Beruflich Dokumente
Kultur Dokumente
Revolutionised Apportion
CHAPTER 5
TRACKERS
Open trackers can be used by anyone by adding the tracker address to an existing
torrent, or they can be used by any newly created torrent.
Some of the allure of private tracker versus a public one are: higher speeds, a
tighter community, and safer downloads. Private trackers implement a strict set of rules,
so generally files containing malware are extremely uncommon. Many private trackers
keep in close contact with each other, so bad users (who trade invites or attempt to fake
their ratio) can be quickly blacklisted. Almost all private trackers implement a passkey
system, where each user is given a personalized announce URL so if there is
unauthorized distribution, it can be pinpointed to the user responsible. Some private
trackers have a higher level of security than others - many sites, such as, the late ScT,
only allow their users to refer to their site as an abbreviation, and never as the site's full
name or URL. Other trackers restrict invites to outstanding members, and many trackers,
to increase security, have gotten rid of the invite system altogether. An example of a
private tracker is pink, which was forcibly shut down in late 2007 by law enforcement
officials.
The downside is that some private trackers are little more than pyramid schemes
inviting donations to enable users to bypass the upload-to-download ratio. In a closed
community it can be mathematically impossible for all members to maintain the required
ratio. Competitions may be offered, with prizes of improved ratios, and forums may
suggest other complicated means of doing this. In the pyramid schemes, 'donations' turn
out to be the only way to continue downloading. Downloads can then be quite expensive.
On the other hand, if they are unobtainable elsewhere, they may be good value for
money. Some trackers use "free leech" systems to improve the users ratio. When
downloading a free leech torrent only the upload gets logged in, the download is ignored.
Usually popular (healthy) and large torrents are offered as free leech. Seeding to a ratio of
at least 1 is still recommended even for free leech torrents.
Trackers are the primary reason for a damaged BitTorrent 'swarm'. (Other reasons
are mostly related damage or hacked clients uploading corrupt data.) The reliability of
trackers has been improved through two main innovations in the BitTorrent protocol.
Multi-tracker torrents
Multi-tracker torrents feature multiple trackers in a single torrent file. This provides
redundancy in the case that one tracker fails, the other trackers can continue to maintain
the swarm for the torrent. One disadvantage to this that it becomes possible to have
multiple unconnected swarms for a single torrent where some users can connect to one
specific tracker while being unable to connect to another. This can create a disjoint set
which can impede the efficiency of a torrent to transfer the files it describes.
Trackerless torrents
The original BitTorrent client was the first to offer decentralized,distributed tracking
using a distributed hash table (DHT), making torrents more independent from the tracker.
Later, Vuze, rTorrent, µTorrent, BitComet and KTorrent adopted this feature. Vuze's
"Distributed Database" feature uses its own form of DHT (Kademlia) which is
incompatible with the official BitTorrent client's implementation; however, support for
the official implementation can be added through the Mainline DHT plugin. Most other
clients support the official DHT implementation.
DHTs form an infrastructure that can be used to build more complex services, such
as distributed file systems, peer-to-peer file sharing and content distribution systems,co-
operative caching, multicast, anycast, domain name services, and instant messaging.
Notable distributed networks that use DHTs include BitTorrent's distributed tracker,
the Kad network, the Storm botnet, YaCy, and the Coral Content Distribution Network.
Decentralization: the nodes collectively form the system without any central
coordination.
Scalability: the system should function efficiently even with thousands or millions of
nodes.
Fault tolerance: the system should be reliable (in some sense) even with nodes
continuously joining, leaving, and failing