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.
aside from movie and music piracy there are legal uses for bittorrent p2p too, like when Linux distros are released the demand is much greater than the file servers can handle and thats where bittorrent plays an important role, i prefer to get my Linux ISOs via bittorrent because it helps others get their ISOs too, for example FedoraCore-3 was released and it came on 4 CDs plus a fifth rescue CD making for a HUGE download, and also offered resume so if you have to log off or have a network problem you don't lose all that data and have to start your download over...
There are also sites that list legal torrents, try File Soup or Legal Torrents for example. These are just two that I remember offhand, I'm sure there are many others as well. Remember, BitTorrent, like any other P2P application, has plenty of legitimate uses. Don't get sucked in by the *AA propaganda machine (not directed towards the parent, just saying that in general).
from right below figure 4: In order to test the integrity of meta-data, we donated to Suprnova an account for hosting a mirror. By installing spyware in the HTML code, we have registered each .torrent download and could have easily corrupt the meta-data. We conclude that using donated resources for hosting meta-data entails substantial integrity and privacy risks.
It all comes down to two things: knowing where to host, and how to maximize your availability.
35% of all net traffic belongs to BitTorrent traffic. The corresponding web and traffic tracker required to power that is inconsequential.
I used to run NovaSearch.net, which was for a time the official search function of Suprnova.Org. I made up roughly half of all their traffic, something on the order of 300k pageviews per day by the end. Availability was indeed a large problem, and always my primary concern. However, my possible availability was much higher than actual availability. By this I mean that Novasearch had the POTENTIAL to be available much more than it was, due to reliance on Suprnova.
When SuprNova went down, NovaSearch (usually, often it could be used as an out-of-date cache when Suprnova was down) went down too, because it didn't get updates. That accounted for most of my downtime, very little of it was actually from issues relating to NovaSearch itself.
Despite all this, NovaSearch, during it's primary operational period, relied on only one dedicated server (A second was added for static content later on, but for transfer cap reasons, not actual bandwidth or load). This highlights the primary problem with Suprnova in regards to their reliability, they rely on donated mirrors, and that reliance has caused them to use an insufficient architechture (Last I heard the core of Suprnova was one single dual xeon server). Had they instead chosen to use a clustered solution that they managed themselves, combined with hardware firewalls and DDoS mitigation technology, the availability then and now would be significantly higher.
Tracker reliability is a much lesser problem. Torrents can easily survive short to medium tracker downtime just by the shear momentum of the users. Once they have a peer list, they can continue communicating with those peers even with the tracker down. And the widespread adoption of various unofficial additions to the BitTorrent protocol have further improved that. One such improvement that enjoys almost universal support among third-party BitTorrent clients is the multi-tracker protocol, which effectively allows trackers to be clustered so that even if all but one of the trackers for a torrent is down, it can continue normally.
Anyhow, this is a long post that sort of went off on a tangent and started rambling, but I thought that I should put a few words in because of the role I played.