Slashdot Mirror
Examining Bittorrent
ToyKeeper and other wrote in with this: "The Register published a detailed analysis of BitTorrent traffic and user habits today, focusing on four aspects: availability, integrity, download speeds, and ability to withstand flash crowds. BitTorrent carries 53% of all P2P traffic (or ~35% of all 'net traffic), and this paper helps explain why. Also included are data about torrent lifetime, network poisoning, response during downtime or attacks, and lots of pretty charts. A few performance problems are revealed, which will hopefully be addressed in future p2p systems." The original paper (pdf) is available.
... again ...
How are we supposed to "mark your words" when you post as an AC? Also, you seem to think that downloading music is illegal everywhere, when it's not. Not everyone lives in the US of BushCo. Also, the servers holding the torrent files are not breaking any laws.From the article:
Well, since, according to El Reg http://www.theregister.co.uk/2004/12/08/brit_net_The stats:
All the above are LEGAL.my journal: scripts for leaching porn baked fresh daily
Yup. All it takes is any.
The legal principal is this: if the {object, device, chemical, drug} has a purpose for which it is legal, then the thing should be legal.
The exceptions to this (guns, marijuana, and other things we've allowed to be banned) prove the rule. The pressure to legalize or ban something evinces arguments about its legitimate uses, and it's these arguments that are persuasive. Saying "We'll do it anyway" is unproductive.
In this case, since downloading Free software is so much more efficient with P2P, it's inappropriate to ban it even if that software is only a small percentage of the service's traffic.
Raise your children as if you were teaching them to raise your grandchildren, because you are.
*HUGE* anonymous investigations!!
.torrent on one tracker. I'm not giving any legal advice here, but if you were to download one file for what you believe to be fair use, then they won't be able to come after you like they did with KaZaa users. Instead of the hundreds of shared files, your IP address is now only associated with one.
.torrent is being shared. Someone would have to look for all new torrents and connect to the tracker and start logging IP addresses for the lifetime of the .torrent, plus who is to say you have the whole file? Are you a criminal for sharing part of file, a chunk that is useless on its own?
I always listen to warnings like that! Y2K, update to SP2, don't download anything from the internet.
BitTorrent is inherently "safer" than any P2P (like KaZaa). Can you be busted for sharing illegal files? Sure. But.. You are at most only in trouble for the ONE copyright violation from one
Could they monitor EVERY tracker and EVERY torrent on those trackers and log EVERY IP address, maybe.. But don't forget torrents are time based, ie. you are only sharing file for a certain percentage of the time that
Wrong. There are no investigations going on to bust bittorrent users. There are investigations going on to bust people doing illegal file sharing, and some of them happen to be using bittorrent.
35% = BitTorrent
35% = other P2P (the article tells you this)
5% = plain old FTP (just a random guess of mine)
2% = email / instant messaging
23% (the remainder) = other (newsgroups?) / plain browsing, of which a significant portion might be pr0n.
(All numbers about as accurate as the results of a Slashdot poll ;-)
Slashdottings may be fun to note, but significant amount of all internet traffic? Don't think so. The low number for mail is because there may be lots of spam, but it's not that big a part of the total amount of data. Download a movie, >1 GB. traffic, and you can watch the content in an hour and a half. A single e-mail is maybe just a couple or tens of KB's, but may keep you busy for a while. And like IM, mostly text-based.
I think this is the basic idea behind the PATRIOT Act, the Clear Skies initiative, the Healthy Forests Proposal, No Child Left Behind act, CAN-SPAM act, etc. The forces of evil are way ahead of you on this.
The study shows how vulnerable BitTorrent is to failures of Suprnova mirrors and trackers. Kill a large .torrent host and you effectively kill the network. Kill a large tracker and you severely cripple it. In comparision, ed2k network is much more resilient to attacks.
.torrent file that needs to be hosted on a large server. You can send an ed2k link by e-mail, IM or post it on Slashdot. Furthermore, ed2k has excellent search capabilities - both via servers (very fast and very efficient) and via distributed Kad[emlia] system (fast and efficient). With the ability to check the filenames and comments for a certain file, you are relatively safe against fakes even when you can't use verified links. Of course, here I deliberately ignore the fact that both networks need "community portals" to inform users about released files, to provide forums for discussion, etc.
/. and elsewhere (as if other networks don't exist), I still don't hate BT. Actually, I tried it again recently and was very satisfied with the download speeds. I don't wish BitTorrent bad. But with the recent developments with police raids on torrent sites and ed2k link sites in Europe the networks will be tested and I am not sure BitTorrent is best prepared for it. Suprnova appears to be safe because of its geographic location, but it still remains a single point of possible failure. I don't think ShareReactor was as critical to the edonkey2000 network.
First, you don't need servers to distribute ed2k links. A short ASCII string effectively replaces a large
Second, the servers play only a secondary role, even if many servers would go down, that would have a small impact on the network because of source exchange. And using Kad it's even possible to operate entirely without servers.
I do not hate BitTorrent, really. Even though I am a long time eMule user and even though I am very annoyed by the apparent popularity of BitTorrent here on
Future Wiki -- If you don't think about the future, you cannot have one.
This paper and all of the recent news articles that provide an estimate for BitTorrent protocol traffic use the same source. A single slide in a presentation by someone from Cache Logic shows BT using 1/2 of all P2P traffic at a "tier 1 ISP." Other sources cite P2P traffic at 66% of all 'net traffic. Therefore, BT is 33%.
I think any estimate made without measurements at many major routers would be suspect. While there is no doubt that BT is quite popular, the evidence presented thus far for the amount of traffic using BT protocol is extremely flimsy. I would take it with a grain of salt.
burris
I have a few comments on your analysis of the BitTorrent protocol... My main criticism is that you are analyzing BitTorrent in combination with pirate web pages as a P2P file sharing system, when BitTorrent's real purpose is to be a file DISTRIBUTION system.
.torrent file) would obviously receive a genuine .torrent file, and the data in that file verifies the data received in the download. It's only torrent file redistributors like Suprnova.org where you'd need to be concerned about pollution.
.torrent file anyway, and a standard download would also be blocked.
:-)
BitTorrent is designed to replace and enhance the performance of a standard http or ftp download server. Where even ten simultaneous downloads can slow the performance of most inexpensive server setups to a crawl, BitTorrent can easily handle ten thousand or more, and in this it is an enormous success.
One necessary element of a true BitTorrent distribution is a dedicated seed server. This server ought to be always working, and should have a significant amount of bandwidth behind it; I'd recommend 30KB/s minimum, but more is better. You complain that seeders are "punished" and this is why torrents die, but while long-term seeders are nice, they aren't necessary. It is better for me as a content distributor to allow people to close their torrent and play with their new download as soon as they'd like to. Having torrents die off when interest fades is an artifact of misuse of this specification.
You worry about pollution on Suprnova.org, and so do I; there's no reason why it wouldn't exist. But as BitTorrent was normally intended this isn't a problem at all. People visiting Blizzard's website to download content via BitTorrent (actually Blizzard uses a modified downloader, but the concept would be the same if users received a standard
You're also concerned about tracker availability. I recommend content distributors run their own trackers, which is an easy task given the numerous types of trackers available, including script-based trackers. There's no reason for a tracker to go down unless the web server goes down, in which case no one would probably be able to get a copy of the
As a sharing method BitTorrent indeed has some deficiencies, but it simply wasn't designed for that. That BitTorrent is being misused for that purpose only testifies to its effectiveness. Perhaps a sharing system with elements taken from BitTorrent will someday arise; I know Suprnova.org is attempting to create one with "Exeem". But please don't badmouth BitTorrent.
Unlimited upload on BT (or any p2p for that matter) is bad because it uses TCP so the upload interferes with the packets sent to tell the computer you are downloading that you are actually reciving packets from it. Since those packets do not get sent, the computer you are downloading from sends slower. It would also be interfering with your HTTP requests and acknowledgements to the web server. (There is a correct terminology for what I am saying, I am just to tired to think of it.) On the other hand, I have found that capping my upload too low does lead to lower download speeds. (If there are not a ton of seeds, my download tends to hover around three times my upload.)
Centralization breaks the internet.