BitTorrent Turns 10

ktetch-pirate writes "On this day, 10 years ago, Bram Cohen released the first bittorrent client to the public. Most P2P protocols have had a rapid rise and then a drop-off as the subsequent 'best thing' has come out, but after 10 years, nothing has bested bittorrent, and it still remains king of the P2P castle. Just when will it be replaced?"

Comments

Pretty much never?

[Haedrian]

"Just when will it be replaced?"

Never? Going distributed is THE way of stopping people from shutting you down. So far only the tracker is fixed (and there are stuff in place to discover clients by seeing the others who you're connected to). So I'd say this is here to stay.

Re:Pretty much never?

[mrogers]

Going distributed is THE way of stopping people from shutting you down.

But ironically, what BitTorrent got right (and it pains me to admit this, because I'm a big fan of pure P2P solutions) was centralising the hard parts - search and peer location - and distributing the easy part - content distribution.

Another area where BitTorrent struck the right balance between pure P2P and pure centralisation was in content curation. Gnutella made it incredibly easy to share a file, but the result was a ton of low-quality, badly-labelled, nearly-identical files. BitTorrent made it just hard enough that only a few, relatively dedicated people would create torrents, and everyone else would just redistribute them. I don't think that was a conscious design decision, but it happened to hit the sweet spot.

Re:Pretty much never?

[wvmarle]

Nowadays there is such thing as "trackerless torrents". No idea how it works, but it works. A while ago I tried to download some torrent, but the single listed tracker in the .torrent was down. Nevertheless soon the download started, like magic :-) And once one peer was found, many more followed quickly thanks to peer exchange.

BT has in a way been replaced several times already. The protocol from 10 years ago has evolved a lot (will the original BT client even be able to handle to current torrents?), with additions like peer exchange, DHT, magnet links and encryption. The idea behind BT is brilliant and simple, and as such will always live on. It solved most if not all problems from the original P2P protocols: the P2P issue itself (too many downloaders on a single peer), disappearing peers (now you have more than one - download will continue from other peers), and overall download speed. The protocol was found to have some problems itself, most notably the centralised tracker, which is also solved now. The problems that remain are the finding of content, for that there is still no solution to the current centralised databases (aka "torrent sites"), and longevity of the content: as soon as the last seeder stops seeding, the file is lost from the network.

And on top of it, it's not owned by a single for-profit organisation like Napster or LimeWire. When that company goes out of business, the protocol is out, and something new is needed. BT will live long I think. It's an open and free protocol, allowing for it to evolve and have people add features to it. There is no "single point of failure" - by design.

Re:Pretty much never?

[klapaucjusz]

Nowadays there is such thing as "trackerless torrents". No idea how it works, but it works.

It uses a technique known as a Kademlia Distributed Hash Table (DHT). It's a rather tricky algorithm, which turns out to work beautifully for this particular application.

--jch

Re:Pretty much never?

[Lanteran]

Comcast, why am I not surprised?

Re:Pretty much never?

[CharlyFoxtrot]

We're in the silver age of music piracy. The golden age was Napster: everyone had their mp3's in folders instead of managed by applications like iTunes and everything was shared by everyone by default. You could find the most obscure songs. To me it was like a preview of what the internet always promised: a huge library where you could access any data (in this case music) that was out there. A little glimpse of the internet's true potential.

Re:Pretty much never?

[The Dawn Of Time]

Shocking, Comcast follow the law and serve the notices. They should be on the side of the little guy who just needs to watch Game of Thrones for free or he might be bored! The ignominy!

Re:Pretty much never?

[Kjella]

I think the legal balance is just as important. The fact that "BitTorrent" isn't one service or one network but that each site offer their files individually and the Ubuntu torrent has nothing to do with those on The Pirate Bay. That detached the technology and those building the tools from the shadier uses of it. Oh, they've rattled their sables a bit but never really had an legal grounds to shut BitTorrent itself down, unlike Napster, Grokster, Limewire and so many others.

And despite the best efforts to shut down torrent sites, many of them still operate very much in the open. The fact that The Pirate Bay has been all over the media and is in the top 100 most popular sites on the web means they've walked a very fine line and come down on the right side - at least for now. You didn't have to look that very hard in the past either to find it, but it was not that obvious to everyone and their dog.

I think something like TPB model is there to stay, if necessary they'll just move it to be a TOR onion site, still centralized but anonymous. Not the content itself as that'd be sloooow, just the site itself. For the moment that is simply not necessary, but there's now other ways should the public torrent sites lose while still keeping the things that made it a success.

Re:Pretty much never?

[rrohbeck]

The problems that remain are the finding of content, for that there is still no solution to the current centralised databases (aka "torrent sites")

That works beautifully with Kad for ed2k so I've always wondered why nobody has implemented that on top of the DHT in Bittorrent.

and longevity of the content: as soon as the last seeder stops seeding, the file is lost from the network.

Well every block of the file has to be on a hard drive *somewhere*. If people don't keep it around then the bits are gone.
I like Freenet's approach: Every user donates a chunk of disk (like 2GB) for anonymous encrypted storage.

Re:Pretty much never?

[Tacvek]

Combining multicast support which is certainly be bettter and more widely supported than it's today is. The reason is looming and it's HD-IPTV. Unicast HD-IPTV is a bandwith consumptive hog and the only practical way is to use multicast for it. That forces many organisations which did not previously bother configuring it reconsider and this will open doors to other applications using it too.

I strongly doubt that. Consider that AT&T uses multicast for their U-Verse service. The problem is that they provide no way to create your own multicast streams. They have no interest in allowing other applications to use it. Why would they? The multicast gives their service a serious cost advantage compared to other live streaming services. Because it is generally viewed as a separate TV service, most proposed net-neutrality regulations would not require them to open it up.

To the best of my knowledge there is no real standards for negotiating multi-cast between autonomous systems. Even if there were, the fact that the packets can multiply inside a network (when they reach a router with subscribers on more than one of the connected (sub)networks) makes setting up peering agreements difficult.

Transit agreements between ISPs generally assume that one packet sent in results in one packet leaving the network, or being delivered to a machine inside the network. With multicast between distinct autonomous systems, that one packet in could result in 100 packets being delivered to in network machines, and potentially one packet to each connected network. How should that be counted for billing?

If each resulting packet is counted, that would require network changes to track how many in-network machines it was delivered to. If it were counted as just the one packet for this network, plus one more packet for each connected peer or transit provider it reached, the required changes would be much smaller.

However in either case that would be really unfair to the sender of the packet, who has no way of knowing how many times an individual packet would be counted. It would also be rough on intermediate networks who may try to track usage and switch packets between multiple transit providers such as to minimize costs. They would only know that they sent one packet of some specific size to a given transit provider who may count it as many more packets.

The whole thing is very awkward. After all despite those issues, ISPs could benefit from the same number of users with the same style of usage resulting in many fewer total packets needing to be sent through the oversubscribed links from the core networks to last mile.

Re:Pretty much never?

[Hal_Porter]

One problem with bittorrent is that it's too easy to find peer IP addresses.

Let's suppose the RIAA connect to the tracker. They can get a list of IP addresses and then contact the ISP to convert those into names and addresses to be sent a subpoena.

Now one way to get around this is to use a VPN. But most VPN providers are not able to provide sufficient bandwidth to get your warez.

Actually imagine the following scheme.

Everyone agrees that they will provide a VPN service when torrenting. Now a few people connect to the tracker and start downloading. They advertise their willingness to act as VPN servers to the tracker. People connect to them and also advertise their willingness to act as a VPN server. At some point the people connected directly can shuffle to the back of the queue, i.e. connect via one or more VPN servers.

If correctly coded it seems like it would almost impossible to track a downloader downloading via VPN to a real IP address. At most could be tracked to one of the VPN servers which would discard their data when they stopped.

Now It's not foolproof - some people need to connect directly to the tracker and they would be traceable.

You could imagine that the people in piracy friendly jurisdictions (e.g. Scandindavia) would volunteer to connect directly to the tracker and tell it they'd act as VPN servers. People in piracy unfriendly jurisdictions (e.g. US/UK) would opt to only connect via their VPN service. The only way to stop this would be to block VPN completely. Even China doesn't do that because big companies love VPN segments to connect their intranets over the internet.

Incidentally that's another happy side effect of my scheme - people in repressive places can easily find a VPN server. Maybe trackers could be set up solely for people willing to act as a VPN server, unconnected to torrents.

In many ways it gives you a completely anonymous internet - basically TOR but mainstreamed.

Re:Pretty much never?

[mandelbr0t]

The problems that remain are the finding of content, for that there is still no solution to the current centralised databases (aka "torrent sites"), and longevity of the content: as soon as the last seeder stops seeding, the file is lost from the network.

I don't see these as problems. The torrent sites can be made anonymous, and have multiple mirrors. If an entire organization is somehow taken down, a new one will pop up very quickly. The fact that you can't search a tracker or a peer for content is a good thing. And, as to longevity, it means that as long as content is popular, there will be a seeder. It's reasonably safe to say, with the popularity of BitTorrent itself, that if a seed doesn't exist, the content just isn't popular.

Re:Pretty much never?

[blueg3]

Sort of. Bittorrent doesn't actually address search at all. Peer identification was centralized initially, which is incredibly efficient. Modern clients use out-of-band methods like DHT and PEX to either get additional clients (same torrent, but on a tracker not listed in the file you have) or to get clients when all available trackers are down (or not listing that torrent). They're much less efficient than a tracker.

But yeah, search isn't part of BitTorrent at all. It works well now that you can sort of run torrent-finding websites without being shut down. Back in the Gnutella days, you needed to provide search in order to have any hope.

Re:Pretty much never?

[hairyfeet]

Actually according to a friend in the state crime lab there could be a serious problem with the Freenet approach. The way it was explained to me was like this: You carry a box from point A to B. you don't have the key to the box, you don't attempt to open the box. the cops pull you over and find there are drugs in the box you are still going to jail whether you had knowledge of the contents or not.

You see the way the current CP laws are set up all they have to do is have a cop download a CP file that is traced back to your IP. It doesn't matter if Freenet encryption keeps you from seeing the file because they are not busting you for possession but for distribution and as I said above you don't have to look inside the box to distribute it.

Is it right? No but personally I think the whole witch hunt, where we are busting people for words on a page or cartoons is total bullshit. But considering in most areas you are looking at as much as 20 years an image for distribution I'd think twice before risking being the one who gets to test their whole "plausable deniability" bit in court.

As for TFA happy Bday BT! Sadly I believe that your best days are behind you, at least in the USA, as the *.A.As are gonna ram through three strikes laws where THEY are judge and jury, and the ISPs can't wait to give us ever shittier caps so they can take those massive profits and buy some more hookers and blow instead of line upgrades. So while I'm sure you'll have a future maybe sending WoW or Steam updates I kinda doubt when the ISPs get done and the *.AAs help them along there will be much bandwidth or anything else left to worry about using BT.

The corps aren't gonna rest until they turn the net into the home shopping network, and after Wikileaks and the Arab spring governments will be happy to help them in their goals. Hell I'm shocked they haven't gotten Nancy Grace and the other talking heads to feed the public some shit about how BT and other P2P protocols are used by nothing but terrorists and kiddie fiddlers. So good luck BT, frankly you are gonna need it.

Re:Pretty much never?

[SoupIsGoodFood_42]

The way things are with iTunes and iOS, along with iCloud coming up, seems more like the silver age to me. Golden would be if Apple continued successfully while also becoming more open source/IP-free or Linux made a finished product as good as Apple. Napster was the bronze age.

Re:Pretty much never?

[Kjella]

It's not really all that tricky. Instead of a hash, let's just say we assign it a date like April 3rd. You know some peers by birthday, and each peer pays extra attention to people with a birthday like their own. So you just ask the closest you know, they'll know someone closer and so your search is passed until it finds people in the right month, week and finally day. You could easily simulate this by placing people in a big circle by birthday, each get a few long connections and many short connections. That way everyone knows how to route the message without a central node. It's a little bit more complicated than that but not much really.

Re:Pretty much never?

[Haedrian]

Golden would be if Apple continued successfully while also becoming more open source/IP-free

Never going to happen. Apple is all about control. Tight control over what you can do, tight control over what you can use with their items.

Re:Pretty much never?

[Lanteran]

Yeah yeah, I'm not saying piracy is right (though for the purposes of screwing the bastards at the RIAA/MPAA out of money, it is), but comcast is one of the ISPs more notorious for monitoring and shaping traffic, and though they all do it to some degree, comcast goes "above and beyond".

Re:Pretty much never?

[froggymana]

Well every block of the file has to be on a hard drive *somewhere*. If people don't keep it around then the bits are gone.

Since the .torrent has the hashes for the data in the torrent couldn't it be possible (not easy, but still possible) to brute force a few remaining bits of a torrent?

Re:Pretty much never?

[jones_supa]

RapidShare links are my current choice.

Re:Pretty much never?

[Pentium100]

And, as to longevity, it means that as long as content is popular, there will be a seeder.

This is why I still sometimes use eMule to download some unpopular TV show or song. It may take a while, but it usually can be downloaded, since with eMule you share all your downloaded files by default, while you have to specifically keep the torrent seeding, even though you do not move or delete the file itself.

Re:Pretty much never?

[Pentium100]

Or you can use TOR. The problem is bandwidth. For now, individual users are rarely prosecuted, so going to an anonymous network would only reduce the bandwidth, since now peers will also be used as relays. Since they are also downloading/seeding the same file, you can just ask them to send you whatever they have (and not fetch a piece they are not interested in to just send it to you) to have full speed but you can do that now, without any modification to the BitTorrent protocol.

Basically, connect to the tracker using a proxy, get peer IP addresses then connect to them directly. The peers don't care that the tracker thinks that you have a different IP. But you could also be connecting to a RIAA "peer".

But what if the RIAA advertises its computers as VPN servers (in your system)?

Re:Pretty much never?

[mrogers]

I think something like TPB model is there to stay, if necessary they'll just move it to be a TOR onion site, still centralized but anonymous.

On that point, it's interesting to see clients like MediaGet and Frostwire 5 incorporating search into the client. If one of the sites they rely on gets shut down, not only could the clients switch to another site at the next upgrade, they could potentially switch to another way of contacting the site (eg through Tor, as you suggested) without the users needing to be any the wiser.

Re:Pretty much never?

[OrangeTide]

if it takes the cops 3-24 months to crack a freenet packet, lacking the key you have plausible deniability about the contents of the packet. the crimelab basically made your case for you.

now if you want to avoid the tedious process of getting dragged through court, I would recommend against using freenet or any other anonymous network or even any p2p network.

Re:Pretty much never?

[Goaway]

Except for the part where they fought hard to get rid of DRM, so that you could use the music you bought from them more freely.

You're letting internet fanboyism do your thinking for you.

Re:Pretty much never?

[klapaucjusz]

F[ine] magnet links, how do they work?

The magnet link contains a hash of the .torrent file, which happens to be the key to search for in the DHT (it's in the btid field of the magnet link). The local peer consults the DHT, and finds some peers that participate in the desired swarm; it then downloads the .torrent file from those peers. After checking that the torrent matches the expected hash, the local peer just does normal trackerless operation.

-- jch

Re:Pretty much never?

[AmiMoJo]

That was my first thought. Editors, can you please learn some basic editing skillz and do you fucking jobs? More often than not the submissions are laden with weasel worlds or flat out misleading summaries. Corrections are rare. I know it requires a tiny bit more effort to check the story but, well, that is what you are being paid for.

The only exception seems to be Taco who usually adds a little bit of balancing explanatory text and maybe straightens out the summary. Learn from him.

Re:Pretty much never?

[AmiMoJo]

The interconnect providers generally charge for a given bandwidth rather than for the amount of data sent. Many ISPs also buy bandwidth that way now because that is the way telcos like to charge for ADSL. BT sells 144Mb pipes and you can stick as many customers as you like on there. Both ends are rate limited so if your multicast packet generated 100 packets going out of the network then it would also use 100x the bandwidth from your allowance.

That is why ISPs will never allow it. When they have 50:1 or worse contention ratios (I hear some of the shittyest services are 500:1 on ADSL) a single user could saturate their pipe with multicast.

What they need is a P2P caching system. One was invented around 2000 but I don't know what happened to it. It would be legal, at least in the UK, as the law specifically states that ISPs are not liable for cached data. If they were none of them would run a web proxy cache. The problem now is that since ISPs started trying to throttle P2P traffic BitTorrent started encrypting packets, which makes them uncachable. They screwed themselves.

Re:Pretty much never?

[Hal_Porter]

Or you can use TOR. The problem is bandwidth.

The people that run TOR discourage torrenting for just that reason. The bandwidth problem is the root of that - bittorrent has a way of sucking up available bandwidth and they want to use the bandwidth for their own nefarious purposes. Given that a swarm of biittorrent clients has a very high aggregate bandwidth why not use it to add anonymity to the torrenters?

But what if the RIAA advertises its computers as VPN servers (in your system)?

Well it's not as bad as you think. The idea is that you stack VPN servers. So each VPN server can't really tell what it's forwarding.

Re:Pretty much never?

[Pentium100]

But the anonymity will only matter wen you connect to the tracker. As I said, since the peers are downloading/seeding the same file, you will get no speed boost by using a peer as a VPN server (to get data from another peer) instead of downloading whatever he has. And that can be achieved today.

Also, while the combined bandwidth is quite high, a lot of users do not have good upload speed - I had 768kbps until about a year ago. If one of those is in the chain, you will get limited speed too.

Re:Pretty much never?

[westlake]

I think something like TPB model is there to stay, if necessary they'll just move it to be a TOR onion site, still centralized but anonymous. Not the content itself as that'd be sloooow, just the site itself.

That would cut deep into its traffic.

Too complex. Too geek. Planet Peer - The anonymous networking community

There are many people who shy away from things like Freenet because they believe they are toxic. Something that might bring the Special Victims Unit to their door.

The Pirate Bay is as deep into the shadows as most are willing to go.

Re:Pretty much never?

[cjb658]

eMule still works. I am finding more and more garbage these days when I use its search feature for video and programs, but for music, it works pretty well.

There are also sites like TV Underground and ShareReactor that provide good ed2k links.

Re:Pretty much never?

[Hal_Porter]

But the anonymity will only matter wen you connect to the tracker.

Not exactly. Suppose you connect to the tracker and act as a VPN server. Suppose the only people that connect to you are also VPN servers with a few peers at the edge of the cloud. Suppose everything is set up so the routing data is in RAM and will be binned once the torrent is done. Also none of the VPN servers in the middle of the cloud can decode what they're passing on.

Now at that point it's extremely hard for anyone to pin any piracy on anyone. It's even hard for the peers to tell what's really going on - once you go through two more VPN servers it's probably impossible.

you will get no speed boost by using a peer as a VPN server

Also, while the combined bandwidth is quite high, a lot of users do not have good upload speed - I had 768kbps until about a year ago.

Well you might compared to using a VPN service. All the ones I've used are painfully slow. Slower than my ADSL upload rate in fact, let alone my ADSL download rate.

If one of those is in the chain, you will get limited speed too.

Well you can ameliorate this a bit by connecting to a few servers and dropping the slowest one periodically. Rather like bitorrent already works in fact - I'm just adding VPN in to obfuscate things.

Re:Pretty much never?

[WNight]

I heard they're chopping what they pay you astroturfers which I guess explains why you're just phoning it in these days.

Will you shill for any authoritarian nonsense or are you fixated on IP?

Re:Pretty much never?

[Pentium100]

Not exactly. Suppose you connect to the tracker and act as a VPN server. Suppose the only people that connect to you are also VPN servers with a few peers at the edge of the cloud. Suppose everything is set up so the routing data is in RAM and will be binned once the torrent is done. Also none of the VPN servers in the middle of the cloud can decode what they're passing on.

So... TOR.

Now at that point it's extremely hard for anyone to pin any piracy on anyone. It's even hard for the peers to tell what's really going on - once you go through two more VPN servers it's probably impossible.

Except that you know what everybody is downloading - the torrent in question. Unless by "tracker" you mean a server that tracks the "VPN servers", in which case you will get TOR.

TOR is slow because of all the relaying, BitTorrent over TOR(BT) will probably be just as slow. BitTorrent is fast because everybody is downloading from everybody, once you start putting peers in chains, the speeds will drop.

Re:Pretty much never?

[rrohbeck]

No. The BT blocks are much larger than the hash so there are (astronomically) many different blocks that map to the hash value and you can't know which one is the correct one.

Re:Pretty much never?

[bbn]

To the best of my knowledge there is no real standards for negotiating multi-cast between autonomous systems.

IPv6 takes care of that. Every /64 subnet also has an associated multicast range. This means every user on the internet also has the ability to multicast... provided the ISP bothered enabling it of course.

Technically there is not reason they should not enable IPv6 multicast. You can not saturate any pipes, as someone writes here. The distribution is optimized such that the stream is only broadcast through ports and paths that lead to active subscribers. If nobody listens, no data is sent. It is the most efficient distribution possible.

If I am sending a 1 Mbps stream (assume that is my upstream limit, so I can't be sending any more), then there will be nowhere in the network where a link will carry more than 1 Mbps of data from me. The total amount of data delivered can become huge of course, but that does not matter. The data is copied only at forks and if a lot of data is delivered it is because a lot of users are requesting it - the bill for that would belong to the users listening to the stream, not the one broadcasting it. 1000 users listening also means 1000 users sharing the bill so there is no problem here.

I do agree that ISPs are likely to sabotage multicast for business reason. They do not want any random user to be able to setup his own HDTV broadcasting operation from his home. They want monopoly on providing that service.

Also ISPs lose half their business. Because multicast is such efficient there will be much less bandwidth used. Currently 1000 users listening to a stream means billing the content provider for bandwidth times 1000. With multicast it first becomes 1000 streams near the receivers so the backbone is carrying several orders less data - that is less billable data.

Re:Pretty much never?

[rrohbeck]

Sooo.. if open, unencrypted P2P is going the way of the Dodo because of three strike laws then isn't that an argument for Freenet?
And if someone on Freenet downloads a packet from my node that happens to be part of a CP file, since I have no control over what goes through my box and don't know about it, doesn't the free haven provision in the DMCA apply? I'm only a provider.

Re:Pretty much never?

[bbn]

Let me expand a little on this. Assuming the following setup:

A content provider at ISP "A" is sending a radio stream.

"A" is connected to backbone provider "B".

There are 100 users listening at each of the ISP's "C1", "C2" and "C3". They are also connected to backbone provider "B".

With the current unicast solution "A" is delivering 300 streams to "B" and getting billed for that. Each of the "C" ISP's are receiving 100 streams from "B" and also getting billed.

With multicast this changes. "A" is delivering 1 stream to "B" and getting billed for only 1 stream. Each of the "C" ISP's are receiving 1 stream from "B". So "B" gets to bill a total of 4 streams compared to 600 streams losing big time. The fact that he can only bill for 4 streams instead of 6 is really nothing compared to the fact that he just lost traffic for almost 600.

The "C" ISP's are very happy though. It is true that somewhere inside their network this stream becomes 100. But it is also true that nowhere in their network will a link carry more traffic than before, so there will be zero upgrades needed. They can pocket all the money they used to pay "B".

If "B" feel really left out he could charge a premium to make up for the 4 vs 6 streams. It will never recover his big loss though.

Now the "C" ISP's are still going to refuse to implement multicast. They are going to make much more money on their IPTV monopoly than by allowing multicast.

Re:Pretty much never?

[hairyfeet]

Only if you are an ISP. Are you an ISP? I think not. There are plenty of things a corporation can do that you the average Joe can't, see limited liability for an example.

The problem is the fucked up way that the CP distribution laws are written. In their current form if you have ZERO access to the package, simply deliver the package, like I said you are on the hook for distribution since the distribution charge does NOT have a single line about you yourself actually using or looking at said package simply that you delivered it and that is what the IP logs prove. Hell I doubt they would even need to crack any FreeNet crypto, simply show they download "x yr old fucking" and traced the download to your IP address. After all they busted a guy not too long ago for a single thumbnail image generated by one of those topsites, and I doubt anyone could really sanely argue one is gonna be using a thumbnail for anything.

In the end because the courts haven't really covered anything like FreeNet that means SOMEONE is gonna get drug into court sooner or later so that the courts will rule one way or another. Now considering on average for a CP case at least it my state you are looking at 2 years before coming to trial, with VERY high bail if you get any at all, along with a minimum of $100,000 to defend yourself unless you are willing to risk your freedom to a "public pretender' that will ONLY push you to please guilty (I know, I used to work for a bail bonds and saw how they operate) the question becomes thus: Are YOU willing to risk 60+ years of your life and 2 years minimum in county simply to find out if FreeNet has a legal leg to stand on? Seems like a hell of a lot to risk on an untested theory and I wouldn't be surprised if guys like the EFF shied away from jumping in on anything involving CP.

Re:Pretty much never?

[Tacvek]

If one packet that becomes 100 is charged 100x, then they are not charging for bandwidth, since that would be a flat cost for an x Mbps pipe. Remember bandwidth is throughput, and is measured in bits per second.

If there is an allowance, then you are clearly referring to data utilization, (the integral of utilized throughput over a period of time) which is measured in bits or bytes.

-------
Let me outline all this.

Consider the following ISPs. BV&V, and Zerivon are huge Tier-1 ISPs.

We have a Large ISP named InternetCo. InternetCo is a Tier-2 ISP.

InternetCo buys transit from both BV&V and Zerivon. InternetCo is also peered with several other ISPs, but their peering agreements cover local delivery and delivery to customers only, no transit to other networks is provided.

InternetCo sells transit to several ISPs. They sell based on data usage too.

PhoneCo is medium sized Tier-3 ISP, who buys transit from InternetCo. PhoneCo sells Internet service to its smaller divisions. The divisions sell to end-users and to small ISPs. The pipes between the divisions and the main PhoneCo network are oversubscribed. PhoneCo sells service based only on the size of the pipes.

CheapNet is a small ISP. They buy transit from a phoneCo division via a DSL line. Their DSL line is oversubscribed of course. They sell only one speed to customers at a flat rate sufficent to cover their flat rate from PhoneCo plus some profit.

-------

Ok. Now lets say some server in the Zerivon network is multicasting some work. Well InternetCo gets a single copy of each packet over the the connection from Zerivon and and makes a copy for each interested customer. PhoneCo gets a single copy of each packet that they send a copy to each division that is interested. The division sends a copy to CheapNet. Because of multicast, only one copy of each packet goes over the oversubscribed pipes. Everybody is happy with this situation.

Now, what happens if a CheapNet customer is sending a multicast stream out. One copy of each packet goes to CheapNet. CheapNet makes a copy for each interested customer, but they don't really care about that. They care only about the fact that they send at most one copy of the multicast packet to the PhoneCo division. The divison happens to have no interested customers, but there are interested customers upstream, so it sends one copy of each packet to PhoneCo. PhoneCo sends copies to interested divisions, and passes a Copy to InternetCo. InternetCo sends a copy to each interested customer, as well as to interested peers, and also to each of BV&V and Zerivon.

Intenernet Co does not like this. First of all, they feel that PhoneCo shouldhave to pay more for packets that result in 50 copies getiing sent to customers, and peers, than for a packet that goes to only one customer or peer. InternetCo though has no easy way to measure how many times each packet is split, so it really cannot charge PhoneCo appropriately. But even so that is not a big deal.

The big deal is that if this stream consisted of 20 GB of total data, PhoneCo payed sending 20 GB of data. However InternetCo payed sending 20GB of data to BV&V and payed for sending 20 GB of data to Zerivon.

If InternetCo is not charging PhoneCo at per gigabyte at the sum of the per gigabyte prices of BV&V and Zerivon, then InternetCo is losing money on this multicast stream. They will not tolerate that very well. While they might be able charge PhoneCo enough, other Tier-2 ISPs most certainly cannot. For example a Tier-2 that buys transit from 5 providers likely cannot charge its customers the sum of the per gigabyte prices of all 5.

---
The big problem here is sending multicast packets in the customer to providers direction. In that direction you can get problems like the one noted above.

However in the other direction (providers to customers) the ISPs are perfectly happy. They don't need to pay anything additional when they duplicate packets to send to their customers. If they charge customers per gigabyte, they are actually making more profit than usual with multicast, since they pay their provider (if any) only once, yet get payed for each of their customers interested in the stream.

Re:Pretty much never?

[Tacvek]

That is all true.

However, there is another problematic case.

Provider "B" sells full transit to provider "A". "B" pays for packets sent to/received from providers "C" and "D", but is not buying full transit from them, so "C" will not deliver packets between "B" and "D", and "D" will not deliver packets between "B" and "C". (My understanding is that many Tier-2 providers have some "B" and "C" where this is the case).

If a user on provider "A" creates a multicast stream, that 100 customers of "C" and 200 customers of "D" want to see/hear, the following happens. "A" pays "B" for the sending one stream. "B" sends the stream to "C" and "D", and pays each for sending one stream.

In unicast "B" pays for sending 300 streams, but "A" pays him back for those plus markup, so he comes out ahead. With multicast "B" pays for sending two streams, but "A" only pays him for sending one stream. Depending on just how much "B" charges "A" for service, "B" could go from making significant money on the streams to losing money on them.

"B" would like to charge "A" for this, but "B" has no easy way to track this, since in the unicast world all "B" dd was count the bytes on the pipe between "A" and "B". Now "B" would need to flag the multicast packets from "A" and count them at the connections to "C" and "D" as well, which they are not equipped to do.

Lets say "B" sets things up so they can double bill if a packet went to each. "A" might be multi-homed, and wants to be able to know how much each provider will charge them, so as to route traffic to minimize costs. (this can make good sense if one or more of "A"'s providers has some form of variable pricing or overage fees). That is fairly easy in the unicast world. However, in the multi-cast world "A" has no way of knowing if a multicast packet sent to "B" will get billed twice or not, making it much harder to optimize things.

Re:Pretty much never?

[bbn]

I get you, but this is probably not such a big issue in practice. You just charge worst case for any received multicast packets. The tier 1 providers do not pay anyone and simply charge double on any outgoing multicast packet.

But this is assuming you are going to route multicast the same way as unicast.

All the smaller ISPs in my country might have deals with larger ISPs just like the case you describe. But they also all have a deal with a tier 1 provider that is willing to route all their traffic. The other deals are just an optimization to save money and improve latency. Such an ISP could choose not to take advantage of their direct connections to large ISPs and send all multicast to their tier 1 upstream. This forces the overall charging structure back to the case I described.

The resulting distribution might not even be any less optimal. There are many equally optimal ways to connect a graph.

Re:Pretty much never?

[gottabeme]

Come on. Apple only did that so they could sell more than their competitors and help put them all out of business. It had nothing to do with the idea of DRM being evil--Apple is all about DRM! Apple IS all about control, and as they emphasize iOS and their app store--which are fundamentally based on DRM--they are moving towards more control for them and less freedom for their users. They also patent everything in sight, whether it's original or not, and use that to stifle competition.

Fact is, as good as their products are (I have an iPhone 4, and had a 3G, but I don't plan to support them with anymore of my business), at their core they are fueled by greed, and they are evil. They're no better than Microsoft in their business practices--they're just as immoral--they simply make much better products.

I think you're being hypocritical: you are the fanboy.

Re:Pretty much never?

[Goaway]

And speaking of internet fanboys...

Re:Pretty much never?

[Tacvek]

Very true. Unfortunately there is no easy way to check how big a problem something like that would be, since ISPs don't go publshing full information about peers, customers, and transit providers.

I do know that in the US there are several Tier-2 ISPs that are larger than many Tier-1 ISPs, and have settlement-free peering agreements with most of the tier-1 ISPs, plus are peered with many other Tier-2 ISPs. Thus they pay nothing to route packets to most of the internet. However, they do not have free routing to the entire internet, which is why they are not Tier-1.

One way this can happen is the Tier-2 ISP in question is not able to peer with one of the overseas Tier-1 providers, like Tata Communications, or DTAG, simply because they don't have a presence at any Exchange points where the Tier-1 in question operates as.

In order to get connectivity then they generally buy transit to that network from one of their peers. (Remember that a standard peering agreement with a tier-1 covers routing to that tier-1 network, and its customers, but does not normally cover routing to other tier-1 networks).

Re:Pretty much never?

[gottabeme]

Fanboy of what?

When...

[smileygladhands]

It will be replaced when our ISP monopolies makes it so difficult to use bittorrent, another way must be created. Destruction brings creation.

Re:When...

[StormReaver]

It will be replaced when our ISP monopolies makes it so difficult to use bittorrent, another way must be created. Destruction brings creation.

We're getting there very quickly. ISP data caps will greatly degenerate the usefulness of Bittorrent for many, many people.

File size range

[tepples]

Once the pirates discovered BitTorrent, I seem to remember that there was a shift from pirating 3-6 MB singles on Kazaa, Gnutella, and the like to pirating 50-100 MB albums. Has BitTorrent since become better at transferring small files, or is it still suited only for large transfers?

Re:File size range

[Dyinobal]

Bit torrent can do large or small files with equal ease. It's just the distributed method of seeding really shines with large files.

Re:File size range

[monoqlith]

Unless the files are zipped into an archive, a lot of clients will let you choose which particular files you download from the torrent and skip all of the other ones.

Re:File size range

[sourcerror]

By the way, when you're starting to download a torrent that consists many files, you can select only the ones you want. You don't have to download the big stuff if you don't want to.
There's nothing in Bittorrent that discriminates against small files.

Re:File size range

[mmcuh]

But to be fair, small trackerless torrents can take disproportionally long to download since it takes a fixed amount of time to get peers from the DHT.

Re:File size range

[tepples]

once you've decided to download the torrent, why do you just want this one single instead of the entire discography in flac and mp3?

Because I want to be able to download other things during the same month without having to pay prohibitive overages.

Re:File size range

[demonlapin]

It works just fine for small files; the problem is that doing so craps up the bittorrent software's window with a bunch of different torrents. Something like the eMule Kad protocol is better suited to handling lots of small files from a user's point of view.

Re:File size range

[Arancaytar]

However, users who don't know how it works will often stuff all those files into a an archive before creating a torrent (usually something stupid like .rar), which firstly is useless as the files are already in compressed formats, and secondly removes any control over file selection.

Re:File size range

[oakgrove]

Has BitTorrent since become better at transferring small files, or is it still suited only for large transfers?

Whatever issue you're having, I'm not seeing it.

Re:File size range

[blueg3]

BitTorrent is reasonably good at small files. Other people seem to be claiming that it's *as good*, but that's not quite true. Clients can be reluctant to try to download the same piece multiple times, so any torrent with a small number of pieces gets lower performance than one with a higher number of pieces. (That is, if the number of pieces is substantially below the number of peers you could be downloading from simultaneously, you can see a performance hit.) Also, the whole web browser -> download .torrent -> start download in client cycle is a bit of high overhead for a single small file, and most clients limit number of shared / downloading items by number of items rather than some measure of size, so for a lot of clients, they'll only download 4-8 small files at once if those files are in their own torrents (without reconfiguration).

Share ratio requirements

[tepples]

BitTorrent might not be replaced until tracker operators learn what an average is. A lot of private trackers require their users to keep their share ratios at or near 100%. But it's mathematically impossible for everybody to have a share ratio greater than 100%. Share ratio is upload divided by download, but across a whole swarm, the sum of upload will equal the sum of download, making the average share ratio 100%. One can't seed unless there's a downloader on the same swarm. So what are people who get in on the tail end of a swarm, where no downloader shows up for days at a time, supposed to do to keep their share ratios up to the tracker's standard?

Re:Share ratio requirements

[Zironic]

Most trackers only require a ratio of 0.8 or 0.7 for that reason.

However I've never had a problem seeding myself into the positive.

Re:Share ratio requirements

[Tynin]

I think this is actually intentional. Most of the private trackers I use that use ratios and penalize you for going below a certain percentage, also include the option to pay them some mount monthly/yearly for premium access where your ratio is overlooked.

Re:Share ratio requirements

[gl4ss]

keeping logs of who shared what or how much is stupid. if you want such systems there's plenty of ratio aware solutions for doing it and plenty of direct connect hubs which have minimum share requirementst too - but such systems are always bustable. also, requiring 100% ratios just means people will keep creating shadow users and upload to themselfs, that's stupid too.

Re:Share ratio requirements

[ktetch-pirate]

Actually, it's worse. Remember, the initial seeder doesn't download, so that's at least one upload without a download. So while the average across the torrent is 1:1, in reality, the peers/users on it will be less than 1, with the initial seeder removed. Also, can we stop calling them 'private trackers' - there's NOTHING private about them. They're LESS private than public/open trackers. They're Registration Required trackers, or Reg-wall trackers.

Re:Share ratio requirements

[Hatta]

Sites with ratios generally have some sort of stimulus program that keeps the credits plentiful. Underground-Gamer for instance has golden torrent weekends, where the most desirable torrents on the site are free leech and 2x upload credit for seeders.

Re:Share ratio requirements

[tepples]

They're Registration Required trackers, or Reg-wall trackers.

I was referring more to invite-wall trackers.

Re:Share ratio requirements

[tepples]

However I've never had a problem seeding myself into the positive.

If you are among the last to download a given file, how many weeks does it typically take you to seed to 1.0?

Re:Share ratio requirements

[Zironic]

Tracker Seeding ratio is usually measured as a total across all files, not individual files. Trying to get 1.0 on all files is a bit silly.

However most files I download reach 1.5 within the hour.

Re:Share ratio requirements

[ktetch-pirate]

Maybe, maybe not. A lot of RR trackers are scene based. If you've got access to a scene dump to make the torrent in the first place, you don't then need to use the torrents to get more stuff also available on the dump.Besides which, it eats into your e-peen 'ratio'

Re:Share ratio requirements

[limaxray]

You are correct - seeding torrents is like a P2P pyramid scheme and the people on the bottom are left holding the bag.

The thing is, this situation is a rare occurrence for most users, and most will be able to seed greater than 1 most of the time. In my experience, the number of torrents you can comfortably seed greater than 1 dwarfs those that you can't. While I have found torrents on private trackers to be typically very well seeded, often to the point of saturation, I've never had a problem maintaining a positive ratio and I usually don't seed more than a day or two.

Re:Share ratio requirements

[KingAlanI]

Yeah, sometimes if I'm amongst the last to download a big torrent, I don't even bother to seed it because I'm not really needed - I'd rather my uploading be focused on the rare/unpopular stuff and/or the torrents I make.

Re:Share ratio requirements

[Rockoon]

The thing is, this situation is a rare occurrence for most users, and most will be able to seed greater than 1 most of the time.

You are failing at the exact math that you are replying to. For every person greater than 1.00, there is at least one person less than 1.00. This makes it impossible for most users to be over 1.00 most of the time.

Re:Share ratio requirements

[tepples]

A lot of RR trackers are scene based.

Which is another part of the problem. I'd even be willing to make my own original works available for upload, but a lot of the ratio trackers don't want them because "the scene" only cares about ten major distributors.

Re:Share ratio requirements

[Zironic]

No there's not. For instance you could have four users with 1.2, 1.2, 1.2 and 0.4. In this case the majority has above 1.0.

Re:Share ratio requirements

[Rockoon]

The grandparent premise is that everyone must maintain near 1.0 on many private trackers. There cant be a 0.4 on said tracker.

Re:Share ratio requirements

[js_sebastian]

The thing is, this situation is a rare occurrence for most users, and most will be able to seed greater than 1 most of the time.

You are failing at the exact math that you are replying to. For every person greater than 1.00, there is at least one person less than 1.00. This makes it impossible for most users to be over 1.00 most of the time.

No offense meant, but it seems you also fail at math. In the limit you just need one leecher downloading huge amounts of content from all other users, to keep them all at arbitrarily high share ratios.

Re:Share ratio requirements

[Ryanrule]

The one I know gives points for just having torrents seeding, with more for older/only seed. These can be exchanged for upload gb or other things like custom forum titles.

Re:Share ratio requirements

[tepples]

In the limit you just need one leecher downloading huge amounts of content from all other users

And in the limit, the leecher gets b&.

Re:Share ratio requirements

[Joce640k]

However most files I download reach 1.5 within the hour.

Ah, you're one of those with massive upload bandwidth who keeps his ratio huge and makes it impossible for normal people to seed.

Re:Share ratio requirements

[Pentium100]

Actually for a single torrent, the ideal ending is one peer with infinite ratio (the original seeder, since he does not download anything), two peers with 0.5 ratio and the rest with 1.

Re:Share ratio requirements

[Pentium100]

Depends on the tracker rules. For example, the tracker may require upload speed at least 10 or 100 mbps to get upload rights. If you don't have it, you won't be able to upload.

Sneakernet

[tepples]

And with numerous ISPs capping home users' monthly transfer in the double digit or even single digit GB, "another way" is likely to involve sneakernet.

Re:Sneakernet

[LordSnooty]

sneakernet

I still want to see some kind of interface between mass storage devices, wireless and content, enabling WLAN content parties. You rock up, your device sees everyone else at the event and starts sharing, perhaps intelligently based on your preferences, Maybe smartphones hold the answer to this, but they may not offer lots of storage. However this could work on a small scale. You could just turn up to the pub and your device would sort out the rest. It could share your own content and look after the security. Could even base it on bittorrent for max efficiency. That would be very cool. Someone make it happen.

Re:Sneakernet

[Lanteran]

United states.

Re:Sneakernet

[qpqp]

I believe, you want something like a Napster combined server/client for LAN that forms an ad-hoc mesh with other clients.

Re:Sneakernet

[fuzzyfuzzyfungus]

I don't know of anything pre-rolled(or, more importantly, enjoying critical mass); but it doesn't strike me as a terribly difficult problem(aside from the critical mass bit).

While the use case is comparatively rare, bittorrent clients will happily enough interact with other nodes on the same LAN. With DHT, you wouldn't even need a tracker(though it wouldn't be rocket surgery to configure the router serving the LAN to have a little captive portal page where people on the LAN could upload .torrents for the benefit of others on the LAN who don't have them(and thus don't participate in the DHT). As long as DHCP is provided, pretty much any vaguely contemporary OS/BT client configuration would Just Work. To avoid spilling onto public internet, and attracting the wrath of the ISP or the Copy Cops, you could either simply not have a WAN connection, or have one; but not route the BT traffic(or shape it to lowest priority, to allow fresh material to come into the local swarm; but not swamp other things).

There are really two main issues:

1. Critical mass: Unless this sort of thing is relatively common, it'd be little more than a cute trick, arranged well ahead of time, for your local LUG or LAN-party or whatever.

2. What client hardware to use: In one sense, this is easy. If it has wifi or ethernet(or even bluetooth/zigbee/IrDA/PPP-over-RS-232 if you are feeling perverse) and can run a bittorrent client, it will work. However, compared to almost anybody's internet connections, LAN transfers will be Fast, but the inconvenience of going out to use bittorrent will be sufficient that it won't be worth it unless everybody brings a bunch of stuff with them, and a bunch more space for the stuff they want. Nothing stops a cellphone user or somebody with a 320GB(total, much of it already used) laptop from taking part; but it'll be the guy who shows up and plugs in his 6-bay NAS, with the (almost always Linux) firmware hacked to include a BT client that will really keep things running...

Re:Sneakernet

[SuricouRaven]

In this case, 'legitimate' means 'Has either business ties with a major ISP, or the money to pay them for the deal.' Netflix is legitimate, but Small Town Studio trying to distribute their low-budget independent movie is not.

Re:Sneakernet

[Nutria]

It's not normal to pay expensive services and that in more than a decade no upgrade has been done to their infrastructures

Are there really non-dialup ISPs that are as slow in 2011 as they were in 2001? Cox cable sure is faster than it was 10 years ago...

Re:Sneakernet

[klapaucjusz]

With DHT, you wouldn't even need a tracker

Actually, you don't even need the DHT. Most BitTorrent clients implement "local network discovery", a protocol extension that allows them to automatically and quickly discover peers participating in the same swarm on the local LAN segment. It's usually disabled by default, for obvious privacy reasons, so you'll want to hunt for the features in your preferences dialog.

--jch

Re:Sneakernet

[fuzzyfuzzyfungus]

How handy. That(again) makes everything except the 'critical mass' part even easier. Heck, there are already plenty of places with public wifi that don't bother, or don't properly implement, client isolation, and I'd imagine that virtually none of them would be configured to do much meaningful filtering of traffic between LAN clients(anybody who cares just seems to use client isolation, or doesn't offer wifi to the public), everybody else either does nothing, or does some basic throttling/blocking to keep spammers and leeches from ruining everybody else's latte-and-email.

You'd still need to get enough people, and enough material, together that it wouldn't be easier just to swap USB drives; but with the feature you describe, and basically any router, you could have a LAN swarm immediately, if you wanted one.

Re:it already is almost dead due to ISP's

[mmcuh]

You seem to be living in the wrong country, or using the wrong ISPs. I've lived in different cities in Sweden for the past decade, and I've never had any problem getting maximum download speeds on a decently seeded torrent, whether that has been a lowly 200KB/s or 1MB/s.

Could be replaced by anonymous network

[mmcuh]

The one thing that could replace BitTorrent as the major filesharing protocol would be a protocol that is more anonymous and harder to track, in case people would get more privacy-conscious in the future (yeah, right). Even then it would probably be something evolved from BitTorrent, like OneSwarm.

It already is almost dead due to stupid users

[ktetch-pirate]

It's my experience (from 6+ years of doing frontline support, plus a few years as a p2p researcher) that most people have bad settings for their clients. They 'tweak' them to try and make things better, but in reality only make things worse.

Re:it already is almost dead due to ISP's

[Sinthet]

Just the opposite happens for me actually. I'm supposed to have 3 mb/s Internet, generally, the most I get via http tends to be 250kb/s, while bitorrent tends to hover around 300/330 for torrents with a good seed/peer ratio. No idea as to why this is, but in my experience, properly configured torrents tend to download faster than direct http transfers.

Re:Notice something about Bram Cohen?

[zill]

Prince of Darkness: "So what do you want in return for your soul?"

Cohen:"An efficient peer-to-peer file sharing protocol that can't be killed by RIAA/MPAA. Oh wait, eternal youth too."

Wi-Fi NAS

[tepples]

I still want to see some kind of interface between mass storage devices, wireless and content

I believe it's called a NAS (not necessarily that Nas). Wi-Fi NAS exists.

Re:Unlikely

[Opportunist]

Erh... the conditions you describe are met. I have virtually endless bandwidth and big content is selling its content online for less than a buck.

In my experience it's that plus slow seeders.

[YesIAmAScript]

There is sometimes a "keepalive" seeder for a torrent, but that seeder won't give more than a few K per second. I guess it's better than nothing, but it's really frustrating.

All in all, most things work quite well. Especially anything that is moderately popular.

Don't expect you'll be able to download everything at max speed though. Sometimes there just aren't enough other people offering it.

Even the tracker isn't centralized.

[satuon]

I would argue that even the sites and the tracker aren't centralized in the sense that there is no single torrent site/tracker - there are many, and nothing prevents anyone from creating a new one. In smaller countries, there are local torrent sites which only the people of that country use and know about. I doubt anyone is going to go after them. When people say that torrent sites are centralized they think about the highly visible targets like pirate bay or isohunt which everyone knows.

10 years & still only 1 broswer supports it, O

[AmigaHeretic]

BitTorrent still doesn't seem remotely mainstream still. I know with Opera you can basically treat a torrent almost like any other download. I'm not sure why other browsers never took this approach. I know for the e-l33t around here you all want a separate client, but for those that just want to download the occasional torrent the browser seems like the logical place to support ahhh...."downloading" of a file. I don't know....

Re:it already is almost dead due to ISP's

[Lehk228]

most http servers limit how fast they will send you data, try downloading from several sources and you should be able to max out your connection

Re:it already is almost dead due to ISP's

[Mashiki]

You seem to be living in the wrong country, or using the wrong ISPs...

Well that applies to Canada, the US, and most of Central and South America then. The only countries that don't are in Asia, and a couple in Europe.

Re:Unlikely

[CRCulver]

A buck per track is still too much. If the prices for audio files are so high, I'd rather pirate in the short-term and, in the long term, save up money and buy CDs for a couple of dollars more so that I get the physical artifact (looks nice on my shelves, serves as a backup).

Re:it already is almost dead due to ISP's

[CRCulver]

It's more than a couple countries in Europe. I live in a small town in Romania, for years already we have fiber to the door and 1 Mbps down on torrents with no throttling. Of course, when I go to Finland several times a year for work, things get even better, but even the European backwoods is better than what you get in most US metropolitan areas.

Re:Unlikely

[SuricouRaven]

You are mostly right, though there is still one area that I think you might want to concede: Staged releases. Pirates like to get things fast, preferably fast enough to see them at the same time as their friends. While you can legally download films now, you can't do it right away - first it's cinema time, and then blu-ray and online. If you live outside of the first release country it's a lot worse, as by the time you can see the film it's already last-months fad and you've heard all the spoilers and missed out the chance to complain about how much it sucks. If the studios are to have any hope of removing the non-financial motivation for piracy, they are going to have to drop the practice of region-stageing releases and just let every country see it at the same time.

The following hand gesture is dedicated to Cox...

[KingAlanI]

I'm not too pleased with Cox's 200GB cap that amounts to only a few percent of what my 15mbps down/2mbps up is theoretically capable of.

Re:it already is almost dead due to ISP's

[SuricouRaven]

As a networker, I understand why that can happen - but only under conditions of very high RTT. Very high. Unless you're getting pings on the order of a second or so, that shouldn't happen.

An overseeded first file

[tepples]

Tracker Seeding ratio is usually measured as a total across all files, not individual files.

True, but if the first file you download at a given tracker is overseeded, then your total across all files will stay low as most downloaders will get their blocks from someone else.

If it's not on NFOrce then forget it

[tepples]

You are forgetting that there is one surefire way to boost your ratio: upload new torrents to the site.

And get banned from the tracker because the torrents are something other than warez-scene releases. Why does Google have over 4,000 results for "if it's not on NFOrce then forget it"?

Another way is to look up the most popular torrents and begin seeding those. This is bit trickier as you need to download the file(s) first, which in turn may eat your ratio.

Not to mention the ISP's cap.

Re:The following hand gesture is dedicated to Cox.

[odirex]

Bro, your ISP is named "c**ks", didn't you think for a minute a company with that name might f*** you?

Re:I hope the concept is never replaced...

[Nutria]

Tor network

Don't they store anonymous other users' data in encrypted format on your hard drive?

Re:The following hand gesture is dedicated to Cox.

[Nutria]

I'm not too pleased with Cox's 200GB cap

Cox has a cap? Since when? (I recently d/l'ed much more than 200GB in a month and they didn't slow me down or charge me extra.)

Re:Unlikely

[nabsltd]

>But "big content" (oh, populism... you are endlessly entertainingly retarded) does allow people to download things legally, although you have to pay.

Where do I send my money to be able to legally download the just-released Lord of the Rings extended editions in full Blu-Ray quality (including audio and all extras), and play it on whatever device I want?

Seriously, if I could do that, I would. But, "big content" won't let their content out in the format that end users want, and that really is one of the largest reasons why there are "pirates". I "pirate" Blu-Ray disks every week, because there is no 100% legal way for me to copy the content to my hard drive, or to copy it to my portable media player.

Re:10 years & still only 1 broswer supports it

[AmigaHeretic]

Yeah, I remember people saying that about FTP. ;)

Why would you want to "download" a file using FTP in a 'Browser", that's what an FTP client is for.

Re:10 years & still only 1 broswer supports it

[AmigaHeretic]

Yeah, and maybe I'm wrong. but if Firefox and Chrome etc, supported in right in the browser and it was easier for people to use it might be used for more legit content. I could be complete wrong about that though.

I think a lot of legit adoption (ie BitTorrent replacing things like RapidShare for everyday quick uploads) has to do with learning how to seed a torrent the first time, learning to use a tracker, etc.

Re:it already is almost dead due to ISP's

[mister_playboy]

With http download services such as megaupload, filesonic etc... I regularly max out my 25Mbps connection while downloading

Are you are getting 25Mbps download from those services as a free user or as a paid user?

Re:Anyone remember the Bittorent test from July 20

[mister_playboy]

It was mentioned in TFA: http://developers.slashdot.org/story/02/03/20/0143248/Finally-Real-P2P-With-Brains

Re:Already replaced by anonymous network

[mmcuh]

Don't all of those use secret protocols and non-free software? That means that there is only ever going to be the one client, which probably would hamper popularity quite a bit. Not to mention that they won't be used by anyone who cares about free software.

In the free world there's Freenet and GNUnet, which are both theoretically very sexy but still very slow. Probably very, very hard to track though.

And will be elimiated

[nurb432]

When our ISPs charge per byte transferred at a totally oppressive rate. Who is going to give up their bandwidth 'just to help out'?

Re:10 years & still only 1 broswer supports it

[theurge14]

Congrats to the Opera fans, but for the rest of us the "browser that does everything approach" died with Netscape Communicator almost 10 years ago.

Re:it already is almost dead due to ISP's

[BKX]

Are you sure you're not confusing kb/s with kB/s? A speed of 3Mb/s will result in downloads of around 250-350 kB/s. That's because your ISP will rate your connection speed in megaBITs per second. Your browser and BT client will tell you speed in kiloBYTES (or megabytes) per second. 8 bits = 1 byte. Furthermore, your connection speed will reflect the total number of bits transferred, including protocol overhead, whereas your browser and BT client will only tell you about throughput (the bytes that were downloaded that actually get written to the downloaded file), which doesn't include overhead, making the ratio between your connection speed and throughput more like 1 Mb/s = 100 kB/s.

Re:10 years & still only 1 broswer supports it

[AmigaHeretic]

Congrats to the Opera fans, but for the rest of us the "browser that does everything approach" died with Netscape Communicator almost 10 years ago.

How many megabytes smaller is the Opera download than Firefox download again ;)

FF Win32 - 13.0mb
Opera Win32 - 9.8mb

Even if I just want to use it as "only a browser" I guess it's still smaller! Interesting. Oh yeah, and on topic it downloads torrents too!

Re:it already is almost dead due to ISP's

[Sinthet]

I'm not confusing it, maybe I wrote it wrong. I know I'm not supposed be to getting 3 megabytes per-second with my plan. My point was that Bittorent tends to give me higher speeds by 50 to 70 kilobytes. The misunderstanding here probably resulted from my improper use of notation.

Re:it already is almost dead due to ISP's

[Sinthet]

This is just my general experience from having this plan for the past three years, I know servers can limit how quickly they send you data. My "speed-test" results used to hover between 280 and 300 kb/s, though I haven't done one in awhile.

Re:it already is almost dead due to ISP's

[Kjella]

Here in Norway the average is now 7.4 Mbps and the mean 5.3 Mbps. About 80% of all households have broadband, yet we are less densely populated than the US (12 vs 31 people/km^2) and the average Norwegian lives in a town of less than 20,000 people. Yet we're still envious of Sweden and Denmark. Last figure is 14% of the population on fiber, but 20-25 Mbps lines are generally available on both cable and DSL. Most new installations are fiber though, which usually means symmetric speed... P2P when everyone is on 25/25 Mbit is way different than 25/5 Mbit.

Re:10 years & still only 1 broswer supports it

[SilentChasm]

And how many addons/extensions are you using in Firefox?

Re:10 years & still only 1 broswer supports it

[AmigaHeretic]

Maybe because the GNU/Linux distributions all include a simple client already? Nobody cares about Microsoft Windows users.

You miss the point. Your GNU/Linux distro has an FTP client built in too. Evertimey a DL link happens to start with ftp: do want to have a separate client open or do you want your browser to just DL the file? I bet 99% of the time when you DL something you don't even know if the DL link you click was http or ftp. Nor should you care. The browser takes care of it for you.

This is similar to how Opera handles torrents. For a quick DL you can just click it like I would any other DL and it's in my download list. Easy peasy.

Should it replace a full BitTorrent client? No.
Should a browser replace an full FTP client? No.

Should a browser be able to download a file quickly off the internet whether it be http, ftp, torrent, etc? Yes.

When I2P was created

[Burz]

one of the first apps for it was i2psnark, a built-in bittorrent client.

So, far from being replaced, bittorrent seems to be moving and thriving in at least one vehicle that can skirt ISP obstruction.

Re:10 years & still only 1 broswer supports it

[theurge14]

I see your point, but in my case I'm using Chrome with AdBlock.

Re:10 years & still only 1 broswer supports it

[AmigaHeretic]

I see your point, but in my case I'm using Chrome with AdBlock.

Chrome installer is like 24+ MB

Re:Unlikely

[Joce640k]

A buck per track...?

Re:it already is almost dead due to ISP's

[Skal Tura]

no, that is incorrect. Most http servers actually does not limit the data rate per connection, address or by any other means neither.

If you get better speeds with multiple connections, it most likely means your connection just sucks.

Re:10 years & still only 1 broswer supports it

[AmigaHeretic]

Try downloading a largeish torrent via it, leave it running for about 60 minutes, clock the average d/l speed - kill opera get utorrent (or a.n.other) to take over the download, watch the d/l speed increase by a factor of 10.

How about a few seconds? I'll give you the link so you can test it too.

I went to:
http://www.icarosdesktop.com/dl.htm

Here is the direct link:
http://www.icarosdesktop.com/icarosfiles/torr/IcarosLive_1_3_0.torrent

Here is a picture of the speed at just 4%
http://img198.imageshack.us/img198/6894/operatorrent.jpg

Now, that particular one is not the fastest torrent out there, in fact I think it is only seeded by a few people, but it is a legit one that I actually do download when there are updates to this particular AROS distro.

That is 866.3 KB/s reached in a few seconds in Opera's built in client. Now I run uTorrnet anyway for things I seed myself longterm, but it's nice to just DL a file and not worry about anything. It's also nice when I have Opera running from a thumb-drive (option from Opera's default installer BTW) that I can just DL a torrent from "anywhere" just as I would any other DL and not have to mess around.

Again, as the article says BitTorrent has been around 10 years. I think it's common sense that you should be able to download a "download" in your browser.

Do I use an FTP client when moving lots of stuff around on my server? Hell yeah. Do I use my FTP client when I want to download a quick file of webpage and the linked file happens to start with ftp: ? Hell no.

Already Replaced

[TheSpoom]

People in the know, who don't want to be sued, have already replaced it (albeit with an older technology).

Gnutella?

[gottabeme]

Isn't it decentralized? Seems to me like it might fit the bill for a LAN.

However, for a WLAN, all the bandwidth is shared, so I'm not sure having a P2P swarm would be any faster than simply using a Samba share. Ethernet, perhaps.

Re:The following hand gesture is dedicated to Cox.

[KingAlanI]

http://en.wikipedia.org/wiki/Winter_War
I thought the Finns and the Soviets didn't get along. ;)

Re:The following hand gesture is dedicated to Cox.

[KingAlanI]

to the offtopic mod:
I thought it was quite ontopic, since ISP caps interfere with making full use of BitTorrent (and other potentially-high-bandwith Internet uses, for that matter).

to the reply posts:
no, I can't really move just to get better Internet.
this is still an improvement over the crap DSL I had in my previous location. (my current building, otherwise a good location, doesn't happen to be wired for DSL or FIOS, else I would have checked out those options; any mobile internet probably would have been even worse, especialyl since I'm not mobile.)

Re:The following hand gesture is dedicated to Cox.

[KingAlanI]

maybe they have different policies in different regions.
I need to contact them soon to figure out how they measure that (and where I can check that), as well as what precisely happens when I go over.