Slashdot Mirror
The Uncertain Future of BitTorrent
javipas writes "The people behind the popular BitTorrent tracker are working on a new version of the BitTorrent protocol that could become the successor to the current one, maintained by BitTorrent Inc. The company founded by Bram Cohen — original author of this protocol — now has decided to close the source for several new features in the BitTorrent protocol, and this "gives them too much power and influence". The new file format would be called .p2p, and would maintain backwards compatibility with current .torrent files."
You can give OSS to the people, but you can't take it back!!
Perhaps that's one of the biggest reasons people should think long and hard about attempting leverage open source to gain popularity and a user base. There's that possibility of the user base forking your work and taking it over if they don't like the direction you're going... and that's exactly what I predict will happen with BitTorrent. And while they're at it, they'll probably go ahead and build into it some anonymity protection.
Anyway, all p2p is based on innovation. Just look at history: Kazaa, napster, eDonkey and thousands of others. Protocols tend to disappear and being replaced by better and more sophisticated ones. Or they just get extensions like eMules Kademlia.
I think we should be happy that somebody's thinking about something new instead just relaying on something that's good but not optimal. Especially now with current climate of litigations and general problems (traffic shaping, etc) with BitTorrent it's time for something more resilient and anonymous.
Just my 2p.
"an experienced, industrious, ambitious, and often, quite often, picturesque liar" - Mark Twain
The PirateBay team is currently developing on a new torrent protocol that they hope will be the next-generation successor to the current BitTorrent file. They say that they're concerned about continuing to use the current standard since BitTorrent has closed the source and hope to be able to create an open-source successor that maintains backward-compatibility with the current .torrent standard. The new standard, currently named .p2p is still in the development phase, but the initial release is planned for sometime early next year. Among the planned new features are responses to the increasing number of spammers and anti-piracy organizations who currently abuse the BitTorrent protocol.
Seriously, would it have been that hard to have waited for a submission that was informative and grammatically correct?
There are some very interesting technologies that can be applied to a new .p2p format while remaining backward-compatible with .torrent files. Such as auto-regeneration of almost-complete torrents via in-file redundancy (small size increase, massive benefit), the possibility of onion routing and obfuscation, new uploading algorithms, that sort of thing.
.torrent specifications -- the old, open one and the new, partially-closed one -- why not go whole hog and fork the thing all to hell? An application should be able to easily handle both.
And honestly, if Bittorrent closes some of the protocol, the features either going to be ignored or reverse engineered. In which case there's already 2 different
[ think ]
After MPAA got Bram Cohen and the UTorrent guy on their pockets, it was a matter of time until they tried to pull such stunts. My bet is that they will try to close a "hole" in the protocol, the impossibility to create a truly private swarm, one where only authorized peers could connect, regardless of the desire of the peers themselves to share the information about the other peers (DHT style). That's the wet dream of people selling content, they could sell access to their content using the bittorrent protocol and nobody would be able to join the swarm without paying.
But there is nothing there people should be afraid., as everybody knows, real innovation on the P2P scene occurs when the interested parts (the filesharers, not necessarily illegal ones) are the real force behind the development, as PEX (protocol encryption) came to prove, now that the cat is out of the sack, there is not a lot of things that Mr. Cohen can do.
By keeping the source closed, he is in fact assuming all responsibility for the actions of his code. If his code allows something bad to happen, we can say with certainty that it's all his fault.
Hey Pirate Bay folks, here's my list of feature requests for the new version of your open source torrent protocol:
ONION ROUTING:
1) Implement Onion routing (aka: Tor / anonymize the sources) as a built in feature.
2) Onion Routing should, where possible, try to use exit points and middle points that have roughly the same amount of bandwidth as you, otherwise torrenting will not become a reality through Onion Routing. So some kind of peer bandwidth algorythm needs to be incorporated.
3) Onion routing should be on by default, and each user should also become an exit point and donate 30% of their bandwidth to this. This will greatly increase the number of exit routers & provide this as a defacto alternative, as opposed to just some obscure security feature for the 31337 (hackers & government homeland types).
4) Individual site upload ratios, should take into consideration that fact that you are an exit point and some portion of that 30% should be counted toward your uploaded bytes ratio (even if traffic is going to other sites)... in other words, help promote torrent security = get bonus points from private trackers.
SIMPLIFY ISP SHAPING BYPASS
Background: Forcing protocol encryption isn't enough these days; some ISPs are shaping or even blocking torrent traffic by methods such as sending TCP RST packets to close a session, or their infrastructure auto-analyzes your encrypted traffic patters and if they are high bandwidth, very encrypted and on for long amounts of time to the same destination you get flagged & shapped (regardless of the fact that you could indeed be doing something legal)
1) There's a page on Wikipedia that lists all the "BAD ISPs" (http://www.azureuswiki.com/index.php/Bad_ISPs). This is a list of ISPs internationally that in one way or another shape your bitorrent traffic (Comcast anyone?). We need to be one step ahead of these ISPs and render their multi-million dollars worth of shaping infrastructure useless - sooner rather than later - sooner so that they can't make up for the ROI on all that gear they purchased. If the ROI fails, the next time engineering dept approach CEO for X dozens of millions more, they will get declined and we (torrent community) will win.
2) This site breaks down "throttling" into 5 different categories or ways in which the ISP can throttle you... each listing the bypass method.
http://www.azureuswiki.com/index.php/Avoid_traffic_shaping#Escalation_of_the_crypto_settings
Note that level 5 (the most aggressive shaping method known so far) is only bypassable by a single client today (Azeurus), utorrent to my understanding can not bypass this.
Anyway my point with these above 2 items is that these facts need to be considered:
1. The number of ISPs throttling internationally is already large and growing larger
2. Your new torrent client needs to simplify bypassing these various levels of encryption so that it can be adopted by the masses. If it is not adopted by the masses (rendering ISP throttling useless), the ISPs will have won.
I don't have time to type more, so please research what other clients out there (beyond just torrent) are doing and borrow ideas from them.
Here's a brief list of intelligent encryption/anonymous software out there to investigate:
RODI: http://slashdot.org/article.pl?sid=05/06/01/1252232
MUTE: http://mute-net.sourceforge.net/
ANTS: http://antsp2p.sourceforge.net/
GNUnet: http://gnunet.org/
I2P: http://www.i2p.net/
FreeNet: http://freenetproject.org/
TOR: http://tor.eff.org/
THanks and good luck!
No trees were killed in the making of this post; however, many trillions of electrons were horribly inconvenienced.
Hell, I think modern piracy would've taken longer to come to their attention if the dumbshits at Napster hadn't tried to make a business model out of it.
If the protocol is open-sourced, I don't care if he writes a closed-source implementation. However, the current protocol that they claim to be writing isn't published on the wiki. They're keeping it a secret... so, screw BitTorrent.
I vote that we write one of our own. I've written a BitTorrent client before, and have written a protocol extension. I'm just beginning to ponder a completely new protocol. Any interest?
Beer is proof that God loves us, and wants us to be happy.
You can't take the sky from me...
Good luck close-sourcing Python code, anyway... reverse-engineering .pyc is beyond trivial. If there's anything really useful in there, it will be reverse-engineered and mysteriously make its way back into the BitTorrent OSS fork, anyway.
Proud neuron in the Slashdot hivemind since 2002.
I hear they're going to have Johnny Dupe in that one...