BitTorrent Client Offers P2P Without Central Tracking

Shiwei writes "While BitTorrent is the most popular P2P protocol, it still relies on several centralized points for users to find the files they are looking. There have been several attempts at making BitTorrent more decentralized, and the latest Tribler 5.3 client is the first to offer the BitTorrent experience without requiring central trackers or search engines. Tribler offers some very interesting technologies; the latest version enables users to search and download files from inside the client. Plenty of other clients offer search features, including the ever-popular Torrent, but Tribler's results come from other peers rather than from a dedicated search engine. Users can search and download content without a server ever getting involved; everything is done among peers, without the need of a BitTorrent tracker or search indexer."

Comments

What

[AnonGCB]

Slashdot UTF fail. muTorrent, or utorrent, not Torrent.

Re:What

[Anonymous Coward]

Or kTorrent

Re:What

[Anonymous Coward]

Or uTorrent (for those who did the smart thing and did not read TFA.)

"Tribler offers some very interesting technologies; the latest version enables users to search and download files from inside the client."

Yes. BitTorrent software is very interesting technologies. However, I don't see how this is news. They even do go into further detail, still no magic new tech just a reference to "peer exchange (PXE)" and "distributed hash tables (DHT)"....

Sure wish uTorrent had some PXE's and DHT's. It sounds like a good idea. Oh.. wait.

Re:What

[GonzoPhysicist]

I always figured it was an SI prefix, e.i. 'micro-Torrent'.

Re:What

[Anonymous Coward]

Hello. This post is likeable, and your blog is very interesting, congratulations :-). I will add in my blogroll =). If possible gives a last there on my blog, it is about the Generic Cialis, I hope you enjoy. The address is http://onlinehealthtip.blogspot.com/. A hug.

Re:What

[asvravi]

He is talking about a million uTorrent.

Re:What

[pckl300]

It is, but nobody wants to look up the symbol every time they type it.

Re:What

[Anachragnome]

"A hug."

Hilarious, heart-warming and creepy, all-in-one.

Re:What

[blackdew]

The "news" seems to be the p2p search protocol, which isn't really THAT new of an idea, and isn't too useful until it's supported by all the major clients and they actually manage to be compatible with each other, which given their track record will happen in a century or so...

Re:What

[ArsenneLupin]

Is this a genuine spam, or just somebody's twisted sense of humor?

Re:What

[sauls]

No, dude, the creepiest part is that blog's post from November 22, entitled Sex is Not Love, and helpfully tagged "painfull Sex".

Re:What

[srussia]

"A hug."

Hilarious, heart-warming and creepy, all-in-one.

FWIW, the text is quite obviously a machine translation from Spanish.

Re:What

[Anonymous Coward]

alt-gr + M

atleast for my icelandic layout

Re:What

[spxZA]

Or anything that uses libtorrent. I modded libtorrent to work as a distributed content delivery system for an ex-employer. Remove the ignoring of unknown keys, add a bit of hackery elsewhere, and anything added into the swarm ends up on all clients. This was...um...1-2 years ago.

Re:What

[AI0867]

I'm guessing that they *did*, but slashdot ate it.

Re:What

[dogmatixpsych]

That's when you know they have won - we can't tell the difference between a spammer and a /. commentator. This is the beginning of the end.

Don't ask me who "they" are.

P.S. Of course it was spam.

Re:What

[bluie-]

For some reason I really enjoy 's' less plurals

Back in Time.

[Anonymous Coward]

So Essentially... bit torrent trackers are turning into Kazaa/Limewire type of software... We are going backwards.

Re:Back in Time.

[Goaway]

Don't worry, nobody will care this time either.

People keep trying to do this. At first we had P2P clients with integrated searching. None of them ever got all that popular. Then BitTorrent is launched, and it didn't have any searching, and relied on the much more familiar and comfortable web for that. It became a huge hit. What do these people do? They think, "Wow! BitTorrent is pretty great! But wouldn't it be even better if it had search?"

Predictably, though, they fail completely, every time.

Re:Back in Time.

[TheSHAD0W]

No, the protocol is splitting, with one fork preferential towards pirates and the other honoring the original aim, the legitimate publisher.

Re:Back in Time.

[Anonymous Coward]

So, decentralization = bad?
Or, search at your fingertips = bad?

I think your comment has failed.

Re:Back in Time.

[Yvan256]

You mean the legitimate publisher who wants to leech my limited monthly cap for their own purposes?

I'm glad Blizzard gives us the option to disable that in their games.

Re:Back in Time.

[pavon]

So you are saying that Napster never got very popular?

The reason that Bitorrent became popular was because it was a faster protocol, and thus worked better for large files like videos and games. It had nothing to do with people being turned off by integrated search.

Re:Back in Time.

[Pentium100]

I think that integrated search is a good idea - now if I want to download something, I have to search on pirate bay, if I don't find it there, I go to isohunt and so on. I'd like the ability to search on all public trackers at once. As a mater of fact I'd also like to be able to search all private trackers (that I am a member of, or optionally can become a member) at once. Centralized search would be bad, but a decentralized search that looks like centralized (you only search in one place) would be great.

And I think that BT became popular not because it relied on the web for searching (you can search for ed2k links too), but because it usually is faster than eMule and similar. Usually people do not keep a lot of torrents seeding at one time so each torrent gets better upload speed than if it was divided among thousands of files in the "shared" folder. The original BitTorrent client even had one window for each .torrent, with uT it's easier to seed 100 torrents, but still more difficult than it would be with eMule, especially if I want to move the files around.

Also, there are private trackers for BitTorrent, that enforce good ratio and thus make downloads faster (since there are more seeders), I never heard of private eDonkey servers and Gnutella does not even have servers.

However, BitTorrent also has its disadvantages. Sharing is more difficult than "drop the file in your 'shared' folder or share the folder that the file s in", users that have/want the same files cannot download from each other because the files are in different .torrents.

Re:Back in Time.

[hoggoth]

No Bittorrent client will be complete until it has an email client built in. A flight simulator would be nice too.

Re:Back in Time.

[Anonymous Coward]

Agreed. If this "step backwards" allows us to relive the glory days of Napster, I'm all over it. More importantly, though, decentralized services are nearly impossible for governments to shut down. Which, in our current political environment, sounds like a great idea.

Re:Back in Time.

[Anonymous Coward]

Napster bad! Fire bad! --James Hetfield

Re:Back in Time.

[jonwil]

Tribbler is also Open Source so the government cant shut it down (like they did to other p2p programs in the past that were not open source like kazza)

Re:Back in Time.

[beav007]

What about a text editor? I hear that EMACs has one now...

Re:Back in Time.

[tirefire]

I'm glad Blizzard gives us the option to disable that in their games.

My sarcasm detector is out of batteries. Does WoW really require subscribers to donate their upstream bandwidth for patches? Honest question.

Re:Back in Time.

[Surt]

Yes, they use a torrent based distribution system for their patches. So, yes, while you are gaming, you are typically using some of your upstream bandwidth to help deliver patches to others.

Re:Back in Time.

[Anachragnome]

"Tribbler is also Open Source so the government cant shut it down..."

I more concerned about Big Business getting their fingers in the pie.

From Wikipedia:
"After a dozen downloads the Tribler software can roughly estimate the download taste of the user and recommends content.[4] This feature is based on collaborative filtering, also featured on websites such as Last.fm and Amazon.com."

The problem is that collaborative filtering drives everyone in the same general direction--it is essentially distilling down one's tastes to the bare minimum. If too many people focus on these "suggestions", less popular torrents will die of neglect. The conspiracy theorist in me says that this exactly the idea--kill torrents, not all, just some. From the perspective of most media outlets, the only good torrent is a dead torrent.

There is also the possibility of gaming the system of collaborative filtering to intentionally steer interest in specific torrents.

I'll stick with TPB. Seed/Peer counts speak volumes.

Re:Back in Time.

[Anonymous Coward]

A truly small price to pay to ensure that the spice will flow when you need your next fix.

Re:Back in Time.

[Anonymous Coward]

Sons of fucking bitches!

Re:Back in Time.

[icebraining]

Torrentz is a free, fast and powerful meta-search engine combining results from dozens of torrent search engines

www.torrentz.com

Re:Back in Time.

[Anonymous Coward]

As the grandparent said: You have the option to disable it. Peer-to-peer patch distribution is enabled by default.

Re:Back in Time.

[neokushan]

Ironically, the day they invent a Bittorrent client that can suck you off will be the day they kill Bittorrent. At least for porn.

Re:Back in Time.

[Kjella]

The reason that Bitorrent became popular was because it was a faster protocol

A lot of that wasn't really the protocol as such, it's that you actually got faster downloads for faster uploads so people turned off all their caps. Napster etc. didn't really reward uploads much, you got the files at pretty much the same speed no matter what. Proper incentives are everything.

Re:Back in Time.

[IKnwThePiecesFt]

Woah Woah Woah. Yes they offer a torrent based system by default but turning that off is as simple as clicking Options->Uncheck peer2peer downloading.

So it's by no means forced and anyone with a bandwidth cap can easily turn it off.

Re:Back in Time.

[AmonTheMetalhead]

I just use google to find where the content is, instead of visiting every torrent site until i find it....

Re:Back in Time.

[Dr_Barnowl]

Yes.

In fact, some games (Lord of the Rings Online, I'm looking at you) install a content distribution service (Pando) where you basically agree to be part of their content distribution network, all the time, and not just while you are gaming. Unless you switch it off in the service panel, of course.

Re:Back in Time.

[Anonymous Coward]

Tribbler is also Open Source so the government cant shut it down

Tribler was developed (and is still maintained IIRC) by a research group at Delft University of Technology led by Johan Pouwelse. He has also appeared as an expert witness in several (at least one) US filesharing cases. As long as the UT continues his funding (right now he's being funded by the EU as well) I don't see Tribler being shut down any time soon.

That said, he is quite outspoken on his views on traditional media (distribution), so I wouldn't be surprised if at some point the UT will get pressured to release/reassign him. Then again, he has tenure...

Re:Back in Time.

[Anonymous Coward]

Is this such a bad thing? Obviously it's not great if you've got a monthly cap but if you've got the option to turn it off, how is this not a good thing?

Re:Back in Time.

[codeshot]

decentralization = hard

Re:Back in Time.

[kainosnous]

You see this as a bad thing, and personally, I agree. However, it's probably time to face reality. In this long September, the internet has taken a new shape: a hive mentality. Gone are the days where we just grab whatever information we come across because it is unique and exciting.

The net has simply become too large for that. These days, people seek features to cut down the signal to noise ratio. This in itself isn't a bad thing, but it's based on taste, and the taste is that of the majority who only care about the latest cute hampster video or what happened on American Idol. It's only natural that some of that "noise" will be useful information. I was on a torrent site the other day looking for programming texts. I was lucky if I could find more than a handful of seeders.

The key is to encourage people in our own niche group to participate in sharing information. Let the masses have Facebook and Twitter, but we must take part in distributed file sharing while there is time.

Re:Back in Time.

[Anonymous Coward]

That's not correct.

The launcher in 4.0.3.13329 does not seed while you are playing, and if you are streaming while playing (have not completely downloaded) the P2P is turned off and your only seed is the webseed (normally Akamai).

Furthermore, you can simply turn the P2P off entirely from the launcher: Options > Downloader Preferences > Enable Peer-to-Peer Transfer.

It may not be visible in the client yet or documented, and could be reset by some updates, but it's also possible currently to set a maximum upload speed in the Blizzard Launcher's torrent client by editing WTF/Launcher.WTF and setting uploadMax to a value other than 0 (it's bytes per second, so 50000 would be 50KB/s, and 0 = unlimited).

Re:Back in Time.

[Raenex]

The reason that Bitorrent became popular was because it was a faster protocol

No, it became popular because Napster was sued out of existence.

Re:Back in Time.

[Hatta]

I disagree. Having separate trackers with their own community was a big part of BitTorrent's success. It brings people together, they actually talk about what they're sharing, and they can organize to put together big projects that just didn't happen before bittorrent. It's not exactly the lack of integrated search that did it, but the lack of search pushed people to the web which is a much better platform for collaboration and communication. If this client doesn't even let people make and read comments on a torrent it really is a step backwards.

Re:Back in Time.

[Cowmonaut]

You forget, most users don't change ANY settings. Only the most "hardcore" WoW players would look that up, and out of the millions in its player base only a small percentage are "hardcore" (small enough that Blizzard has changed the game to be playable for 10 man groups instead of 60 man groups).

Re:Back in Time.

[Goaway]

Since you didn't seem to get the point, let me summarize:

Decentralized, search: All P2P clients except BitTorrent. All failed.

Centralized, no search: BitTorrent. Wildly successful.

Re:Back in Time.

[Goaway]

It was the only P2P client at the time. It was also never as popular as BitTorrent.

Re:Back in Time.

[Anonymous Coward]

judging by the protocol distribution of DMCA notices I process I'd say they are all in the same boat, all failed or wildly successful or both or neither. But all in the same boat. None hide from the agents of the rights holders and all are in significant use.

[anonymous for obvious reason]

Re:Back in Time.

[Atzanteol]

Is it 1997?

Re:Back in Time.

[pavon]

That's interesting. I never got into private trackers, so I had the opposite experience. I always liked how Napster let you browse people's shares. I would search for an artist I had heard about and then while downloading I would look through their collection and find all sort of other music that I had never heard of, and would download some of that as well. I discovered more bands that way then anything else at the time.

When Napster was shut down, I tried a few other packages, but the built-in search and public trackers didn't really have the same discoverability that Napster had. Shortly after sites like Myspace and later LastFM filled the role of try-before-you-buy, which is all I really used filesharing software for.

Re:Back in Time.

[sexconker]

This.

People seem to think that being decentralized will stop the lawyers from shutting down P2P or suing individuals.

You can never really decentralize - you always need to connect to other peers. The best way to find peers is via a central server that maintains a list of peers. It's possible to set it up so that you can "trust" this list of peers. In reality, however, P2P networks live or die by content availability and speed. How do you up those? Add more peers. If you had to wait 6 weeks to get an invite to a private tracker, that doesn't mean the RIAA/MPAA/etc. did the exact same thing.

If you take out the tracker, how do you find peers? People talk about distributed hash tables and peer exchange. No one really says the obvious - you go looking for peers. And from there you can find other peers. So how do you start looking? Where do you send your initial packets? You've either got a centralized server to handle the bootstrapping, or you just spam subnets, becoming the center of your own network map. This is akin to walking down the street yelling "Anyone got crack for sale???". If this is what you're doing, you've got no guarantee that you won't get fucked. In fact, you make it easier to piss off someone who wasn't even looking for you.

Being decentralized does nothing to make an individual anonymous, safe, etc. But surely it will help the network as a whole, right? Wrong. Instead of going after trackers and sites, they'll go after the initial distributors. For movies and games, there are only a handful of guys supplying 90% of the content. I guess music is easier for the average schlub to rip and share, but who the fuck wants the shitty music they put out today? Honestly. If you take out the ability to easily go after big sites, they'll just ramp up their efforts to go after the sources. They'll always focus on the easiest targets. They USED to focus on the sources (and did so successfully) until Napster came out.

So in the end, decentralization may protect the network, but it doesn't protect the users. As a downloader, you can feel relatively safe. As a source, you'll have a bigger target on your back. As a downloader, if sources are targeted, you'll have less to download.

As for trusting other users for searching?
Do we not remember Kazaa, LimeWire, BearShare, etc. etc. ? I hope your friends and family all run selinux.

Re:Back in Time.

[sexconker]

My sarcasm detector is out of batteries. Does WoW really require subscribers to donate their upstream bandwidth for patches? Honest question.

Require? No. You can turn it off.
If you want to get the patch before the next patch rolls out, then yes.

Re:Back in Time.

[nabsltd]

The conspiracy theorist in me says that this exactly the idea--kill torrents, not all, just some. From the perspective of most media outlets, the only good torrent is a dead torrent.

Except that the most popular torrents are the ones that media companies most want to shut down: new releases, blockbusters, current TV shows (which would be tricky to win a lawsuit), etc.

None of the media companies really care if people are downloading things like DVDs released 10 years ago, or 25-year-old CDs. For many people, though, those are some of the most important torrents, as you sometimes can't legally purchase the content any more.

Re:Back in Time.

[kheldan]

uTorrent has a Tetris game built in, you'll just have to be satisfied with that.

Re:Back in Time.

[funfail]

What makes you think that the only reason of BitTorrent's success is its lack of search? I remember that the switch from eDonkey/eMule to BitTorrent was solely because of BitTorrent protocol was much faster.

Re:Back in Time.

[funfail]

Do you have an upload cap?

Re:Back in Time.

[Yvan256]

The downloads and uploads are added together for the monthly cap of 35 GB.
http://www.telebec.com/english/magasinage_ligne/asp/internet/sympatico_hv.asp

Re:Back in Time.

[Surt]

I don't think that it is bad, myself, merely that it is a fact.

Re:Back in Time.

[dissy]

You mean the legitimate publisher who wants to leech my limited monthly cap for their own purposes?

So asking if you would mind helping out with spreading the cost of bandwidth around, where you are free to say yes or no, is now 'leeching'?

Remind me to never ask you for a favor.

You sound like the type who would respond to "Could I please use your cellphone for just two minutes to make an emergency phone call" with the response "Well I dunno, I only get 700 minutes a month, and even though I have plenty of minutes left, I still feel you somehow owe me money for those two minutes you want to use"
Or worse, you claim I can use your phone, and later try to hold those two minutes I 'owe' you over my head for the rest of time...

It's not like anyone is forcing you to leave the torrent running, especially in small legitimate publishers cases where they only have one or two files to distribute anyway and there is no such thing as ratio!

Re:Back in Time.

[Anonymous Coward]

Then what would make the Bittorrent client distinguishable from EMACS?

Re:Back in Time.

[Anonymous Coward]

AFAIK the torrent feature is/was enabled by default and those publishers won't care/can't check if my monthly caps are busted by their actions.

Do you really think they would pay part of my internet bill even if I had proof that they're the ones who wasted my bandwidth?

Peers Peering Particularly at Profitless Peers.

[Anonymous Coward]

What about man-in-the-middle peers?

Re:Peers Peering Particularly at Profitless Peers.

[Anonymous Coward]

man in the middle is good, at least when the man is me and I'm in the middle of a Kim Kardashian/J-Lo ass sandwich! Yum yum!

Re:Peers Peering Particularly at Profitless Peers.

[biryokumaru]

You really should get an account... like PizzaAnalogyGuy or something.

Another Victory

[Anonymous Coward]

Oh, hurray. Another victory for cheap entertainment consumers over the people who work to entertain them. Pretty soon, there will be absolutely no point in creating entertainment since it can never get compensated by the people getting entertained. Cheap-asses unite!

Re:Another Victory

[biryokumaru]

Yes, because the only reason anyone would ever create anything is to get a paycheck.

Re:Another Victory

[mr_bubb]

Yes, I pity the poor, hardworking, impoverished studio heads and record company execs, who have always been the primary beneficiaries of royalties. The Beatles, as a group, got a farthing per record, BTW.

Re:Another Victory

[icebraining]

It's the end of the world as we know it... and I feel fine!

Re:Another Victory

[Anonymous Coward]

NO, No and Yes, its not the only reason of course but getting a paycheck is surely as important as any other reason, you fucking ninny tard.

Now enjoy the fruits of your fucking idiot, slefish and childish need to acquire others property without compensation and watch as your stupid little world crashes all around as it is now doing

Re:Another Victory

[mcgrew]

Tell that to Cory Doctorow. I've slightly edited the quote for brevity, and the emphasis is mine. If you want to read the whole text, it's in the forward to Little Brother. The link is to the entire text of the book.

I recently saw Neil Gaiman give a talk at which someone asked him how he felt about piracy of his books. He said, "Hands up in the audience if you discovered your favorite writer for free -- because someone loaned you a copy, or because someone gave it to you? Now, hands up if you found your favorite writer by walking into a store and plunking down cash." Overwhelmingly, the audience said that they'd discovered their favorite writers for free, on a loan or as a gift. When it comes to my favorite writers, there's no boundaries: I'll buy every book they publish, just to own it (sometimes I buy two or three, to give away to friends who must read those books). I pay to see them live. I buy t-shirts with their book-covers on them. I'm a customer for life.

Neil went on to say that he was part of the tribe of readers, the tiny minority of people in the world who read for pleasure, buying books because they love them. One thing he knows about everyone who downloads his books on the Internet without permission is that they're readers, they're people who love books.

People who study the habits of music-buyers have discovered something curious: the biggest pirates are also the biggest spenders. If you pirate music all night long, chances are you're one of the few people left who also goes to the record store (remember those?) during the day. You probably go to concerts on the weekend, and you probably check music out of the library too. If you're a member of the red-hot music-fan tribe, you do lots of everything that has to do with music, from singing in the shower to paying for black-market vinyl bootlegs of rare Eastern European covers of your favorite death-metal band.

Same with books. I've worked in new bookstores, used bookstores and libraries. I've hung out in pirate ebook ("bookwarez") places online. I'm a stone used bookstore junkie, and I go to book fairs for fun. And you know what? It's the same people at all those places: book fans who do lots of everything that has to do with books. I buy weird, fugly pirate editions of my favorite books in China because they're weird and fugly and look great next to the eight or nine other editions that I paid full-freight for of the same books. I check books out of the library, google them when I need a quote, carry dozens around on my phone and hundreds on my laptop, and have (at this writing) more than 10,000 of them in storage lockers in London, Los Angeles and Toronto.

If I could loan out my physical books without giving up possession of them, I would. The fact that I can do so with digital files is not a bug, it's a feature, and a damned fine one. It's embarrassing to see all these writers and musicians and artists bemoaning the fact that art just got this wicked new feature: the ability to be shared without losing access to it in the first place. It's like watching restaurant owners crying down their shirts about the new free lunch machine that's feeding the world's starving people because it'll force them to reconsider their business-models. Yes, that's gonna be tricky, but let's not lose sight of the main attraction: free lunches!

Universal access to human knowledge is in our grasp, for the first time in the history of the world. This is not a bad thing.

In case that's not enough for you, here's my pitch on why giving away ebooks makes sense at this time and place:

Giving away ebooks gives me artistic, moral and commercial satisfaction. The commercial question is the one that comes up most often: how can you give away free ebooks and still make money?

For me -- for pretty much every writer -- the big problem isn't piracy, it's obscurity (thanks to Tim O'Reilly for this great aphorism). Of

Re:Another Victory

[jfengel]

Hands up in the audience if you discovered your favorite writer for free -- because someone loaned you a copy, or because someone gave it to you?

Wow, that's so deeply misleading as to count as an argument the other way. The obvious followup is, "And how many of you would have paid for it if it were available for free?"

Obscurity is the biggest problem of the vast majority of artists, but free downloads don't change that. Now you're just free among everybody else who's free. Publishing a book on paper or a CD in a music store makes you less obscure: you're among the handful of artists picked to do that. Somebody bothered to invest in a physical artifact. That's how you get famous enough that a non-trivial number of people share your work.

That model is failing fast, and yes, we need to find a new model. But artists who are already famous from the old model have less than zero experience with the new model; they have a deceptive and misleading opinion based on a position that very few artists ever get.

Re:Another Victory

[Anthony Mouse]

Wow, that's so deeply misleading as to count as an argument the other way. The obvious followup is, "And how many of you would have paid for it if it were available for free?"

You can borrow basically any book you can think of, for free, from a public library. People still buy those same books, do they not?

Re:Another Victory

[biryokumaru]

Right. I'm crazy because I don't think the entire world will fall apart if TV studios stop being so profitable. I mean, who will design and build our roads and cars and airplanes and houses and computers then?

Re:Another Victory

[jfengel]

Borrowing a library book isn't really "free", in the libre sense. It's due back at a certain date. You can only get it during the hours that the library open. If you lose it, you are REQUIRED to pay for another copy, so it's not even really gratis. And there are only a finite set of copies, so if the book is popular, you may not be able to get it at all.

Finally, I dunno what library you go to, but my library is far from having every book I want.

People buy books because they want to be more free with them. They're willing to pay for the convenience of having it any time they want. Even to loan it out, but it's still only a single copy: when you loan it out, you don't have it.

If unrestricted digital copies were allowed, none of that would apply. It may be that this is what the world is coming to regardless, so we may have to deal with it. But it's ridiculous for a famous author to claim that you can give it away free and somehow make it up in volume.

People may well pay for the convenience of print, something that still requires a real printing and binding machine. But if the text is given away, nobody will have reason to pay for anything above the printing cost.

Non-famous authors have problems getting noticed even if their books are free. Especially if their books are free, since they don't stand out from the millions of others. Print on demand doesn't change that. A commitment to pre-printed paper copies in a bookstore is something that helps an author stand out, and it's self-serving for Doctorow to forget about that.

It's not the only route, but the key is the investment of money: things that require no investment will always have to compete with the vast world of other free things, mostly junk.

DHT?

[Anonymous Coward]

Wasnt that the point of DHT?

Re:DHT?

[Per Wigren]

Yes, if you only read the headline. This client also has decentralized search.

To clarify

[PhrostyMcByte]

The summary (and TFA) is misleading. This client isn't the first to support trackerless downloading. Most clients support DHT and PEX, and have for some time. You just need a single peer to bootstrap yourself, and you're good to go.

What Tribbler has done is created a P2P torrent search engine. I'm not sure if they're the first either (I swear I remember reading about some other client with P2P search a couple years ago), but it does appear they put some thought into making their feature set more user-friendly, with categorization ("Channels") and such.

Re:To clarify

[DamienRBlack]

Wouldn't this experience be a lot like the old eDonkey 2000 experience. The problem with no centralized servers is that no one pays attention to ratio and the like. In fact I don't think most eDonkey users ever thought in terms of ratios. Also, there is no place to go to request stuff and ask for new seeds. The reason I switched to torrents, (and it took me a long time), is because of the centralized tracking. Sure, the popular stuff is usually available, but try to get something obscure that you can only find online and you are probably screwed. At least that is how eDonkey always was. Now eDonkey had some servers, but my point is that I feel like the experience would be the same as sharing on the eD2k network. No comments, no tracking, no ratio enforcement, no one pulling fakes and spam. eD2k was a hazardous wasteland but of mines. How would any of this be addressed with peer-to-peer torrents?

Re:To clarify

[Shikaku]

I'm not sure if they're the first either (I swear I remember reading about some other client with P2P search a couple years ago)

http://en.wikipedia.org/wiki/EXeem

It failed however, due to the initial version having spyware and subsequent loss of users.

Re:To clarify

[Dunbal]

While the ratio concept is a good one, it's totally borked by ISPs providing things like 3Mbps down and 64kbps up. No matter how hard you try, you'll never get a good ratio.

Re:To clarify

[Nikker]

You could create a shared directory on each client just like eDonkey did and hash them ... just like eDonkey did ;) Since the requests are going to your machine you could really return any results you wanted as long as it has the corresponding hash map it can send and receive.

Re:To clarify

[6350']

This is blazingly untrue. Most users don't have a synchronous line, yet with only moderate effort ("stay connected after you're done" and "actually pay attention to how much you have downloaded and take it easy") these millions have no reall issue maintaining a ratio. You're Doing It Wrong.

Re:To clarify

[Chuck Chunder]

"stay connected after you're done" and "actually pay attention to how much you have downloaded and take it easy"

I think the other factor is likely to be "get onto the torrent early". If you are limited to a tiny up speed like that I don't expect many clients are going to bother you much if there are plenty of other available peers that'll serve faster.

If you are there near the start when the seed may be bottle necked your 64k will be valuable and will remain so until there's more seeds than demand.

Re:To clarify

[the_other_chewey]

What Tribbler has done is created a P2P torrent search engine. I'm not sure if they're the first either (I swear I remember reading about some other client with P2P search a couple years ago)

Yup, the eMule/aMule network has had serverless search via the Kad Network for years. Works pretty well.

Re:To clarify

[Nursie]

Bootstrap is the interesting issue.

You can't have a situation with no server involved, ever, unless you're distributing the software on a friend-to-friend basis. There *has* to be a root node or selection of root nodes that the software knows about when it's installed, unless they have sufficiently advanced technology that it's indistinguishable from magic. Or they use some sort of brute force search....

Sure, once a node is online and given enough other nodes stay online enough of the time, it would be possible to have a persistent network.

I suppose you could do something like search google for random torrents, join in, test the folks you connect to for being part of the decentralised network, grab network info from there etc. It still uses google as a central reference point but it would be more robust than having some sort of hard-coded 'peer tracker' server, or using any sort of brute-force port scan of the internet.

Re:To clarify

[SilentChasm]

I would say the advantage of eDonkey2000 (or eMule now) is the lack of ratios per file. You share large numbers of files and download large numbers. What gets uploaded is what's needed and requested by others, not necessarily a specific torrent you want to get a higher ratio on.

The no comments/fake filtering/requests/reseeds can be mostly solved the same way as Bittorrent has solved it, with a link site/forum community.

The other major advantage of ed2k is that there won't be two separate swarms for the same exact file like Bittorrent. Bittorrent really needs some kind of standardized hashing method per file, even if it's just added data in a torrent with the original Bittorrent hashing remaining intact. A problem I've seen a few times is where two separate (old/rare) torrents are of the same file but they both have no seeds, only partial availability, so neither one finishes even if together they would have the whole file.

I've actually had better luck with rare stuff on ed2k than Bittorrent because of that lack of (unnecessary) duplication.

eMule Collections (file with list of links) or Magnet Links (uri with hashes/filenames) are kind of my ideal, a hash based system for finding stuff not dependent on any site staying up.

Re:To clarify

[Anonymous Coward]

Myself,the very reason I dont leave things to seed is i get annoying letters from my ISP threatening to cut my service when they catch me usually due to prolonged seeding.I pay very good $$ for both up and down speeds and should be able to share freely with other single use consumers.So without centralized tracking would I be able to leave my seeds?Im not sure if you're mising the point or am I???

kooky

Re:To clarify

[Anonymous Coward]

You just need one torrent link, right? Like this Tribbler software perhaps?

Re:To clarify

[icebraining]

Have you tried changing the ports and forcing encryption (disabling legacy connections)?

Re:To clarify

[Nursie]

Well, the TFA says - "Users can search and download content without a server ever getting involved"

And I was trying to figure out how that might work. If it's "no servers after the first time" or "no servers after a tribler peer has been found", that's fine, but you need something that fits the description of a server at some point in that. I was just thinking aloud at methods that might not require any user action and work straight off without requiring a specific Tribler central server.

'cos you can't just utter the magic words "In the cloud!" and have it work...

(not that they mention 'cloud', it's just my pet hate at the moment)

Re:To clarify

[Anonymous Coward]

Not quite true. Network effects are interesting. If it gets big enough, it can be self-sustaining. e.g. the simplest solution is to only need a root server the first time you connect & establish root nodes in the network that act as anchors. Thus, you'd need all root nodes that you've known about + root server to disappear before you lost the network. Losing the root server might affect the addition of new members (but then you can pre-load the client with a root servers list or establish a rendevouz mechanism using a third-party service).

Re:To clarify

[toQDuj]

Cloud is just the latest word for "outsourcing". Although they do not want to call it that, because in many cases, outsourcing simply does not work or is not cheaper.

Re:To clarify

[Anonymous Coward]

I was just thinking aloud at methods that might not require any user action and work straight off without requiring a specific Tribler central server.

Any set of peers would do, so a central server is not actually needed to bootstrap. You could be seeded by any peer, using any active torrent. Sure, they probably include a seed file with the distribution, but it would just be for convenience. Without a central server, those seedfiles would expire, so the version you've downloaded 6 months ago wouldn't probably work without you finding a new seed/seed file/torrent. This is how it might work, I don't know how it actually works.

Re:To clarify

[discord5]

Bootstrap is the interesting issue.

It is indeed. The bittorrent DHT solution is based on Kademlia (or the BEP for Bittorrent specifically). If you google a bit you'll find a few papers and some interesting things, including attack vectors. I'm implementing a version of Kademlia at the moment to have nodes in a network find other nodes for accepting work in a distributed environment, and bootstrapping the thing is "the weakest link". You could set up multiple bootstrapping nodes, but suppose that a network failure takes out your access to the bootstrapping nodes you're basically humped when you want to join the network.

Sure, once a node is online and given enough other nodes stay online enough of the time, it would be possible to have a persistent network.

What you're referring to is known as "churn" and provided the network is large enough it becomes less of an issue. There's actually an interesting paper on churn in Kademlia, but I'm sorry to say I can't seem to find it anymore. As it turns out, Kademlia is quite resistant to churn provided there are enough nodes in the network. Not really surprising, but it's a nice read if you want to know how resilient the network is without having to test it yourself.

I suppose you could do something like search google for random torrents, join in, test the folks you connect to for being part of the decentralised network, grab network info from there etc. It still uses google as a central reference point but it would be more robust than having some sort of hard-coded 'peer tracker' server, or using any sort of brute-force port scan of the internet.

My current solution involves trying multicast and if that fails a broadcast to find other nodes if the bootstrap servers are down. This usually allows you to find other nodes on the LAN (depending on the network configuration) and once you have a single node in the network you can start doing lookups. Few admins like the idea of you "scanning" the network to find other nodes, understandably. Of course, this method of discovering nodes has disadvantages as well, for instance you could have a segment blissfully unaware of another segment in the network until the bootstrap servers come back online.

However for typical bittorrent use I doubt that this particular strategy will work. The work I'm doing has little to do with bittorrent, or how P2P is traditionally seen by the bulk of its users. I doubt that the implementation I'm working on is directly usable for another problem than the one I'm working on.

Re:To clarify

[Anonymous Coward]

(I swear I remember reading about some other client with P2P search a couple years ago)

The Gnutella network had distributed search; Fasttrack used supernodes to direct search queries -- still distributed, but less distributed than Gnutella. I'm sure there were other options too (Ares, Soulseek?)

Re:To clarify

[cyclomedia]

This

I've tried leaving certain torrents sitting there seeding and get zero leechers ever, presumably because I have bugger all bandwidth.

Re:To clarify

[kvezach]

There have been proposals to make use of partial matches (and papers about how to do so effectively), but to my knowledge, nothing has come of them.

Re:To clarify

[Anonymous Coward]

Just as you mentioned, brute force is the only other option available.

Ways to make this easier would be to separate IPs per owner, then region, then country, and so on.
Then you just travel up the sets from your IPs owner to IPs continent if you find no broadcasts. (or you can even manually select where to scan if you know the area the trackers are)
IP files get sent around so a network is built up over time and there we have a fully (traditional) serverless network.
Options to tweak the number of IP checks per period, max IP connections, and all the other usual stuff for brute force scanning.

Build some onion routing straight in to it and bham, harder to crack open as well, more-so if it ends up in mainstream software, which it might sooner or later because governments are being paid off by media companies to try and restrict such things. (doubly retarded, it seems)

Of course, this is open to abuse just like any other serverless system is: companies + lots of money = lots of rogue nodes with false information.

Re:To clarify

[Crayon Kid]

The problem with no centralized servers is that no one pays attention to ratio and the like.

Then they should implement the concept of "authoritative peers". Peers that have a secret key that gives them special powers over the swarm, such as kicking them out or regulating their participation.

It would also solve another problem with torrents: incremental updates. Today, if you want to do a minor alteration to one file out of 100 in a torrent, you have to publish a different torrent, tell all the peers to switch to that one and to start building up to 100% again. An authoritative peer would just push deltas on everybody.

Re:To clarify

[zero0ne]

Don't forget, all the hardcore private torrent sites will ban you if you enable all those extra "features" like DHT and PEX.

Re:To clarify

[Anonymous Coward]

Or one could check RIAA/MPAA/government server for which IP's are banned or otherwise shunned and search amongst those :)
Big media and torrents are meant to work together (big media just doesn't want to say that out loud)

Re:To clarify

[Anonymous Coward]

There are a lot of torrent clients out there. A lot. You can express the number as a percentage figure of connected machines, I bet.

Do you think the bandwidth involved in trying to get DHT info from 900 or so in order to bootstrap with a degree of likelihood that borders on certainty is a problem? I don't...

Re:To clarify

[Hatta]

It's probably more a lack of leechers than anything else. Private trackers reward people for seeding, so most of the stuff is overseeded. Most of them have occasional free leech periods or other programs to stimulate the credit economy, look into them. Generally it's only in the beginning that you have ratio troubles on a private tracker.

Re:To clarify

[mounthood]

Bootstrap is the interesting issue.

You can't have a situation with no server involved, ever, unless you're distributing the software on a friend-to-friend basis.

Not true. You could just send UDP packets randomly until you find a peer. It's a ridiculous strategy, but it could work for a large persistent cloud of peers. Optimizations could even make it practical (like remember 1000 peers when you close and try them when you restart.)

Re:To clarify

[Anonymous Coward]

No, incremental updates are a case of using the wrong tool to solve your problem.
Rsync is probably more appropriate than bittorrent, in this case.

Re:To clarify

[Anonymous Coward]

I've been thinking of the bootstrapping problem for quite a while now. If you have a case where a new user crash-lands in the middle of nowhere (but somehow with internet access), then there is no way to bootstrap that. Thankfully, none of us fall into that case.

There are really only two ways to bootstrap: Know who to contact already, or have someone nearby who you can ask for more information.
DNS bootstraps by having a pretty static set of 13 root servers defined: well-known IPs which stay in service for many years.
IPv6 bootstraps via link-local multicast ("anyone have this IP? Nope? It's mine!" and talking to the local router via that IP to learn the network and route(s) outside link-local)

DHTs, in general, use the first one: well-known places to hit up for information (such as the bootstrap routers hardcoded into bittorrent clients), or like emule, they give you an option to enter a bootstrapping node (acquired out-of-band, like your friend's IP or from a website)

Re:To clarify

[SilentChasm]

The Merkle thing still seems to deal with blocks of a whole torrent instead of blocks of files that are in a torrent. Similarly with the other, it seems to be more focused on a new hashing method involving trackers knowing what pieces a user has in order to hand out peers that better meet user needs (as in giving a user a list of peers that have the pieces they need, as well as peers that need what the user has).

Neither one really solves the problem of swarms being divided simply because someone else made a torrent of the exact same file. Unless they're file based I don't really see how they would.

eMule collections are nice in that (the plaintext ones at least) are just:
ed2k-link
ed2k-link
ed2k-link
etc

Re:To clarify

[Anonymous Coward]

No, all the hardcore private trackers will ban you if you use a client which lets you override the "private torrent" bit in the torrent file.
Well-behaved clients automatically disable DHT and PEX for torrents with the private bit set.
All public torrents make use of those functions just fine. No need to alter any settings in your client.

Re:To clarify

[kvezach]

It deals with redundant pieces of a whole torrent rather than with files, true. But say that you have two torrents: one contains file A and B, and another A and C. Then the rolling hash (and possibly Merkle idea, too) will detect that both torrents have a significant chunk in common, and let the users who need the A-portion connect to both torrents.

The advantage in comparison to per-file links is that it could also detect common chunks within files. If one of the torrents is an ISO of Debian i386, and another is an ISO of Debian amd64, the file-sharers can share the common parts of the ISOs with each other. It generalizes the solution to go beyond just file redundancy.

Re:To clarify

[Anonymous Coward]

Xeem it was called, if I recall correctly.

Ok, but.

[MrQuacker]

But how does moderation work then?

With a large public tracker like PirateBay there are mods and members who weed out and delete the malware, spam, and bad torrents that are on the tracker. Wouldn't a distributed system like this actually make it easier for "bad" content to get uploaded? Its like Limewire all over again.

The idea here seems to be that "you cant stop the signal". But I am not sure how they get around the fact that you don't have to kill the signal, just garble it.

Re:Ok, but.

[mysidia]

But how does moderation work then?

Throw in web of trust?

Send in $1 to get your 'reviewer' certificate signed by some trusted entity. Sort search results by number of signed positive reviews; and then number of downloads which "reviewer nodes" saw occuring.

Re:Ok, but.

[biryokumaru]

Well, if ten peers show up, maybe it's not so good. If ten thousand show up, maybe it's legit.

Re:Ok, but.

[MrQuacker]

That's a measurement of a torrents health, the seed/leech ratio, not quality. So while the 1000peer file might be done faster than the 10peer file, its a moot point if you're downloading static or malware.

Re:Ok, but.

[biryokumaru]

Except 10,000 people probably won't be retarded enough to just keep around garbage files that didn't work when they downloaded them.

Way to miss the point.

Re:Ok, but.

[Dunbal]

So if you want to shut the whole thing down, you go after the "trusted entity". Sort of destroys the point of decentralizing.

Re:Ok, but.

[mr100percent]

Too easy to fake. Look at how many limewire spam files are supposedly with 100+ peers

Re:Ok, but.

[cheekyjohnson]

Common sense and an antivirus software could be used to weed out a majority of the malware. As for the spam and bad torrents, I have no idea.

Re:Ok, but.

[Nikker]

You could still rate at the clients machine and propagate the torrent signature to all the other clients there by distributing the ratings in almost real time on already discovered client/hosts. Meta moderations could also be propagated to blacklist. Since this is added to the torrent you can sync hashes to ensure everyone is on the same page. YMMV

Re:Ok, but.

[Chuck Chunder]

You don't need a single "trusted entitity", a web of trust is based on your own prior experience and what others around you will vouch for.

If you have downloaded a torrent signed by someone before and been happy with it your software might be happy downloading more from them without warning. If you haven't seen anything from that person before your software might poll your peers to see if they will vouch for it and ultimately give you a choice one way or another.

Various key servers could be set up to serve trust information but would not present a critical point of failure or (for dodgy torrents) be at much legal risk because they wouldn't be serving anything remotely related to other peoples copyrighted information.

Re:Ok, but.

[Crazy Taco]

So if a botnet gets on with 10,000 users, then everyone on the Internet will believe it's therefore a legitimate file? Or if something new comes out and two people have it, so no one else downloads it, how can it become big enough to appear "legit"? There's no good way to solve these issues, so you are pretty much guaranted that there will be a flood of malware. Not that I care. I hate piracy. But the point stands.

Re:Ok, but.

[Pseudonym Authority]

Except they probably will.


Also, botnets.

Re:Ok, but.

[Anonymous Coward]

Digital Signatures

Re:Ok, but.

[Merls the Sneaky]

Antivirus is useless if the user doesn't even know what a file extension is, let alone a friggen .exe. People will click files labelled pop trash.mp3.exe.

The thing about common sense. Common sense simply isn't that common.

Re:Ok, but.

[Anonymous Coward]

Youre completely out of line!!!ok theres perhaps millions of people willing to give a credit card number for a mere $1.But most of those people I would call 'sucker'

      my opinion

kooky

Re:Ok, but.

[MrQuacker]

If you ever used one of the free trackers you'd find out that a lot of the fake/spam/malware files posing as movies or tv shows have thousands of fake peers. This is to trick people like you that think there is strength or knowledge in numbers.

So just because a file says it has 17543 peers, it really doesnt.

Re:Ok, but.

[Anonymous Coward]

There should definitely be a moderation system in place, and should be as easy to use as the facebook like, but of course extend it with a little more detail,like/dislike, or good quality/bad quality/malware

If there's no server, then I'm guessing that each peer will need to have his/her rating on his/her own computer within the torrent file you download? If you download the file from another which has already rated the file, either the rating is cleared, or invalid until you rate it and you overwrite the old rating.

I'd also love it if the program would discreetly remind you to rate the data you've downloaded.

If this is an option, then there's no real reason why we couldn't even go further with the facebook style, what if you could add certain peers as "friends" (both have to accept friendship status), and you could see what they're downloading, and you could add comments or tags (I'm guessing a database linking the torrent file with the comments on the users machine?), or grab the torrent as well, and so on.

Like this though, but it would be nice to have some kind of a rating system which makes sense.

Re:Ok, but.

[Kjella]

Didn't email and pgp prove once and for all that WoT and convenience is pretty much direct opposites? You don't get a straight answer, you get a bunch of complicated information and choices on who to trust. Also, many people need to do it in order for it to function and most people won't so people won't think twice about downloading a torrent with no trust. Those who try will have huge problems with malware makers and MAFIAA goons trying to poison the well. Not only can they create as many fake identities as they want, for example they can upload versions that aren't obviously borked, people will vote them good at first then get marked bad by others even though they were honestly trying to contribute. They can give good trust information a while making others trust them, then turn into crap so you need some form of trust revocation.

All in all, it's ugly to make it work. Not that it can't be made to work technically, but those who manage to work with this probably have no problem finding good sources of material anyway. But you'd lose 90% of the file sharers who'd probably just be lost and download just as much crap as on a trustless P2P network.

Re:Ok, but.

[Dr_Barnowl]

The experience with email and PGP just proves that most people are boneheads. The only places I see PGP (well, GPG) signatures are on software development mailing lists.

"email" is the wrong metaphor. It misleads people into believing their messages are secure, because they are used to their mail being enclosed in an envelope - just like the one that basically every email program with a GUI depicts somewhere. It's more like a postcard. One that gets delivered via a network of disreputable postmen, some in the employ of enormous and sinister organizations, some just out for themselves. Encryption is the envelope, but people are happy to just keep sending postcards.

I have the right tools installed, but honestly? The only time I even *sign* messages is to get a little kudos on those developers mailing lists. I don't bother encrypting mail because in general for the vast majority of recipients you're going to have to

• Explain to them why email is insecure (it can be hard to persuade them of this)
• Find PGP compliant tools that integrate with their chosen email solution (since most people use webmail, this is going to be hard)
• Talk them through installing the tools
• Talk them through key generation
• Talk them through key management, distribution, and trust
• Explain to them that yes, if they forget their password, they are *never* getting those mails back
• No, I can't decrypt them, even though I encrypted them, because they are 'crypted with YOUR key
• Explain public-key crypto

If your chosen correspondent doesn't give up at some stage of this process, then you are either very persuasive or you have something important to hide. Or they are technical. It's an excellent measure of how much people value their privacy - most do not value it enough to have to expend significant mental effort to protect it. It remains to be seen how much this perception will be effected by the ongoing corporate harvesting of what would, in a "real world" social network, be considered private material.

Re:Ok, but.

[Anonymous Coward]

I find Freenet's built in message board a nice exercise in WoT. Essentially, with complete anonymity of users, their first version got spammed to hell and became unusable. Once they introduced a mandatory WoT, it started to work. Anonymity was the factor that drove both developers and users to embrace it.

Re:Ok, but.

[AI0867]

You can automate the trust, or rather, make it implicit.

Whenever you favorite a torrent, you place some trust in both it and the publisher. That gets propagated.

I don't know the specifics, but I do know that people at my university have worked on this for years. (tribler isn't new, this is just a new release that has the P2P search)

Re:Ok, but.

[moonbender]

If you're at a point where you need a botnet to make a spoofed file look legit, you're doing pretty well. Using a botnet is a lot of effort for fairly little gain. However, AFAIK there are (or were) other ways of faking the number of peers without resorting to "brute force". Still, looking at the number of people associated with a torrent is a moderately good metric of its quality, though it's obviously not the only.

Re:Ok, but.

[mysidia]

So if you want to shut the whole thing down, you go after the "trusted entity". Sort of destroys the point of decentralizing.

It doesn't matter if a 'trusted entity' gets taken down for their mere act of revealing which items are what they claim to be.

Their existing reviews are already out there. If their keys become compromised, this gets reported by others, so people know to stop trusting the compromised entity.

Re:Ok, but.

[kent_eh]

Or if only 10 peers show up, it's obscure and not well seeded due to a lack of popularity...
Like many of the things I look for. (video of Zappa concerts from the '70s, old Canadian kids TV shows, etc)

Re:Ok, but.

[nabsltd]

With a large public tracker like PirateBay there are mods and members who weed out and delete the malware, spam, and bad torrents that are on the tracker.

Both private and public trackers are filled with "bad" torrents and malware, and there's usually a lot of spam in the torrent comments.

There are sites with search engines plus private trackers who claim "no malware", yet right in the comments of some torrents there are instructions about running the downloaded files in a VM because of the trojan/spyware/whatever.

Re:Ok, but.

[mysidia]

So if you want to shut the whole thing down, you go after the "trusted entity". Sort of destroys the point of decentralizing.

Can you point to a law against publishing a list of filenames, SHA1 checksums, and 5 or 6 "Yes/No" questions?

It could be a service with another primary purpose, for example "Identification of unknown files"... common files on computer systems could be identified/reviewed by unrelated parties.

The same tech could be used not only for P2P networks, but for "web safety"... for example, you download a file from any website.. 6 months later you forget what it is... have a 'metadata service' to find more information about the file for you, based on a large mass of checksums, descriptions, and "original filenames"

XXXXX.xyz xxxxxxxxx
"Is the filename meaningful?"
"Does the filename accurately represent the nature of the content?"
"Is the quality of the content reasonably acceptable?"
"Is the content complete, without undue truncation, omission, or error?"
"Is the item an advertisement or sales pitch?"
"As far as you can tell, is the file with this SHA1 sum free of viruses/malware?"
"Does the item seem to be free of unnecessary or content not disclosed by the filename or description file inside the archive?"

Re:Ok, but.

[cheekyjohnson]

Then that's not really a problem with the software, now is it? It's a problem with people who are too lazy and idiotic to at least learn not to run random executable files on their computer.

Re:Ok, but.

[Anonymous Coward]

This actually works really well for PD, a relatively new anonymous p2p network. I joined the network to download obscure movies and stuff I couldn't find on bittorrent and remember thinking it was going to be like gnutella / kaza all over again. What I found is that the quality is just as good as bittorrent. I think it's because there's a built in rating and signature system. You also don't get those files that have 50 keywords in them because the way PD works is you have your normal file and then a bunch of keywords associated with the file only by the client software.

It's also a lot faster than gnutella ever was. The reason I think is because bandwidth is "distributed" among all the peers. On gnutella you could only upload stuff you'd already downloaded so if you download a lot of unpopular stuff you don't really contribute a lot back to the network (your upload capacity is never used if noone wants the same files). On PD it has a cache of lots of pieces of files that are in demand so no matter how obscure your downloads are, you're always using your upstream capacity to help the network. Obscure content is also kept longer (forever, even) because it gets spread around on peers that would normally not download it... unlike torrents which "die" given enough time.

I still prefer torrents as my primary download method because of the community surrounding torrents (fansubing etc) but I think in theory distributed networks like PD can be more efficient.

Wow. p2p is turning net into a huge cloud

[unity100]

this, dns-p2p, and etc are turning the internet into a truly decentralized, uncontrollable, REAL cloud as it should have been from the start.

i, for one, am not suprised that the ones to save net freedom, are ending up being people who have been accused of piracy. after all, if it is not detrimental to the control of private interests, why villify something in mass media, right ...

Re:Wow. p2p is turning net into a huge cloud

[Anonymous Coward]

this, dns-p2p, and etc are turning the internet into a truly decentralized, uncontrollable, REAL cloud as it should have been from the start.

Not really. The physical topology of the internet's infrastructure is not close to that of a decentralized cloud.

BitTorrent is a really inefficient protocol, as it reduces use of the biggest "pipes" at the cost of increasing use of the smallest "pipes". The only reason it's been so successful is that it allows content provides to reduce their own costs, even though by doing so they are increasing the costs of others by far more.

It's similar to the classic Tragedy of the Commons in that everyone doing what's best personally for them makes everyone suffer.

Re:Wow. p2p is turning net into a huge cloud

[Anonymous Coward]

Only at the application layer, unfortunately. There's still a definite hierarchy of autonomous systems handling the border routing. Mesh networking is the only way to really decentralize, and that is just not as fast. Luckly, it's getting there. However there will always be big pipes to plug in to to get to where you want to go fast. And since most of those are going to be owned by a big corporation or a government, well, there's always encryption.

Re:Wow. p2p is turning net into a huge cloud

[Nikker]

If we were able to have higher throughput behind the ISP the bit torrent protocol would be much more efficient but that's only because of the low ISP > Client ratio. If there are 1M users on any ISP it would be quicker getting what I'm looking for from one of them rather than the actual source. Which is also why caching on large scale networks are beneficial.

Re:Wow. p2p is turning net into a huge cloud

[Anonymous Coward]

Uncontrollable? The IP addresses of everyone in the torrent are still perfectly available by joining, just as they were before. This changes nothing.

Mod parent -1, overrated.

Re:Wow. p2p is turning net into a huge cloud

[icebraining]

Why can't you do it on an application level, based on allocated IP lists? Back when we had a distinction of national and international traffic here in Portugal (international capped to 10GB, national unlimited) somebody made a eMule fork called Blowfish which had an IP based filter, and would let you download only from you isp or nationally.

The same system could be used to give priorities instead of simple blacklisting.

Re:Wow. p2p is turning net into a huge cloud

[ScentCone]

people who have been accused of piracy

Not that any of them actually do ever rip off anyone's work or spread it around, of course.

This whole thread is full of comments tap-dancing (badly) around the fact that the real driver here is the ability to avoid getting busted for ripping off copyrighted works. Yes, the technology is compelling (if, as described, a huge magnet for malware and bot-net bullies running the show), and has legitimate uses. But this particular twist on the torrent landscape is driven hugely by the fact that central servers pointing the way to legit, legal distro torrents and freely released entertainment, etc., isn't what millions of people are actually trying to get their hands on.

On that front: I'm not sure how this helps the guy who wants to be harder to find as he rips and distributes pirated content. If it's robust enough for him and his million best friends to spread around a feature film ISO ten minutes after it's been burned, then it's robust enough for the copyright holders to see it all happening and write down IP addresses, too.

asics and mac

[adelamanda]

moncler vest, moncler shiny thomas, moncler bulgarie, moncler vest moncler shiny thomas moncler bulgarie asics gel shoes running shoes asics asics running asics running WOMEN MINI COOPER women Whizzer Lo Kid women Mexico Mid Runner http://blog.19lou.com/index.php?action-userlink-linkid-187882 http://www.asicsshoesshop.com/ our email:asicsshoesshop@hotmail.com ASICS GEL-STRATUS 2.1, Kanuchi, mexico 66, mac makeup, mac brush, http://www.mac-wholesale.com/ our email:mac-wholesale@hotmail.com asics gel shoes asics running 180 color eyeshadow mac makeup

and a torrent that hasn't disabled DHT/PEX...

[SuperBanana]

Most clients support DHT and PEX, and have for some time. You just need a single peer to bootstrap yourself, and you're good to go.

You also need to have a torrent that doesn't disable DHT and peer exchange. This has to be the most irritating "feature" of BitTorrent; the torrent file controls whether they're allowed. I understand why it exists and it'd be OK...if tracker administrators didn't abuse the hell out of it. I've lost track of how many times I've found a fairly new torrent that had a dead tracker.

Re:and a torrent that hasn't disabled DHT/PEX...

[Anonymous Coward]

Most clients support DHT and PEX, and have for some time. You just need a single peer to bootstrap yourself, and you're good to go.

You also need to have a torrent that doesn't disable DHT and peer exchange. This has to be the most irritating "feature" of BitTorrent; the torrent file controls whether they're allowed. I understand why it exists and it'd be OK...if tracker administrators didn't abuse the hell out of it. I've lost track of how many times I've found a fairly new torrent that had a dead tracker.

It's a valuable feature if you want to allow private trackers to operate.

It's the fault of the site that disables it if they're going to disable DHT and not provide any incentive to actually seed.

at some point

[oliverthered]

at some point all digital media will be available for unlimited download by subscription, and eventually via tax. [at least for the people]

A bit like the library, or maybe youtube (as in you pay for the goods that advertise of youtube ergo you pay advertising 'tax' on the goods)

If you can download it for free from youtube, how many times are you going to watch the same damb thing, is it really worth making everyone download it every single time?

Excellent Work You've Invented Gnutella

[phantomcircuit]

Giant waste of time, bittorrents benefit is from the community bitching about bad torrents, you cant do that without a web of trust or a trusted third party.

Re:Excellent Work You've Invented Gnutella

[aiken_d]

Are you kidding? Once I get my hacked client together to return Rick Astley videos for every search any peer does, there will be even more complaining.

Re:Excellent Work You've Invented Gnutella

[TheScreenIsnt]

Giant waste of time, bittorrents benefit is from the community bitching about bad torrents, you cant do that without a web of trust or a trusted third party.

But torrent search sites are neither webs of trust nor trusted third parties, right? They assist in separating the wheat from the chaff simply by allowing the wisdom of crowds to accumulate through posts linked to a torrent file.
Seems to me that this "what are people saying about this torrent?" functionality is consistent with complete decentralization; the community can in principle be built around the client instead of some site. Who hosts the discussions? Everyone who participates in them. Seems as cloud-able as anything else. Mods *can* be replaced by user ratings and associated karma, I believe.

Re:Excellent Work You've Invented Gnutella

[QuoteMstr]

Mods *can* be replaced by user ratings and associated karma, I believe.

And how, pray tell, do you differentiate legitimate users from malicious ones? Karma works because a trusted third party maintains it.

Re:Excellent Work You've Invented Gnutella

[Anonymous Coward]

Giant waste of time, bittorrents benefit is from the community bitching about bad torrents, you cant do that without a web of trust or a trusted third party.

I would like to appreciate the great work done by You
Generic Viagra Viagra Online Cheap Generic Viagra Generic Cialis Generic Levitra generic propecia Generic Viagra careprost Lumigan

Re:Excellent Work You've Invented Gnutella

[TheScreenIsnt]

My *hypothesis*, (note "I believe") is that users themselves can serve this function by rating comments. Pretty simple, really. You can imagine any number of integration schemes that connect these ratings to a reputation value.

Sure, you can be pessimistic and say "never work", but the sociology of online communities is a young science. An anarchical web is important to me, so these questions are worth my energy and optimism.

Re:Excellent Work You've Invented Gnutella

[Anonymous Coward]

I think you're targeting the wrong community here, mate. There is very little use for Viagra in these parts, and that's not because the males are especially healthy.

Re:Excellent Work You've Invented Gnutella

[Anthony Mouse]

My *hypothesis*, (note "I believe") is that users themselves can serve this function by rating comments.

I believe the point is that someone has to administer the rating system. Otherwise what stops spammers from creating a million accounts and then rating their spam as legit and legit stuff as spam?

There are distributed ways to do this though, e.g. trust users who behave like you do: Have every client generate a public key and then sign the message with the corresponding private key when the user marks a message as spam or legit. Then you flag stuff as spam or legit yourself and your client keeps track of which keys signed a flag for things you flagged the same way, and by this method your client knows which keys to trust when viewing new data which you haven't reviewed yet.

Been done along time ago

[bobjr94]

What about apps like emule ? They have been around for almost 10 years now and have no centralized database or tracking

Re:Been done along time ago

[Truekaiser]

been poisoned to near death by people who are paid by media company's to fill them with fake files and Trojans.

what would be cool

[oliverthered]

youtube, DVB etc.. plugins for instance ripping or downloading or recording and sharing.

point, rip and share.

maybe some darknet intergration,

voip, IM integration.

record that call with you boss and blast it out over the internet.

Still IP data available

[elashish14]

There's still the fact that IP data is available. Any user on the network will be broadcasting their activities making them vulnerable. Protecting users' anonymity is just as important as decentralizing any part of the network. In my opinion, this is the most important aspect of P2P that needs to be fixed. Not that I have any novel ideas on how that can be done....

Re:Still IP data available

[Anonymous Coward]

Onion routing amongst every user. People torrent over Tor all the time, which is absolutely killing the system for the uses it was intended for initially.

But if you built onion routing straight in to torrenting clients, problem solved.
Yes, you will have to deal with the "rogue node" stuff, and it could become an even bigger problem if this were to happen, but that is the only solution i can think of.
Encryption, of course.

Of course, doing this just aids terrorists, pedos and the "enemies of freedom"(...), so BAN IT, BAN IT HARD, ITS EVIL, IT WILL RAPE YOUR CHILDREN AND BLOW YOUR HOUSE UP!

Re:Still IP data available

[rich_hudds]

Everyone I knows just has a dedicated VPN from a different country to their own. With a firewall setup to only allow bittorrent traffic through when the VPN is connected you're pretty much anonymous without faffing about with TOR. You can get a dedicated VPN for about £5 a month.

Re:Still IP data available

[Asic Eng]

Maybe each client could accept requests to transfer bytes from another client. The requests would have a format like "please get me byte x for torrent y from peer z". When you download a file, then for each byte your client would randomly pick another client to make the request.

This way if you serve a file you'd have no idea who actually downloads it - you'd only get requests from random clients which are not actually downloading the file.

You can further complicate this by not making the request directly, but instead add a counter to the request. Each client decrements the request counter by one and forwards the request to another randomly chosen client. Only when the counter is zero will it actually request the byte from the server.

Of course that would add a lot of overhead - probably better not to make it on byte boundaries, but rather use larger blocks. Also the file server isn't really protected. I suspect there are probably lots of better techniques available already.

Re:Still IP data available

[cpghost]

Not that I have any novel ideas on how that can be done....

How about the protocols underlying Freenet, GNUnet et al? There are actually plenty of anon-P2P proposals and prototypes out there, they're just not as popular yet.

The future.

[onefriedrice]

It's becoming ever so popular to complain about ICANN or otherwise feel that a decentralized internet is the solution to our problems. I'm not a prophet, but even I can see the future on this one. The ones who will benefit the most from a completely decentralized DNS and/or P2P system are the ones who control the biggest botnets within the network. The rest of us will be so inundated with garbage that the internet will essentially become completely useless.

That's not to say that ICANN and especially the RIAA et al. aren't problems, but I don't see this becoming a viable solution. So I'm a skeptic, for now.

Re:The future.

[Jane Q. Public]

And you make exactly the same mistake that RIAA and friends make: just because criminals will use it, does not mean that ONLY criminals will use it.

Criminals have always been quick to pounce on new technology. That's no excuse for killing the technology so that nobody else can use it. I doubt very much the picture is as gloomy as you make it out to be. And even if it is: prior restraint is not the answer.

Re:The future.

[Pentium100]

No, centralized control of the internet is a bad thing. Also, why should the US be in control, why not $VeryReligiousMuslimCountry, China or North Korea? I'm sure they would like to shut down some sites too.

And botnets can cause problems in the current situation too. However, I still think that properly implemented decentralized DNS is a good thing. A completely decentralized P2P system that's actively in use will make torrent sites obsolete and make it harder for US companies to take down the files.

We have garbage even now, but currently it's harder to get around it, since the garbage is part of the centralized system (domain takeovers of WikiLeaks, some torrent sites etc).

Criminals can abuse almost any technology, but that does not mean that everybody else should be prevented from using it:
Do you use encryption to do your banking? Do you know that terrorists use encryption too?
How about anonymous networks (tor etc)? Terrorists also use them.
A knife is useful to cut food. It is also easier to kill someone with a knife than just bare hands.
A car is useful for going long distances. It can also be used to deliver illegal drugs or run someone over.

Re:The future.

[icebraining]

Criminals can abuse almost any technology, but that does not mean that everybody else should be prevented from using it:
Do you use encryption to do your banking? Do you know that terrorists use encryption too?
How about anonymous networks (tor etc)? Terrorists also use them.
A knife is useful to cut food. It is also easier to kill someone with a knife than just bare hands.
A car is useful for going long distances. It can also be used to deliver illegal drugs or run someone over.

That wasn't parent's argument - it was that it would be impossible to use due to being flooded with crap. And it does happen, just look at eDonkey networks. I stopped using it when 6 in every 10 files were fake.

Re:The future.

[ArsenneLupin]

cryptographic signatures

Re:The future.

[Pentium100]

Interesting, a few months ago when I used eMule to find a certain song, I managed to get the real file on the first try. Actually, other than the download speed, eMule is quite OK, especially for rare files (ones that nobody bothers creating a torrent for).

So what's missing?

[Chuck Chunder]

I don't think it's a waste of time in the sense that it's a step in a potentially useful direction.

If you want to add trust capabilities to the mix there are non-centralised ways to do it such as allowing digital signatures.

What's the difference?

[Scrapz]

So, how does this make it any different from Limewire, or Napster? I understand that there's the added benefit the swarm, but... really, how is this any different? It'll get driven to the ground with junk and viruses. I'll stick to my private tracker, TYVM.

nice

[VincenzoRomano]

And once you'll get cryptography enforced and a few more TCP/IP tricks I wonder how riaa will stop p2p.

Re:nice

[Magada]

Outlaw crypto and those TCP/IP tricks. Make possession of software that uses these technologies prima facie evidence of conspiracy.

Re:nice

[thijsh]

You have been disconnected from the interwebs until further notice for suspected possession of copyright infringement technology.
Infringing technology detected on your PC (by method of our complementary rootkit) includes (but is not limited to):
- FileZilla
- Putty
- TrueCrypt
- PGP
- And last but not least: mTorrent (the evilest)

We hope you enjoy your offline existence banned from the interwebs.
Frankly, we think you got off easy and deserve much worse you terrorist pirate scum.

Regards,
MAFIAA lawyers

P.S. Hahahahah, suckers, you thought this was a game... Thanks for helping us get all this $$$$$ at the price of your freedom!!!

Re:nice

[Magada]

Yup. Something like that. Only it would be local police coming to arrest you and impound your computers, acting on an anonymous tip from concerned citizens in the RIAA.

Re:nice

[cpghost]

You have been disconnected from the interwebs until further notice for suspected possession of copyright infringement technology.

You mean, possessing a computer (smartphone etc...) will be forbidden in the not so distant future? Makes sense. In their distorted world view, all that would be allowed is a device that can't be programmed by non-authorized personnel. Imagine CS students and programmers needing a special and expensive MAFIAA/Government license + frequent audits to be allowed the extraordinary privilege of writing code!

Re:nice

[Magada]

Pretty much what Stallman was writing about all those years ago, when the technology to enforce such a draconian regime didn't exist yet.

Of course, most everybody labeled him a wingnut.

p2p web pages

[EricX2]

Maybe I'm behind the time with this type of technology, but is there a way that this could be used to access web pages like wikileaks to prevent servers from being shut down or blocked by dns?

Good Idea, Bad Implementation

[CobaltBlueDW]

A distributed search feature is great, but the Trible software is bereft of features. I recommend sticking with uTorret, and hoping uTorrent developers implement a similar feature.

Indulge me with the answer to this, please...

[Anonymous Coward]

Legitimately speaking, why exactly would one require a search engine inside a bittorrent client? I mean, if you want a torrent of something that's actually on the up-and-up, couldn't you just google for the website that seeds it, and click on the torrent link at that site?

Re:Indulge me with the answer to this, please...

[Magada]

Because you can't rely on google or the website being there when you need them. The Wikileaks conspiracy is a case in point. Their DNS provider, their money transfer companies and their hosting company tried to make them disappear. So far, Google is working as intended, but for how long? Also, organizations with fewer resources might wither and die under such attacks.

Re:Indulge me with the answer to this, please...

[taucross]

Better yet, couldn't you just include the torrent hash encoded in base64, upload to a free website, let Google cache it and have them act as the indexer? It's not like they can do anything illegal.

Re:Indulge me with the answer to this, please...

[mark-t]

If the website with the legitimate torrent doesn't have links to infringing content, what reason would there be for someone to try to make it disappear?

There's plenty of websites that provide non-infringing content via torrents, yet I cannot think of a single one where the necessity of a dedicated torrent search engine would be remotely necessary in order to find it.

Re:Indulge me with the answer to this, please...

[Magada]

How about Wikileaks? It's hosting content that has NOT been declared illegal anywhere in the world. Yet it is under constant attack.

Re:Indulge me with the answer to this, please...

[mark-t]

A bittorrent search facility would not, in any way, alleviate any of the issues that wikileaks may be having. While they may certainly benefit from using bittorrent for content distribution, how does having a specialized bittorrent search facility assist in that regard? What legitimate content can you find with a specialized bittorrent search engine that you can't find at the web site that would have seeded the content in the first place (which is generally easily findable with a generic web search engine), other than artificially contrived examples that are little more than a proof of concept of something that nobody has any real reason to utilize because there are other equally practical options available, but the seeders of those files are more hung up on the proof of concept than they are on actually distributing the content?

And hey.... the very nature of wikileaks seems to be almost specifically orchistrated to make certain types of people nervous. Whether or not such people deserve to be made to feel this way is irrellevant. How does, for example, Slackware having a torrent of its latest Linux distro cause anybody else to get on edge? And of course... you don't exactly need a torrent search engine to find it either.

Re:Indulge me with the answer to this, please...

[Magada]

Having a torrent search engine would do away with Google and the Web altogether, when it comes to distributing content.

Your attitude towards Wikileaks notwithstanding, it's quite obvious that certain types of currently-legal content are being or can be suppressed.

Let's take your example: Slackware is not illegal today. If Microsoft decides it's time to kill linux, they can start a patent lawsuit, seeking an injunction against distribution first.

If an injunction is granted, Slackware would become illegal to distribute before the trial even starts and would continue to remain so for the duration (or even afterwards).

Rinse. Repeat. Microsoft has hundreds of patents it could claim to have been infringed upon. The claims do NOT have to be solid. Just solid enough to go to court.

Are you willing to do without Slackware, if Microsoft so wishes?

Re:Indulge me with the answer to this, please...

[mark-t]

You've sidestepped my point again... that there is absolutely no legitimate content that even remotely necessitates having a search engine specific for torrent files, other than cases that have been artificially contrived to prove a point and where the seeder cares more for proving that point than they do for actually distributing the content, because there are probably at least a dozen other ways that would have been possible to find it without using a bittorrent specific search engine that could have been just as practical, and far less likely to be associated with illegitimate activity, even though the actual material may not be illegitimate itself.

Re:Indulge me with the answer to this, please...

[Magada]

Excuse me for being blunt, but I am beginning to suspect obtuseness on your part.

Your principal claim is
"that there is [...] no legitimate content that [...]necessitates having a search engine specific for torrent files other than cases that have been artificially contrived"

Not so. Depending on your definition of legitimate, of course.

Case in point:

The movie "The Departed" is banned in China.

Please tell me, what would be easier, if you want to see it and are living in China?

Having a search facility within your torrent client, or trying to circumvent the various technical means that the Chinese have for censoring the web, using no more than a browser?

Re:Indulge me with the answer to this, please...

[mark-t]

Uhmm..."The Departed" is a copyrighted movie, not freely redistributable content.

It's ironic that the example you've given as a supposed legitimate example isn't something that people can legally distribute via bittorrent in the first place anyplace that Copyright law is upheld, China or otherwise.

Re:Indulge me with the answer to this, please...

[Magada]

Ironic, you say? I find it more ironic that, although nobody can legally buy "The Departed" in China, you still whine about copyright infringement.

Want an example involving public domain stuff? How about any Falun Dafa video then? Stuff about Tibet or the Dalai Lama? Anything about Tiananmen square?

You sound like a hacker. Time to get with the program, really. I'm aware of freenet, but I'd rather not have to host CP just so I can safely explain what absolute tossers my country's leaders are.

Re:Indulge me with the answer to this, please...

[mark-t]

It may not necessarily be easy to get access to such information with China's borders, but you've failed to illustrate how having a bittorrent specific search engine would make it any easier to find it. The only thing you've given an example where having a bittorrent specific search engine would be advantageous is with copyrighted content that isn't supposed to be legal for anybody to be freely distributing... comparing that to censorship is a bit unreasonable because copyrighted material *IS* readily available, you just have to pay for it.

Of course, as for the notion of comparing copyright to censorship, I'd like to point out that content under the former is legitimately available, even if for a cost, while content subject to the latter is not... at any cost (although having money may be helpful in acquiring it, acquiring the censored material is still against the law).

It is not my position to justify censorship, by the way... but you introduced the subject into this when I was talking about copyrighted material. It's merely my intent to show that copyright and censorship are not the same.

Re:Indulge me with the answer to this, please...

[Magada]

The only thing you've given an example where having a bittorrent specific search engine would be advantageous is with copyrighted content that isn't supposed to be legal for anybody to be freely distributing... comparing that to censorship is a bit unreasonable because copyrighted material *IS* readily available, you just have to pay for it.

I specifically point out "The Departed" is NOT legally available in China, at any cost, and you come up with this claptrap.

I then point out examples of public domain works which are being censored and which could benefit from a distribution network that can neatly bypass the Great Firewall of China and the dependence on Google's censored search results altogether. You choose to ignore me.

You, sir, have been wasting my time.

Re:Indulge me with the answer to this, please...

[mark-t]

You're missing my point... again.

Although the public domain examples you cited could benefit from a distribution system such as bittorrent, none of those examples factually *DO* benefit from a bittorrent specific search engine, even within China's borders. I asked you for a factual example, not a theoretical one... and the only factual example you could list was one that isn't freely legally redistributable in the first place.

I'd ask you to try again, but as you feel responding to me is a waste of your time, I guess we're done.

Re:Indulge me with the answer to this, please...

[Magada]

none of those examples factually *DO* benefit from a bittorrent specific search engine

That's just because such an engine does not exist yet.

Are you arguing that the types of content in the examples I gave you are not being distributed via torrents and so would not benefit from an in-client search engine? 'Cause if you are, you are dead wrong.

Here is a Google query that you cannot perform from within China. There are many others, of course.

Re:Indulge me with the answer to this, please...

[mark-t]

First of all, you did not use a torrent specific search engine to find that. Secondly, I cannot help but wonder if you really could not possibly find that information within China without a torrent-specific engine. I'd be willing to bet that you could... you just might have to get a bit more creative than just using google (after all, we are talking about government censored information here). Finally, there are many torrent search engines... it's just that they all search specific torrent sites, and most are not truly peer to peer (there are some peer to peer search engines, but I do not know of any that are bittorrent specific). Further, the amount of porn and spam and other tripe on the services that do use peer to peer search engines is so high that most legitimate content ends up getting lost in the noise, and given that historical precedent, I don't think it's very likely that a future bittorrent specific peer-to-peer search engine is liable to help matters.

Regardless... my main point remains... that to date, no truly legitimate purpose has yet been found for which a torrent specific search engine would actually be beneficial other than artificially contrived examples created by people who are more interested in proving a point than distributing the content, or else simply hypothetical situations.

If this should ever change, as you seem to believe it can and will, I'll be happy to admit I was wrong. Somehow, however, I don't think you or I are ever going to see such a time.

Re:Indulge me with the answer to this, please...

[Magada]

No true Scotsmen, eh?

big whoop!

[Cyko_01]

shareaza has been doing this with torrents for years! complete with file ratings and comments

Not new

[Per Wigren]

From what I can see, it's pretty much OneSwarm, but without the anonymity.

Great!

[dorinmouss]

It is another example of the development of torrent which will never die though a lot are fighting against it.

Sounds very remenicant of Kazzaa/limewire

[lastrogue]

Dunbal already mentioned this, but I didn't see anyone commenting on this. Isn't this just LimeWire or Kazzaa in another skin? Seems kinda like a good way to spread crap malware. Is my logic flawed?

Democracy needs P2P

[chipwich]

Thomas Jefferson said, "Information is the currency of democracy". The WikiLeaks drama is showing us how readily our own politicians will abandon core values of democracy in order to avoid embarrassment. It also clearly demonstrates that we live in a world where our personal communications can readily be disrupted at the whim of private corporations under pressure from these same politicians. The entertainment industry has tried to criminalize peer-to-peer technologies for years, but what is happening with WikiLeaks makes it more essential *now*, than ever before, that we adopt open source peer-to-peer technologies on a large scale. Perhaps the most important of these is The tor project which permits private and anonymous communications. Democracy cannot exist if people cannot speak freely without fear of reprisal. The more TOR relays that exist around the globe, the more immune we all are to the government/corporate censorship we are witnessing. Do your part in ensuring your digital rights by running a relay and becoming part of the network.

Bittorent Without tracker...

[mrops]

...welcome back to 2005 and enjoy Gnutella

Re:Bittorent Without tracker...

[wealthychef]

Searching is nice, but you also need community ratings to identify trojans/spam/crap/excellence and reward uploaders with props to encourage them.

But wasn't this problem solved before?

[Rexdude]

Before torrents, there was Kazaa, Gnutella,Limewire and eDonkey. (They still exist). All of these support searching peers from within the client without any intermediary website. I always wondered that bit torrent seems like a step backward - since it relies on websites and trackers that can be shut down, or seized and have the users traced from the logs.
So why is BT so popular as compared to the earlier services? Is it a more efficient P2P protocol? After all you still require a client to download and you still need to open ports on your firewall to allow traffic, same as with the others.

Re:But wasn't this problem solved before?

[SilentChasm]

I think it's because of the number of downloads at a time.

You typically have only a handful of torrents running vs many files on the other networks.

One at a time downloading: you see nice fast speeds. Lots of files downloading: you see slow speeds all around even if you're going at the same total speed for all files as torrenting would.

That and the somewhat verifiedness you get from a torrent you get from a trusted source as opposed to searching in the other clients. The thing I don't get is why people assume that just because the search is there, it's the only thing that can be used. I would argue ed2k/magnet links and the like are easier than torrents in that they are just links rather than files. You click on the link, it downloads. Find a trustworthy indexing site (comparable to a bittorrent indexing site) and you've got a fairly reliable system that doesn't go away when the tracker does.

It's also not limited to the people who downloaded the exact same torrent as you but to everyone who is looking for a file with the same hash. Why did they use blocks instead of file hashes in bittorrent?

Torrents don't seem to last as long either. They start out fragmented and, rather than sharing everything they have, only a few are active at a time. It's worse for the general availability of files. Per torrent ratios I think mess things up: uploading to get a 1:1 on a file with 1000 seeds is not nearly as important as uploading anything on a file with 0.

Upload practices

[KingAlanI]

The vast majority of my uploading has come about with the torrents I started, where it *has* to be me providing the uploading bandwidth, even at my crappy ~50KB/sec. (I get better upload bandwidth with garden-variety HTTP uploads) Talk about getting in on the torrent early.

For my rare torrents, I'm still a useful seeder, but I've not really needed after that for the big ones. Similar applies for the stuff I download.
Nevertheless, I keep the torrents around (unless it's something I want to relocate to fit my sorting system). However, I sometimes am only seeding the new and/or rare stuff (which I mark in uTorrent with the 'Initial-Seeding' flag for easy classification, not mention that Initial-Seeding actually does get my new stuff out a little faster)

initial distribution

[KingAlanI]

Would distributing the .torrent's in usual person-to-person manners fall under "on a friend-to-friend basis" as you phrase it? Sure, a server might technically be involved (from the email site or whatever), but I don't think that's what Tribler meant, as you would agree with your example of Google as the server. And not necessarily (put only the .torrent in a sneakernet, etc)

Yeah...

[KingAlanI]

Even if the initial seeder isn't trying to be malicious - torrenters make mistakes too, and TPB comments once let me know about some technical issues with a particular file, which I did get fixed.

Health tip?

[Anonymous Coward]

The only weird old tip this guy has is the one in his pants.

Re: Torrent clients disable DHT/PEX...

[xiando]

You also need to have a torrent that doesn't disable DHT and peer exchange. This has to be the most irritating "feature" of BitTorrent; the torrent file controls whether they're allowed.

It is true that there is a flag in .torrent files which can kindly asks the client to disable peer exchange. Some BitTorrent clients will disable peer exchange is this flag is set, many do not care at all and exchange peer information anyway. It is in any case the client software you are using, not the torrent file, which actually disables DHT...

Regarding churn

[Anonymous Coward]

Since neither of us have the paper handy,
In a nutshell, Kademlia isn't affected much by churn because its buckets of known nodes prefer long-lived ones.
If I (as a Kademlia node) know that you exist (I have you in a bucket), you will never be displaced by a newer node unless you become unreachable to me.
The paper found that the longer a node has been around, the better chance it will STILL be around in the future. Not exactly rocket-science, but it's a handy bit of knowledge.

cryptographic reputation

[Anonymous Coward]

a system whereby you trust an entity based on a greater trust of some other closer entity who claims to trust them .