08/31/08 12:55:48 changed by josef
ΒΆ
I've written a patch to support magnet links now. You need to check out svn rev 1065 of libtorrent/rtorrent, and get http://ovh.ttdpatch.net/~jdrexler/rt/experimental/dht-pex-static_map.diff and http://ovh.ttdpatch.net/~jdrexler/rt/experimental/magnet-uri.diff then in the directory that has the libtorrent and rtorrent subdirs you've checked out, do
patch -p0 < dht-pex-static_map.diff
patch -p0 < magnet-uri.diff
and recompile both.
It uses the official magnet protocol from Bittorrent BEP-0009 which is incompatible with Azureus and so far only supported by uTorrent 1.8+, so it'll only work if there are recent uTorrents in the swarm. It supports magnet links in both the old style base32 encoded hashes as well as the recommended URL-encoded hashes.
Note that if there is one or more tracker URLs to use for the download, it must be present as "tr=..." argument in the magnet URI, because there is currently no way of adding trackers in rtorrent afterwards, so without that it'll use DHT and nothing else.
After opening a magnet URI, it will add a meta download to download the actual torrent info. When that is complete, it is replaced by the real torrent. The meta data is saved in your standard torrent download directory, you can delete that after the real torrent has appeared, or you can keep it in case you need to open the same magnet URI again.
Mike Chelen on 19 Nov 08Existing studies on BitTorrent systems are single-torrent based, while more than 85% of all peers participate in multiple torrents according to our trace analysis. In addition, these studies are not sufficiently insightful and accurate even for single-torrent models, due to some unrealistic assumptions. Our analysis of representative BitTorrent traffic provides several new findings regarding the limitations of BitTorrent systems: (1) Due to the exponentially decreasing peer arrival rate in reality, service availability in such systems becomes poor quickly, after which it is difficult for the file to be located and downloaded. (2) Client performance in the BitTorrent-like systems is unstable, and fluctuates widely with the peer population. (3) Existing systems could provide unfair services to peers, where peers with high downloading speed tend to download more and upload less. In this paper, we study these limitations on torrent evolution in realistic environments. Motivated by the analysis and modeling results, we further build a graph based multi-torrent model to study inter-torrent collaboration. Our model quantitatively provides strong motivation for inter-torrent collaboration instead of directly stimulating seeds to stay longer. We also discuss a system design to show the feasibility of multi-torrent collaboration.