data caps "as a way to align consumers' use of the network with what they pay," Netflix said.
So what say you, Soylentils? Do data caps discriminate against online video providers?
Do data caps negatively impact other types of Internet usage?
Is online video the bulk of the data you consume through your Internet connection?
Do you have a data cap? If so, what is it and how often do you exceed it? If you do exceed it, what steps does your ISP take in response?
(Score: 2) by jmorris on Tuesday September 13 2016, @09:52AM
Cheap huh. Go price a gigabit link into a major peer with full commercial grade SLA, no use limits and and a contract that includes the ability to resale. Now do the math. Take a small town, figure up how many cable modem customers you are going to sign up, these days 50mbps seems to be entry level. How many end users you can service on that gigabit uplink becomes THE economic question. If a large percentage start hitting 30-40% average utilization you are screwed. It gets worse. Netflix tends to peak in the same prime time as home Internet usage so that 30-40% total usage (which would include dead hours only used by bittorrent/usenet users) doesn't reflect how many you can load on before you start getting complaints. You need a fatter pipe before you get your first crate of modems into the field. Question: Will you be pulling in enough from customers to pay your bandwidth bill? No, not at 1gbps. If you survive long enough to scale you will but even the big systems with massive economy of scale are starting to worry about the trend lines.
The Netflix in a box is the solution that has kept things viable this long. But Netflix isn't the only player. Generic CDNs like Akami have also done wonders. There are limits though. And no, you can't optimize Bittorrent because any attempt would void the whole "We don't know, we don't want to know; if you send a C&D we will do what we have to." aspect that so far has kept the ISPs out of legal problems. Note that I am saying that even more effort to attempt to create the illusion of unlimited bandwidth is the best shot of a new opportunity right now. But you can't simply expect bandwidth to actually become a free resource.
(Score: 2) by TheRaven on Tuesday September 13 2016, @03:57PM
The Netflix in a box is the solution that has kept things viable this long.
The OpenConnect appliance helps a lot with your problem of peak load. It turns out that the vast majority of traffic from Netflix (which, itself, is around a third of total US Internet traffic) is from a relatively small selection of shows. Most of these will be cached.
And no, you can't optimize Bittorrent because any attempt would void the whole "We don't know, we don't want to know; if you send a C&D we will do what we have to."
BitTorrent will prefer the seeds with the lowest ping times, so will keep traffic on network if it can. If ISPs gave more symmetrical links, then they'd likely see a decline in off-network traffic from BitTorrent.
sudo mod me up
(Score: 0) by Anonymous Coward on Wednesday September 14 2016, @08:30PM
I actually have priced it out, and am strongly considering it as our only available monopoly granted ISP is amazingly terrible
Pricing to have a fiber line run:
Having 100/100 fiber run to my address, unlimited bandwidth 24 IPs costs ~1400/mo
Having 1g/1g fiber run to my address, unlimited bandwidth, 256 IPs costs ~2200/mo
This is a guaranteed connection, with guaranteed speed, and a 2 9's reliability on the connection.
As a business; ignoring equipment and setup costs (which are non-negligible, but not relevant to this discussion) maintaining a GURANTEED 5/5 connection (which could easily be sold as 50/50, 5/5 would only come up in the extremely rare situation where ALL subscribers are using all balanced bandwidth simultaneously) to subscribers would cost:
$70/pers for 20 people
$11/pers for 200 people
This is for a private individual contracting to a fiber company for a small scale deployment. Something massive like a multi-national ISP (Bell/rogers/power utility) that has right of way on existing power lines and roadways is going to get there bandwidth and all the other benefits of being one of the monopoly providers of internet access I can't see how they are failing to provide unlimited bandwidth- what are they doing with the extra $40/person/month (assuming they can't get a better rate than what I can for 200 subscribers).
(Score: 2) by jmorris on Wednesday September 14 2016, @11:03PM
Uh huh.. You are right that you are in the 'looking at this' phase. Spin again. Call them back up and get their price with a proper 5 9's SLA and resale rights. Two nines is barely basic business class service. If all of your customers can expect to have three days of downtime per year, just from the upstream, before worrying about downtime because their link is busted, you won't have happy customers. You got a price on business grade service, which presumes activity will drop off right in time for 'prime time' as employees log out of the cloud to go home and fire up Netflix on their home connection.
Now assume you get carrier grade service at those prices anyway. You think you will service two hundred customers on it? Netflix recommends at least 5Mbps for HD quality. That is assuming ONE person is watching in the home and nobody else is doing anything. That is the whole point behind my observation, that people are now expecting to have every member of the household streaming at the same time on various devices. So figure HD on the main TV and a pair of SD streams plus some other random activity and you start needing to provision around 10Mbps per subscriber as the bottom end. Doubles what you were assuming as cost per sub and the trend is for ever higher usage. You better be able to do some serious caching to make up the difference and will still be running on the ragged edge of pissed off customers complaining about "Buffering..." messages.
So to review, you will be selling basic service for $30-40/mo and spending $11-20 servicing the cost of the upstream. The physical plant isn't free though, it will cost at least $10/mo/sub. Doesn't leave much left for running the business does it? Priced a support desk lately? Planning on paying taxes? You can get some lift on the per sub revenue by renting most of the idiots a cable modem for $10/mo that costs you $25/ea in bulk; that does help the dismal ISP economics a bit. With those first 200 customers you better clear at least $5/ea/mo or you can't even afford to hire yourself, even just part-time, forget funding expansion.