BitTorrent: Netflix should defeat ISPs by switching to peer-to-peer

"Decentralized technology" could bring equality back to the Web, CEO says.

Read the whole story

 

[type unknown]

Bengie25[/url]":zfuvn9gc]

KarolofNine[/url]":zfuvn9gc]1. ISP's begin throttling user connections based on bandwidth usage. (~2009, bandwidth caps for wired conenctions)
2. Heavy users begin paying for usage
3. Users reduce usage to cut costs
4. ISPs begin throttling commercial connections based on bandwidth usage. (~2014, Netflix v. Comcast, et al)
5. Companies begin paying for usage
6. Companies reduce usage to cut costs (By switching bandwidth load to users)
7. ISP's begin throttling user connections based on bandwidth usage. (~2019, predicted)

Depends on what you mean by "heavy" users. By 2019 1gb-10gb fiber should be the norm and 100gb-10tb uplinks for ISPs will be common. "Low usage" may be measured in tens of terabytes. Plenty of room for 4k-3d streaming.

While I don't disagree with you about technology moving forward, it would be a good question to ask 'where' these networks will be providing 1-10gb fiber connections.

This is where I think Facebook and Google are on the right path with wireless service via drones or balloons over target market areas of the world. It will help to eliminate part of the streaming issues (from a tech standpoint only, not worried about business model yet) by re-assigning the last mile connection to someone other than a well entrenched telecom.

Now just work out the business model (plans, rates...) and whether or not there needs to be any govt oversight (regulation) specifically for this type of last mile wireless service. At that point, we may see more competition from vendors and better choices for the average user when it comes to access.

The last mile thing really sucks. You are a hostage to whomever owns the lines in the ground and building.

 

[AngelZero]

The last mile thing really sucks. You are a hostage to whomever owns the lines in the ground and building.

You're pretty much held hostage regardless due to anti-compete agreements, monopoly leases, etc.

 

[Rookie_MIB]

I wonder if Netflix would engineer a Netflix box, capable of playing the video files and of course, containing some local storage for pre-downloading content. Then, it would just be a matter of (sometime in the future if they did decide to implement some P2P capability) a simple software update to provide that P2P capability.

Such an arrangement would be quite effective, and if it earned anyone with no caps who were willing to provide bandwidth for seeding a discount on services, I'm willing to bet people would be more than willing to help out since it would also help out with video quality.

There would have to be some sort of algorithm to handle distribution as to where the files are located to make sure there's sufficient bandwidth floating around, but since it's all a software issue, it should be pretty solvable. Heck, Netflix has 30 million users, while the biggest peer count on GoT was 190,000 or so. That's a pretty big difference and provides a metric sh*t-load of potential bandwidth.

Lets say of those 30 million users, 10% decide to sign up, and each one provides a maximum of 256Kbit of upload bandwidth (1/4 of an Mbit to avoid creating issues for the end users). That works out to 750Gbit of bandwidth - a pretty significant chunk.

 

[Solidstate89]

FeldmanSkitzoid[/url]":10zv79t0]It's a good idea, but...sigh...why can't I just pay by the megabyte?

I really hope my sarcasm detector is broken because your post sent it off the charts...and not the other reason.

 

[Maltz]

IMO, this whole thing makes a lot more sense if the throttling and bandwidth caps have more to do with keeping people from canceling their television subscriptions than anything to do with network congestion. Consider:

- ISP's started implementing caps when available bandwidth is higher than ever before. Why haven't such caps been in place for years?

- ISP's try to charge television competitors like Hulu and Netflix, but not other high-bandwidth things like Apple/Microsoft software updates or backup software, the latter of which hits the customer's upstream where there is even LESS available bandwidth.

- In the case of my ISP, at least, there is no option to pay more if you go over the cap or to upgrade to a cap-less plan (short of going to a MUCH more expensive business plan). You just get a notification letter, but I've never heard of anyone who was actually disconnected. If you're really serious about network congestion, you would throttle these people or charge them more to cover the costs of upgrading the network. But if you're trying to keep them as television customers, you just want to make them think they couldn't cut the cable without losing their internet, too. But you don't want to ACTUALLY cancel them, because when they switch internet providers, you would likely lose the TV revenue as well.

One question I have is whether non-content-provider ISPs are also implementing caps and trying to charge Netflix et al. I haven't heard of any that are doing that, but just about every ISP of any size these days grew out of a content provider, so I don't know.

 

[Bengie25]

type unknown[/url]":3ls5s9mv]
While I don't disagree with you about technology moving forward, it would be a good question to ask 'where' these networks will be providing 1-10gb fiber connections.

My city went from 1mb DSL 5 years ago to 1gb fiber now being offered. It's still expensive, but their prices have been coming down.

My brother lives in a town of 800, but most are farmers and spread over a wide area. He went from $100/month for 1mb/0.25mb DSL to 40/40 fiber with no price change in just the past 2 years.

In Minnesota, they're running 1gb fiber to log cabins in the woods and 100mb to the farms.

As far as I can tell, all the areas that don't have incumbents are getting 1gb fiber. It seems that small towns and cities are getting faster Internet because incumbents don't care about them, but the issue is word is getting out via family members that these small towns and cities are getting faster and cheaper Internet than the big cities.

Expectations will start to rise.

Not to mention local cities are starting to have fiber initiatives. Since incumbents are so set on not using fiber yet, the smaller ISPs are getting the job done and it's wreaking havoc on the incumbents, but they still don't care because they're still smaller cities.

edit: I forgot to mention. I've been following news for GPON roll outs and case studies. Even though the reported stories are only a fraction of the number of fiber deployments actually happening, there are about 10 such stories per year. When you look at it all together, you're talking about 300k+ people getting access to 1gb fiber per year, at least.

This has been going on for years now, we're talking about millions of users by now having access to 1gb fiber, even if not currently offered by the ISP. A sizable chunk of the USA is getting fiber, it's just not going strait to 1gb/1gb like Google Fiber, because most ISPs can't handle going from 1mb DSL to 1gb fiber in a single roll out. But the infrastructure is ready and waiting.

 

[Trem]

I haven't seen this mentiond yet but Voddler has offered this in Sweden since around 2010, it's evolved a bit but at first it was mainly for rental movies/episodes. Not sure if they're using BT but at least a similar p2p solution.

Some info around this at http://voddlertalk.voddler.com/om/

 

[rutebega]

jdale[/url]":2br5yki7]
I think the more difficult part of the problem is getting people to participate in the P2P network, since it will increase their own bandwidth usage. Perhaps Netflix could offer their service for free to anyone with a gigabit connection who agrees to peer...

What about service discounts based on bytes uploaded? A free month for every 50GiB, or whatever Netflix deemed reasonable?

 

[phoenix_rizzen]

Ostracus[/url]":2qzg9tci]The only problem I see with P2P aside from the start-up delays, is the out-of-order nature of the process. Something that conflicts with the in-order nature of streaming.

Popcorn Time users don't seem to have any issues with streaming-over-bittorrent.

 

[H2O Rip]

A legal bittorrent application that would reduce the ISP's ability to extort peering agreement fees? No way that'll get anywhere!
(said in sarcastic tone...but depressingly literal at the same time)

 

[curiosus]

Lets all mariachi![/url]":1x1ccfwj]

curiosus[/url]":1x1ccfwj]Dumping your bandwidth costs onto your users is not a good business model. Especially when those users have bandwidth caps from the terrible ISPs you are trying to avoid. The only way for this to end happily is for Netflix to fight, not pass the problem along.

That depends on your intent. If Netflix were to work on an alternate infrastructure behind the curtain, make no announcements, very quietly pass out client updates and then flip the switch at the beginning of a month 2 things would happen. One, a ton of users would notice a significant spike in video quality, making them very happy. Two, at the end of the month Comcast, not expecting the switch, would end up sending out exponentially more overage bills, triggering consumer outrage from a very large group of people all on the same billing date before Comcast even figures out why, immediately escalating something that most people don't know / care about into a high profile media story affecting "everyday Americans". That's how policy change happens.

Thats probably a really bad idea in practice, Comcast will just blame Netflix and its customers. Unless the "average" american understands the situation fully all they would see is Netflix making hidden changes that drastically increased their bills. If all that anger ends up on the ISPs side then you permanently killed neutrality.

 

[knbgnu]

jandrese[/url]":3jbarvfq]

Andrew Norton[/url]":3jbarvfq]

jandrese[/url]":3jbarvfq]

mikehild[/url]":3jbarvfq]
The reason this won't work is because way too many people use Wifi these days, and doing P2P over Wifi kills the signal for everything else.

Really! Why hadn't someone told me this before???!?!?!?!

Oh wait, I know, because it's not true

you see that link about comcast being caught? Rob Toplowski spotted it initially, I led the verification for TorrentFreak (which prompted the EFF to do their report several months later). Many of the test boxes used wifi.
P2P data is just data. it's not going to kill your wifi system any more than any other traffic. the only exception is if you have a (now) old router, with bad UDP handling, then assuming you use a lot of DHT and µTP, it can run out of ram. But that can happen even with it wired.

Try running an even moderately popular torrent over Wifi while people are browsing or gaming and they will notice, especially if you live in an apartment where several other people are running lots of data over their Wifi links. It is very easy to build up big queues on Wifi and choke off normal traffic.

It's probably not the Wifi that is saturated, since anything remotely modern is at least theoretically much faster than a consumer connection, but the fact that multiple devices are using it. It has nothing to do with p2p and nothing to do with Wifi. Wifi often has multiple devices sharing a connection and p2p is good at saturating a line, especially if there are not limits set. However, you can do the same thing with a wired router and a lot of fast xdcc bots.

 

[Bengie25]

rutebega[/url]":1ldn9h3w]

jdale[/url]":1ldn9h3w]
I think the more difficult part of the problem is getting people to participate in the P2P network, since it will increase their own bandwidth usage. Perhaps Netflix could offer their service for free to anyone with a gigabit connection who agrees to peer...

What about service discounts based on bytes uploaded? A free month for every 50GiB, or whatever Netflix deemed reasonable?

Netflix' average bandwidth custom per use is about $0.3/month of your $8 bill, and almost $4 for content licensing. All in all, Netflix currently pays about $0.01/GB. To break even on the $4 for licensing costs, you would need to upload about 400GB. If they actually wanted to make a worth-while profit, you'd need to upload closer to 1TB.

That ignores that their bandwidth costs closely follow whole-sale bandwidth costs, which are dropping about 50% per year. By next year, you'll need to upload 2TB, and that's assuming their licensing costs haven't gone up.

Actually that data that I read was a bit old already, so you may need to upload 2TB right now and 4TB next year.

 

[type unknown]

Bengie25[/url]":2jpwfhcp]

type unknown[/url]":2jpwfhcp]
While I don't disagree with you about technology moving forward, it would be a good question to ask 'where' these networks will be providing 1-10gb fiber connections.

My city went from 1mb DSL 5 years ago to 1gb fiber now being offered. It's still expensive, but their prices have been coming down.

My brother lives in a town of 800, but most are farmers and spread over a wide area. He went from $100/month for 1mb/0.25mb DSL to 40/40 fiber with no price change in just the past 2 years.

In Minnesota, they're running 1gb fiber to log cabins in the woods and 100mb to the farms.

As far as I can tell, all the areas that don't have incumbents are getting 1gb fiber. It seems that small towns and cities are getting faster Internet because incumbents don't care about them, but the issue is word is getting out via family members that these small towns and cities are getting faster and cheaper Internet than the big cities.

Expectations will start to rise.

Not to mention local cities are starting to have fiber initiatives. Since incumbents are so set on not using fiber yet, the smaller ISPs are getting the job done and it's wreaking havoc on the incumbents, but they still don't care because they're still smaller cities.

edit: I forgot to mention. I've been following news for GPON roll outs and case studies. Even though the reported stories are only a fraction of the number of fiber deployments actually happening, there are about 10 such stories per year. When you look at it all together, you're talking about 300k+ people getting access to 1gb fiber per year, at least.

This has been going on for years now, we're talking about millions of users by now having access to 1gb fiber, even if not currently offered by the ISP. A sizable chunk of the USA is getting fiber, it's just not going strait to 1gb/1gb like Google Fiber, because most ISPs can't handle going from 1mb DSL to 1gb fiber in a single roll out. But the infrastructure is ready and waiting.

So who owns these fiber lines? Since there are some community based initiatives, I would assume that the communities have a heavy hand in how service is implemented if they invested any money to help get the fiber there in the first place. Or are the smaller providers operating on a CLEC style model?

 

[AM16]

This would be great, except that ISPs already throttle p2p and they would use their traffic shaping tools to further kill p2p. S

With the kind of standards out there and the money the ISPs and cable companies are making, any reasonable person would think that our speeds and prices would match if not surpass the rest of the world.

If these ISPs never upgrade their infrastructure (they really have no incentive), I am really sure that we will see a huge decline in online related content purchases. Due to bandwidth and data caps.

I am more surprised that companies whose whole business model is based on online services aren't enraged and lobbying against it.

I'm staying with DVDs for a long, long time.

 

[knbgnu]

I would guess that Netflix probably considered this before they even had a streaming option available, but they nixed it because the MPAA has so many controlling Luddites.

As for concerns about caps, the answer to that is to actually hold ISPs responsible. Caps only allow ISPs to get away with not upgrading their infrastructure.

lee_machine[/url]":171s5eg5]A service like this should work really well for very popular streams but what about the unpopular ones? Netflix would still have to have fast dedicated hosting. That said, having a decentralized method of distributing content would easy the load coming directly from Netflix.

Yes, they would have to be seeders themselves, but it would drastically reduce the load and bandwidth coming from them, which would solve the Comcast problem.

Bengie25[/url]":171s5eg5]

KarolofNine[/url]":171s5eg5]1. ISP's begin throttling user connections based on bandwidth usage. (~2009, bandwidth caps for wired conenctions)
2. Heavy users begin paying for usage
3. Users reduce usage to cut costs
4. ISPs begin throttling commercial connections based on bandwidth usage. (~2014, Netflix v. Comcast, et al)
5. Companies begin paying for usage
6. Companies reduce usage to cut costs (By switching bandwidth load to users)
7. ISP's begin throttling user connections based on bandwidth usage. (~2019, predicted)

Depends on what you mean by "heavy" users. By 2019 1gb-10gb fiber should be the norm and 100gb-10tb uplinks for ISPs will be common. "Low usage" may be measured in tens of terabytes. Plenty of room for 4k-3d streaming.

Key word here is 'should be.' Those should be slow connections by now, but Comcast et al have avoided proper investment in infrastructure because they are in environments without meaningful competition.

 

[Bengie25]

type unknown[/url]":2mk8ib7o]

So who owns these fiber lines? Since there are some community based initiatives, I would assume that the communities have a heavy hand in how service is implemented if they invested any money to help get the fiber there in the first place. Or are the smaller providers operating on a CLEC style model?

The most common story type that I have read is a small single ISP in a small town with little red-tape. This is probably because less political controversy surrounding these kinds of cases without an incumbent trying to spread FUD with a smear campaign.

My personal experience has also includes some cities where Charter or Comcast are the local incumbent, but a smaller ISP rolls out fiber and you suddenly read about reports of lines of people standing outside of local offices to return TV and network equipment to Charter or Comcast because they can't compete with the faster speeds and lower prices of fiber.

I myself have found myself in a situation like that. I went to return my cable modem and had to wait in line for almost a hour because so many people wanted to drop Charter for the new fiber Internet. I finally got my turn and asked the desk lady if it's always this bad, and she said it was ever since the local ISP announced fiber. She said a lot of people were doing early cancellations of their contracts and didn't care about the fees.

 

[erendorn]

lee_machine[/url]":1b5db7n5]A service like this should work really well for very popular streams but what about the unpopular ones? Netflix would still have to have fast dedicated hosting. That said, having a decentralized method of distributing content would easy the load coming directly from Netflix.

Well, as long as Netflix servers themselves seed all their catalogue, this won't be an issue, as they would be able to seed as fast as they currently stream. So the p2p part is just a bonus for them and/or their users.
Possibly even for the internet providers, if the netflix servers are far from some swarms.

jdale[/url]":1b5db7n5]
I think the more difficult part of the problem is getting people to participate in the P2P network, since it will increase their own bandwidth usage. Perhaps Netflix could offer their service for free to anyone with a gigabit connection who agrees to peer...

Priority streaming, faster download, higher quality, free shows, exclusives, keeping the movie files, price reductions... There's many incentives available for an opt-in system.

 

[kvndoom]

AngelZero[/url]":2enhkmw2](note I do not actually work for an ISP)

From an ISP's perspective: block the protocol(s), throttle the number of connections, problem solved. Oh and guess what, you'll still have to buy our product! Suck it, plebes!

No amount of dancing around different ideas gets around the fact that ISPs control things. Either we fix that root problem, perhaps through nationalizing the ISPs, or we deal with the consequences such as the ongoing Net Neutrality stuff.

To paraphrase Jeff Goldblum, "BitTorrent... finds a way."

 

[Gasaraki]

Ostracus[/url]":28dzog7y]The only problem I see with P2P aside from the start-up delays, is the out-of-order nature of the process. Something that conflicts with the in-order nature of streaming.

QoS would fix this problem.

 

[Tundro Walker]

jdale[/url]":16toovg3]

Ostracus[/url]":16toovg3]The only problem I see with P2P aside from the start-up delays, is the out-of-order nature of the process. Something that conflicts with the in-order nature of streaming.

Streaming always requires buffering. If the stream is unreliable, you buffer more. This is a solvable problem, as long as you can accumulate enough total bandwidth.

[bold]I think the more difficult part of the problem is getting people to participate in the P2P network[/bold], since it will increase their own bandwidth usage. Perhaps Netflix could offer their service for free to anyone with a gigabit connection who agrees to peer...

I think that will solve itself.

If there's a popular show, and you join a torrent for it and you get it delivered faster than waiting for a stream ... it doesn't take long for a good thing to catch on.

The main issue is that there will be jackasses that want to be ultra-stingy, just like in other torrenting situaitons. That one jerk that wants to torrent down, but tweaks his system to not seed (either via a file, or compiling some open-source torrenting prog himself that making it nothing but a leecher).

As KarolofNine pointed out above, though, ISP's will squash this... all they'll do is start a) suing customers as "content providers / uploaders" that are violating their subscription TOS, or b) throttling upload speeds, or c) finding this as a new revenue stream to cherry-pick by selling folks "upgraded torrenting packages!" that just have unthrottle upload speeds.

The ISP's control everything, so no matter what happens the ISP's will do something about it to hold everyone hostage. Trying to change things just gives them another avenue to charge more for a "service" that folks are using to get around the clamps they're putting in place.

"Oh, you want to torrent netflix! Ok, we'll start finding ways to charge you extra for that, too! Congratulations on us being so kind as to notice your 'needs' and 'work with you' on them."

The FCC has had their chance, and they're not only bungling it big-time, but consumers think they're nothing but a corporate puppet joke. The FTC needs to step in on all of this and start making some real decisions.

 

[cf18]

It wouldn't work for many devices that support Netflix but have very little to no storage e.g. smart TV, Wii. Without any space to cache content they would be unable to contribute to the P2P swarm. Some older devices like WD TV may not enough CPU and RAM to do both video playback and encrypt P2P traffic at the same time.

 

[Gasaraki]

jdale[/url]":a8v2i544]

Ostracus[/url]":a8v2i544]The only problem I see with P2P aside from the start-up delays, is the out-of-order nature of the process. Something that conflicts with the in-order nature of streaming.

Streaming always requires buffering. If the stream is unreliable, you buffer more. This is a solvable problem, as long as you can accumulate enough total bandwidth.

I think the more difficult part of the problem is getting people to participate in the P2P network, since it will increase their own bandwidth usage. Perhaps Netflix could offer their service for free to anyone with a gigabit connection who agrees to peer...

It could be an automated thing. Anyone with a Netflix client running at the time is also uploading just like how Bittorrent works now.

 

[anonArs]

Ostracus[/url]":1r34r4jv]The only problem I see with P2P aside from the start-up delays, is the out-of-order nature of the process. Something that conflicts with the in-order nature of streaming.

The central repository of content will still be needed since not everyone will be streaming everything at the same time. So the central repository can always be used to grab parts of the video that aren't going to make it in time via P2P.

 

[Gasaraki]

jonsmirl[/url]":1bph9n71]Use a blend of the two technologies. Start off streaming straight from Netflix. Simultaneously use P2P to start asking for blocks a minute or so further into the stream. As you start collecting blocks send word up to the Netflix servers that you don't need the blocks any more. First step would be for Netflix to implement off-line caching.

Neat thing about a scheme like this is that if the P2P piece disappears you are just back to the old system. Also, given the hit driven nature of movie viewing, you might see 90% of Netflix's transit traffic disappear since P2P would satisfy everything internally to the ISP's network for things like the House of Cards initial binge.

Of course P2P servers on mobile are bad. That is why Spotify is getting rid of them on mobile.

This is how Blizzard does it. They have a master seeder that starts the upload and as you download a patch, other peers join in to seed and you start using more of those instead of the master. If Blizzard can do it with patches, I don't see why if wouldn't work for videos also.

Netflix can create seed servers all over the country so even if no one else is watching that movie or show, there would be enough Netflix owned seeders to seed that movie out.

 

[haneefmubarak]

Um... I think YTS already has this shit on lock with Popcorn Time.

 

[glugglug]

Andrew Norton[/url]":174e71if]

jandrese[/url]":174e71if]

mikehild[/url]":174e71if]
The reason this won't work is because way too many people use Wifi these days, and doing P2P over Wifi kills the signal for everything else.

Really! Why hadn't someone told me this before???!?!?!?!

Oh wait, I know, because it's not true

Actually he is sort of right, but for an indirect reason, and not because its P2P persay.

Wifi works like a hub, while wired ethernet for the past 2 decades at least is typically switched.
On a hub you have collisions if multiple users are simultaneously transmitting, and they both have to retransmit and things degrade badly. So it only works well when the bulk of the sending is being done by one device. Currently that works out because 99% of users just download, not upload, so the access point is doing the sending to various wifi devices, with the individual devices just periodically sending tiny pieces of data such as acks or HTTP request headers. Since only one device (the router) has non-negligible transmit time, there are few collisions. P2P while watching a video the router would be sending you a piece of the video to watch while your playback device is sending back a previous piece of the video to share. Collisions all over the place. Expect less than 1/3 your already shitty WiFi performance.

 

[compuguy]

Zak[/url]":2frl0e1l]Another problem I see is that the BT protocol is blocked by many institutions. If you're sitting in the lab all night watching an experiment and try to kill the time with Netflix, you'll be out of luck.

Or even worse, sysadmins may think your computer is infected with a virus. Seriously that is what happened to me when running a bitcoin client at work....

 

[Deleted member 192806]

Bengie25[/url]":194kbld3]

rutebega[/url]":194kbld3]

jdale[/url]":194kbld3]
I think the more difficult part of the problem is getting people to participate in the P2P network, since it will increase their own bandwidth usage. Perhaps Netflix could offer their service for free to anyone with a gigabit connection who agrees to peer...

What about service discounts based on bytes uploaded? A free month for every 50GiB, or whatever Netflix deemed reasonable?

Netflix' average bandwidth custom per use is about $0.3/month of your $8 bill, and almost $4 for content licensing. All in all, Netflix currently pays about $0.01/GB. To break even on the $4 for licensing costs, you would need to upload about 400GB. If they actually wanted to make a worth-while profit, you'd need to upload closer to 1TB.

That ignores that their bandwidth costs closely follow whole-sale bandwidth costs, which are dropping about 50% per year. By next year, you'll need to upload 2TB, and that's assuming their licensing costs haven't gone up.

Actually that data that I read was a bit old already, so you may need to upload 2TB right now and 4TB next year.

Sounds like...Crashplan.

 

[svgpithon]

If they have the cash, maybe they could talk to Google about expanding Fiber.

 

[Deleted member 192806]

ldillon[/url]":loiibjk6]

lee_machine[/url]":loiibjk6]A service like this should work really well for very popular streams but what about the unpopular ones? Netflix would still have to have fast dedicated hosting. That said, having a decentralized method of distributing content would easy the load coming directly from Netflix.

Netflix has an army of servers on different backbones. When I have slow Netflix, it's usually because one of the interconnects is overloaded. P2P would move some of the load in-network, reducing the down-stream load on the interconnects and distributing the load to all available paths. Less popular titles should benefit from the lower connection at the interconnects, even if all of their content comes through them.

If this encourages more symmetrical Internet connections, that's another win.

I think it's not so much the P2P load balancing, as it is the P2P "whack a mole" sources at different addresses that the article's getting at.

 

[SockPuppetGovernment]

I P2P all the hi-rez HD multi-channel content that I want (and that is a LOT) safely through a $6-a-month privacy-proven VPN. It's not for "research purposes", it's illegal-as-fuck-piracy and frankly, I don't give a goddamn. I buy blue rays when I know for certain it's a movie I'll like and want to keep, and most of what I dl is TV, that I already pay for with a $100+-a-month DTV satellite bill and I have Amazon-Prime as well. It's just easier and more convenient to P2P than anything else, I get what I want, the way I want it - no FBI/interpol warning I can't FF through, no DRM at all, no commercials, practically no waiting, as much cheap storage as I can buy (just got a 4Tb drive for $150, or truecrypt zips in the cloud, whatevs!), foreign shows NOW, watch on any device (usually LAN to xbox) etc. etc. etc. The reasons for keeping the satellite (with it's very limited DVR storage) are getting fewer and fewer, and I'm sure I'll STILL sleep like a baby without it.
These politicians and rules and greed will ruin the net - if we let them - for those not tech-savvy enough to find an immediate work-around for anything they throw at it. P2P for netflix is a fantastic enough idea to get me to buy the service if they do it, but I'm not signing up for something that is about to be fucked-up by the fuck-ups. I feel sorry for the millennials, 'cuz a sick FB profile is not what I mean by "tech-savvy".

 

[Major_Kusanagi]

I can't help it, 'Colonel Klinker' comes to mind when I read his name.

 

[Newmusicmark]

Spotify already did that and now they got so big they turned it off (https://torrentfreak.com/spotify-starts ... rk-140416/) so if it doesn't work for Spotify, I doubt it would work for Netflix. They even had the creator of uTorrent working with them when they started so I'm sure this CEO doesn't know what he's talking about.

 

[cwbecker]

Newmusicmark[/url]":33oumx1p]Spotify already did that and now they got so big they turned it off (https://torrentfreak.com/spotify-starts ... rk-140416/) so if it doesn't work for Spotify, I doubt it would work for Netflix. They even had the creator of uTorrent working with them when they started so I'm sure this CEO doesn't know what he's talking about.

You do realize that there's an order of magnitude more data in an HD video stream than in an audio stream, right?

 

[wagnerrp]

nijave[/url]":6dk2e578]Of course a lot of consumer broadband packages have a relatively low upload bandwidth. We have 16/1Mbls from Time Warner that's unmetered and unthrottled. Of course, about 70% of that 1Mbps is constantly consumed by online backups. At 100% upload utilization, our latency takes a massive hit. It's my understanding it has something to do with how cable works and the fact it wasn't originally designed for 2 direction data transfer.

Cable gets ~40Mbps/channel downstream, and ~30Mbps/channel upstream, but it's really just a question of how many 6MHz channels out of the ~900MHz spectrum are allocated for upstream, downstream, and television.

 

[HonorableSoul]

Andrew Norton[/url]":4hisyw2m]

jandrese[/url]":4hisyw2m]

mikehild[/url]":4hisyw2m]
The reason this won't work is because way too many people use Wifi these days, and doing P2P over Wifi kills the signal for everything else.

Really! Why hadn't someone told me this before???!?!?!?!

Oh wait, I know, because it's not true

you see that link about comcast being caught? Rob Toplowski spotted it initially, I led the verification for TorrentFreak (which prompted the EFF to do their report several months later). Many of the test boxes used wifi.
P2P data is just data. it's not going to kill your wifi system any more than any other traffic. the only exception is if you have a (now) old router, with bad UDP handling, then assuming you use a lot of DHT and µTP, it can run out of ram. But that can happen even with it wired.

Since you so eloquently stated that the whole exception is an old router, How many routers on the internet are old? All you have to do is look at the decline in wireless router market statistics for the last 5 years to understand most folks have at least a 5 year old router which have the same problems with P2P, because honestly most non-IT savy folks are not going to be upgrading their modem. So most folks aren't upgraded with the latest router technology.

The original posters comments on P2P traffic hosing up a home nework are valid for a >50% population of the internet.

 

[Bengie25]

HonorableSoul[/url]":1qq5imqa]

Andrew Norton[/url]":1qq5imqa]

jandrese[/url]":1qq5imqa]

mikehild[/url]":1qq5imqa]
The reason this won't work is because way too many people use Wifi these days, and doing P2P over Wifi kills the signal for everything else.

Really! Why hadn't someone told me this before???!?!?!?!

Oh wait, I know, because it's not true

you see that link about comcast being caught? Rob Toplowski spotted it initially, I led the verification for TorrentFreak (which prompted the EFF to do their report several months later). Many of the test boxes used wifi.
P2P data is just data. it's not going to kill your wifi system any more than any other traffic. the only exception is if you have a (now) old router, with bad UDP handling, then assuming you use a lot of DHT and µTP, it can run out of ram. But that can happen even with it wired.

Since you so eloquently stated that the whole exception is an old router, How many routers on the internet are old? All you have to do is look at the decline in wireless router market statistics for the last 5 years to understand most folks have at least a 5 year old router which have the same problems with P2P, because honestly most non-IT savy folks are not going to be upgrading their modem. So most folks aren't upgraded with the latest router technology.

The original posters comments on P2P traffic hosing up a home nework are valid for a >50% population of the internet.

Even on my $160 Netgear 3700 would hose up on BitTorrent. I had to turn the connections down, otherwise it would start giving me packet-loss.

 

[wagnerrp]

Tundro Walker[/url]":1vnhe9x6]The main issue is that there will be jackasses that want to be ultra-stingy, just like in other torrenting situaitons. That one jerk that wants to torrent down, but tweaks his system to not seed (either via a file, or compiling some open-source torrenting prog himself that making it nothing but a leecher).

The main issue is that torrenting, as with all P2P traffic, is inherently extremely inefficient. The internet is not designed to see a lot of upstream traffic, or even a lot of traffic in general, from consumer end nodes. Managing a swarm means a whole lot more computational usage, and a whole lot more throughput as data unnecessarily travels over many extra hops to get from end node to end node.

If you want to more efficiently use bandwidth, use your buffering to allow the use of multicast, and get ISPs to properly implement multicast. Run popular media out over multicast channels, segmented and repeating. Have your client pull and cache segments for the content you want to watch, and backfill any missing data using unicast.

 

[nmonk]

nijave[/url]":7fhyoqs6]Of course a lot of consumer broadband packages have a relatively low upload bandwidth. We have 16/1Mbls from Time Warner that's unmetered and unthrottled. Of course, about 70% of that 1Mbps is constantly consumed by online backups. At 100% upload utilization, our latency takes a massive hit. It's my understanding it has something to do with how cable works and the fact it wasn't originally designed for 2 direction data transfer.

Because they wanted to be able to charge customers more for uploading, and peers more as "we're not sending data to YOU".