Karagiannis et al. [8] quantified the influence of BitTorrent
on ISPs using two types of real measurement data, the
BitTorrent tracker log for RedHat v9.0 and the payload packet
traces collected using a monitor installed on the access link
of the network of a residential university in USA with 20,000
users. While the paper considers the more general issues of
peer-assisted content delivery and its impact on ISPs, we
are specifically concentrating on their measurement based
evaluation and observations about BitTorrent. They indicated
that BitTorrent is locality-unaware which severely increases
the ISPs’ cost, typically resulting in the same content being
downloaded into the same ISP multiple times from external
peers. They also found that users requesting the same content
have a 30%-70% overlap in their lifetimes, consequently
helping each other in the download. An additional 20%-40%
improvement in terms of cross-ISP downloaded content can
be effected by giving incentives for users to stay around after
they complete the download.
Andrade et al. [20] focused on measuring the cooperation in
BitTorrent. In their measurement, they defined three metrics:
(1) free-riding ratio, which is the percentage of free-rider
peers, (2) seeding ratio, which is the percentage of seeds in
a torrent, and (3) sharing ratio, which is the total uploaded
data divided by the downloaded data. They evaluated the
metrics by collecting logs in BitTorrent communities such
as bt.etree.org, and piratebay.org. Then, they measured the
downloading time for free-riders and non free-riders, and the
amount of free-riding and low-sharing behavior in BitTorrent
communities: etree and easytree. Their results demonstrate
that BitTorrent’s TFT protocol does discourage free-riding
successfully. However, if there are a large number of seeds
in the torrent, the TFT mechanism may not work effectively
to prevent free-riding.
Legout et al. [24] conducted experiments on private torrents
and collected data from peers in a controlled environment.
They focused on the choking algorithm in BitTorrent to
observe peers’ individual behavior during the downloading
process. Their experiments were performed on PlanetLab.
Karagiannis et al. [8] quantified the influence of BitTorrenton ISPs using two types of real measurement data, theBitTorrent tracker log for RedHat v9.0 and the payload packettraces collected using a monitor installed on the access linkof the network of a residential university in USA with 20,000users. While the paper considers the more general issues ofpeer-assisted content delivery and its impact on ISPs, weare specifically concentrating on their measurement basedevaluation and observations about BitTorrent. They indicatedthat BitTorrent is locality-unaware which severely increasesthe ISPs’ cost, typically resulting in the same content beingdownloaded into the same ISP multiple times from externalpeers. They also found that users requesting the same contenthave a 30%-70% overlap in their lifetimes, consequentlyhelping each other in the download. An additional 20%-40%improvement in terms of cross-ISP downloaded content canbe effected by giving incentives for users to stay around afterthey complete the download.Andrade et al. [20] focused on measuring the cooperation inBitTorrent. In their measurement, they defined three metrics:(1) free-riding ratio, which is the percentage of free-riderpeers, (2) seeding ratio, which is the percentage of seeds ina torrent, and (3) sharing ratio, which is the total uploadeddata divided by the downloaded data. They evaluated themetrics by collecting logs in BitTorrent communities suchas bt.etree.org, and piratebay.org. Then, they measured thedownloading time for free-riders and non free-riders, and theamount of free-riding and low-sharing behavior in BitTorrentcommunities: etree and easytree. Their results demonstratethat BitTorrent’s TFT protocol does discourage free-ridingsuccessfully. However, if there are a large number of seedsin the torrent, the TFT mechanism may not work effectivelyto prevent free-riding.Legout et al. [24] conducted experiments on private torrentsand collected data from peers in a controlled environment.They focused on the choking algorithm in BitTorrent toobserve peers’ individual behavior during the downloadingprocess. Their experiments were performed on PlanetLab.
การแปล กรุณารอสักครู่..
