Jump to content

utorrent P4P Feature Project (Please Say NO!!!!)


TG

Recommended Posts

Is utorrent planning a P4P project?

Please tell me it is not mainly for the reason in this article.

http://torrentfreak.com/uncovering-the-dark-side-of-p4p-080824/

Mainly because MPAA is already got there heads into it.

And it seems to be more focused towards local peers not worldwide (Bad news)

It would be great to see development isn't going that way.

Link to comment
Share on other sites

P4P support was scrapped long ago when the obvious bias of their results proved that it would offer no benefit to users.

Until ISPs allow for higher network-local speeds, any attempts to bias in favor of network-local peers won't be implemented.

And besides, if there ends up being a large number of ISPs with higher network-local speeds, we will more than likely implement our own system to take advantage of that.

Link to comment
Share on other sites

It's not about favoring. There are no methods in-place for this type of detection to HAPPEN AUTOMATICALLY. Hopefully the inroads BitTorrent, Inc. is making with Comcast regarding "network management" bears fruit for all the people who harp on this type of feature. With the largest cable provider in the continental U.S. working with (or hopefully soon not outright blocking p2p with sandvine technology) the intelligent staff, I have hope for the future :D

Link to comment
Share on other sites

Local peer preferencing is just another word-phrase for "disconnect any peers that AREN'T local"...or at least don't upload/download with them except as a last resort. This would rapidly accelerate the fragmentation of torrent swarms into separate groups that don't share with each other, resulting in the torrent going unseeded sooner than it would otherwise. It's almost like another level of firewalled and unfirewalled BitTorrent clients.

It would be nice if local peers could be tried first to connect if spotted in the peer list for a torrent. However best I can tell, uTorrent currently isn't very good at determining what its ip is...so it could find similar ips or same ISP ips.

Link to comment
Share on other sites

Exactly Switeck, Any way in the article in torrent freak does not say who is in change of putting p4p out in to the work, is it bittorrent inc?

I am just glad that utorrent isnt bothering with it.

But if bittorent inc own utorrent and if they are in charge of the p4p project that means utorrent will have to do the project no question.

Link to comment
Share on other sites

If the speed is higher locally wouldn't that result in those connections being favoured because of their speed?

As to ISP's - you pay them for the bandwidth, and they should deliver, even if it's to the other side of the world. If they don't want that, they should change their contracts.

Link to comment
Share on other sites

If you're connected to local peers, they should have low latency and respond favorably to you uploading to them...if they're allowing decent upload speed OR don't have limit local peer bandwidth checked.

BUT...ISPs are intentionally sabotaging this by throttling all BitTorrent traffic (especially upload) much lower than normal line speeds and then complaining it's BitTorrent's fault when BitTorrent clients seek out distant peers/seeds because local ones suck.

Also, the way things are currently uTorrent doesn't go looking for local (same neighborhood/area) and semi-local (same ISP) peers. If the local peer discovery pings fail, it's pure hit-or-miss...mostly miss.

On torrents with few peers and seeds, a local peer is unlikely. On large torrents, the local peer ip may not get tried before per-torrent connection max is reached...or the local peer (or uTorrent) is firewalled...or one only allows encrpytion and the other has encryption disabled...or they're both seeds...or one has nothing to offer the other.

It's easy for local peers to miss each other or gain little even when they find each other. So uTorrent has to determine its internet ip quickly (even in advance of starting a torrent) and then try ips at least in the same part of the world first (ARIN, APNIC, RIPE, JPNIC, LACNIC). Just trying the same starting ips (where A=same as you in A.x.x.x) would be nice if lots of ips arrived at once from tracker updates and peer exchange.

Link to comment
Share on other sites

  • 4 months later...

Very nice situation in Ukraine.

40% of ISP connect clients up to 100Мbps, but it is a speed in local network, between subscribers (their quantity can be more than 20000 in Kiev). In UA-IX (95% of Ukrainian ISP connected in one point for an exchange by a local traffic, but have very different channels in the world) speed can be 50-80Мbps between ukrainian peers, and in the world - 5,10 or 20М (different tariffs). That is already there is a considerable prize in speed when it is possible to find people from this network or, at least, from UA-IX.

In practice when I start utorrent, I some times stop torrent and I start it again - for such 10-20 starts almost always is peer UA-IX...

It is just necessary to think very well over algorithm, probably, it is not necessary definition local ip, BUT over testing (by turns) to speed of each seed/peer, and a choice as much as possible fast...

Link to comment
Share on other sites

Just some more articles about P4P:

http://broaddev.com/2008/11/10/comcastic-p4p-80-speed-boost-for-p2p-downloads/

Comcastic P4P: 80% Speed Boost for P2P Downloads

http://arstechnica.com/news.ars/post/20081103-comcastic-p4p-trial-shows-80-speed-boost-for-p2p-downloads.html

Comcastic P4P trial shows 80% speed boost for P2P downloads

(almost a duplicate of above)

One of the actual P4P tests:

http://www.ietf.org/internet-drafts/draft-livingood-woundy-p4p-experiences-02.txt

Comcast's ISP Experiences In a Recent P4P Technical Trial

...And there seems to be some questions about the details and validity of the test:

http://www.ietf.org/mail-archive/web/p2pi/current/msg00883.html

...With replies:

http://www.ietf.org/mail-archive/web/p2pi/current/msg00884.html

Ongoing, the "story" seems to continue with the ALTO (Application-Layer Traffic Optimization) mailing list:

http://www.ietf.org/mail-archive/web/alto/current/maillist.html

And here...with more craziness about uTP:

http://www.ietf.org/mail-archive/web/ledbat/current/maillist.html

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...