BitTorrent, not the movie
p2p news special / p2pnet:- The MPDA (Motionless Picture Disinformation Association) is proud of its record in closing down BitTorrent sites.
It all looks good in lamescream media headlines but in real life, its efforts are about as noticeable as a fart in a thunderstorm and in fact, from a movie industry perspective, they’re doing far more harm than good.
The MPAA’s opposite number in the music industry, the RIAA (Rip-off Industry Association of America), has done a great job publicizing p2p file sharing, introducing it to millions of people who, without the RIAA’s sue ‘em all marketing campaign, would never have heard of it.
P2p and the p2p networks are here to stay and one of the primary applications is BitTorrent.
Yesterday, we featured a really good article on its creator, Bram Cohen. Written by Fortune senior editor Dan Roth, it offers insight into Bram himself, and on the process which brought BT to life.
What it doesn’t do, however, is explain what BT is.
Regular p2pnet poster TurboGeek breaks it down >>>>>>>>>>>>>>>>>>>>>>>
Bit Torrent is akin to a huge food fight where you get pieces of everything offered thrown at you, and there are no lunch ladies ; )
The basic idea of Bit Torrent is that instead of downloading a file in strictly bit-sequential order from a single source over single TCP connection, you get pieces of the file from many different sources in the torrent. The set of peers participating in a given torrent is called a ’swarm’. Each piece is assigned a number (similar to the file numbers that NTFS assigns to each of your files, as opposed to using it’s filename.) Each client sends and recieves pieces in accordance with an algorithm based on Pareto efficiency theory. (Look that up on Google, along with Vilfredo Pareto, it’s progenator.)
The client on a peer’s computer then takes the pieces collected, keeps track of them and assembles them in the correct order to reconstitute the file(s) comprising the download.
Generally, there’s another entity that’s required called a ‘tracker’. This merely keeps track of who (IP address) is in the torrent and which specific pieces they’ve already amassed. The tracker does NOT handle any of the object data in the download. (This is why the cartels have a difficult time trying to pull the plug on trackers located in many foreign countries. There’s no legal basis for taking action against the tracker as it is not handling any of the allegedly infringing data.)
A practical example: Suppose a teacher has a class of 30 students and is going to distribute a 10-page handout that there’s insufficient time to collate or staple. They are just 10 piles of 30 copies on the teacher’s desk. The teacher could pick up one pile and proceed to individually hand each student a sheet of paper, returning to the desk for the next pile when the first is completely distributed. This would be tedious for both the class and the teacher.
So, the teach takes each of the ten piles and hands a pile to one of ten different students and tells them to give one to each student, including themselves. Now there are 10 distributors. The handout gets distributed much more quickly.
A student now suggests that each student who’s holding a pile of copies splits his or her pile into thirds and gives the other two-thirds to students who weren’t given a pile of copies to distribute. So they distribute their copies. Each student now has to check each desk to see if that desk already has a copy of the page they’re distributing. If so, they go to the next desk. If not, they leave one of the copies on that desk. When they’re empty handed, they return to their seat.
Mass confusion reigns in the classroom for a few minutes. Invariably, some pages have gotten stuck together or a few students have not gotten one or more pages for some reason. They raise their hands and the other students look through their copies for duplicates and get them to where they belong.
This final phase is called the ‘end game’ where the algorithm goes out the window and the object becomes to get the last few pieces to a peer who’s on the verge of completing the torrent. (At which point fireworks go off and one hears the strains of “Stars and Stripes Forever” being played in the distance. Well, it feels that way after being on a 9GB torrent for two weeks with one seed on four peers and it finally completes.)
In BT, a peer with a complete copy of the torrent is called a ’seed’. Since they don’t need any pieces, they only distribute copies of the pieces. This is called ’seeding’. ‘Leechers’ are peers without a complete copy of the torrent. They’re still receiving (leeching) data, but are also sending copies of the pieces they do have to other peers who don’t.
The key feature of the BT ‘process’ is in the selection of which piece to send.
Theoretically, a peer should send the piece for which the fewest copies exist in the torrent and are therefore considered ‘rare’. Each distribution of such a piece makes it less rare.
Using the school analogy again, think of it in terms of the cafeteria. In the normal client server mode, the students stand on line in a quiet and orderly fashion holding their lunch trays, inching forward to where the lunch lady in the ubiquitous hairnet shovels a large scoop of the day’s sludge onto the tray, the student moves to the next lunch lady, and so on, while the rest of the class is still waiting on line.
Bit Torrent is more akin to a huge food fight where you get pieces of everything being offered that day thrown at you at once by the other students, and there are no lunch ladies (servers). You just try not to get hit with the same piece of food twice.
===============
Something you think we should know? tips[at]p2pnet.net
First they ignore you, then they laugh at you, then they fight you, then you win
- Mohandas Gandhi
Tired of being treated like a criminal? They depend on you, not the other way around. Don’t buy their ‘product’. Do bug your local political representatives. Use emails, snail-mail, phone calls, faxes, IM, stop them in the street, blog. And if you’re into organizing, organize petitions, organize demonstrations and then turn up on your local political rep’s doorstep, making sure you’ve contacted your local tv/radio station/newspaper in advance.
October 19th, 2005 at 9:24 pm
I read this yesterday, shortly after you posted it. It made a lot of sense in a very untechnical way. So much sense that almost anyone can relate to it and see the similie that is drawn without having to understand terms in computereze. Having to deal with unfamilar terms while learning new concepts makes a huge learning curve for the uninitiated. Your explaination avoided all of that while driving your point home.
Good job.
October 20th, 2005 at 1:48 pm
The beauty of this system is never while you are leaching do you have a full copy being serviced out to be nailed with as a host of a file. Most ISPs try to block ports these programs use, but that is rapidly becoming moot as the software simply tries every port out there till if finds one and re-tasks around the problem.
The Other thing that makes this method secure is the addition of a program called PeerGuardian 2 This program blocks all the current Anti P2P companies IPs from interacting with your system for ANY reason. Thus even if they are port scanning you looking for files, they have nothing to capture as evidence. Currently there are 814,485,789 IPs blocked in this way. and more everyday.
Next generation Bittorrent clients are building this into their software and also building anti-fake torrent checking into their search tools. The only thing that could happen is the program developers themselves could be shut down, but that would do nothing to stop the existing millions using the system now on the open source versions.
October 21st, 2005 at 11:06 am
Thanks for the compliments. Nearly everything having to do with computers can be explained using analogies to other domains, simply because computers behave so predictably. I’m talking in the abstract sense here as in “If the machine is in state X and executes Y resulting in Z, it should do that every single time.”
The basic idea behind Bit Torrent has been used before. A simple example is downloading a web page with several pix, several graphics, regions of text, etc. Your machine will open several TCP connections to the server. Each connection requests one such element of the web page. The browser consumes to data provided, organizes it the way it should be in RAM and then renders the page on the screen. In most browsers, you see the rendering process as the alt-imgs and alt-txts get replaced with the actual content.
Download ‘Accelerators’ do much the same thing. They ask for a download as several ‘chunks’ on the theory that if one chunk has problems, the other chunks will keep downloading while the problem is negotiated between the two machines.
In the first case, we may be talking about perhaps 20-60 elements and in the second, perhaps up to 10.
People routinely use a BT client to snatch an entire DVD. If the torrent is using 512kB pieces, a full DVD is going to contain over 9000 pieces. (4.7GB / 512kB (0.5GB) = ~9400)
What makes BT so technically interesting is the scale and the issues that arise as one deals with larger numbers of pieces being distributed to a larger number of peers who have greatly varying bandwidths across the population that runs from a T3 to a 14.4 dial-up.
What makes BT so newsworthy is Hollywood’s manic-depressive reaction to it. (We hate it! We love it! We love hating it! We hate ourselves for loving it. We love how much we hate ourselves for loving to hate it so much….)
They know it’s a perfect way for them to distribute on the internet, but how to own and control such a thing they have yet to figure out. Unfortunately, those are the aspects of anything new that they consider first, resulting in much hang-wringing and hyperbolic hysteria. whatever was ‘new’ gets either lost in the distraction or mangled with clumsy attempts at ’securing content’
–TG.
October 21st, 2005 at 11:35 am
Yes, there is a fairly significant unanswered legal question about whether distributing ‘pieces’ of material, each of which is practically impossible to tell what work it’s from by examining that one piece, is infringement. Furthermore, there is no way that anyone can tell for sure that you got the entire torrent unless you hand around to seed. People abandon torrents partially completed for all sorts of reasons.
The ISPs problem is not only with BT, it’s with all p2p. 6-7 years ago, when the ISPs were engineering their broadband networks they made the assumption that the traffic was primarily going to be asymmetic. The believed that the average customer was going to download far more data than they uploaded. So they concentrated (if I might refernce my analogy from the post) on hiring more lunch ladies and buying bigger serving spoons. They assumed people would eventually be downloading movies, but they would be doing so from a gargantuan server array connected to a 10 ExaByte RAID Farm sitting on a big fat pipe tapped right into the backbone, not from each other.
P2P is an inherenetly symmetic activity. For every byte of data you download, someone has uploaded it, and likely through the little straw that’s given each broadband ‘residential’ customer. That byte is much more taxing than a byte coming from an uber-server that is one hop from the backbone.
The ISPs main problem is that p2p red-lines their backhauls that get the data from their customers to the internet. If you get in trouble with your ISP, it’s from how much you are uploading, not what and how much you are downloading. It’s estimated that approximately 70% of the upstream traffic now if p2p related.
As far as tracking goes, many clients are now implementing DHT (distributed hash table) which essentially implements a distributed tracker, thus there is not single place to go to snoop around and see who is snatching what. It’s far less susceptible to sabotage as it’s quickly apparent if a peer is trying to pawn off corrupt DHT data as the table is extremely redundant. Any attempts at deliberate corruption with find that peer being swiftly shunned by the swarm.
As for attempts at censoring software developers, that’s highly unlikely to happen as BT has numerous legitimate, non-infringing uses. So far Bram has not said one word regarding the ultimate uses of BT, but has said things like “It’s really dumb to use BT for downloading music and movies because it’s not anonymous.” That should keep him on the clean side of the Grokster decision.
–TurboGeek