 swhx7Premium
join:2006-07-23
Elbonia | reply to funchords
Re: Legal P2P will be fostered; illegal P2P will be punished Trackers only connect the peers with one another. This could help to keep connections within the ISP's network if (a) there were seeders within the ISP's network (b) the ISP kept enough torrent files on the tracker to answer a lot of the demand (c) subscribers were willing to use the ISP's tracker instead of a remote one.
But the gain would be only small, For more reduction of peering costs, the ISP would have to cache content, or in Bittorrent terms, provide seeds. This would make ISPs the infringers, not merely conduits, in the case of infringing files. It would still leave customers going off the network whenever they want something that's not on the ISP's tracker, or when they don't trust the ISP. |
 funchordsHelloPremium,MVM
join:2001-03-11
Yarmouth Port, MA
kudos:5
1 edit | I think you misunderstand what a tracker is. A tracker isn't a website that hosts .torrent files or that has information about the contents a download.
A tracker is simply an HTTP machine that accepts queries containing an infohash and outputs IP addresses and port numbers related to it.
When your client contacts a tracker, that tracker doesn't know whether you're downloading a song or a picture, an album or a movie.
Many trackers are attached to web sites, those web sites do contain .torrent files and information about the download -- and they even exchange data with the tracker function.
So when I was talking about a tracker, I'm simply talking about the machine that identifies who is in the swarm.
said by swhx7:Trackers only connect the peers with one another. This could help to keep connections within the ISP's network if (a) there were seeders within the ISP's network Everyone who is downloading has data available to upload. There doesn't have to be a single on-net seeder in order to relieve stress at the network boundaries.
Note, the proposal is to add "tracker" to the list of trackers, not replacing the existing trackers on the list.
(b) the ISP kept enough torrent files on the tracker to answer a lot of the demand (See above -- trackers don't store .torrent files, they only accept the infohash calculated from the contents of a .torrent file.)
(c) subscribers were willing to use the ISP's tracker instead of a remote one. (see above -- we're talking about adding to the list of trackers, not replacing it.)
I completely agree that the subscriber should have the free choice of using such a tracker. If a greater proportion of more local peers generally improves download speed -- which it should -- then I suspect downloaders will choose to do so unless ISPs also start to mess with the privacy of their users.
But the gain would be only small, This is essentially what Pando is doing to get 250% to 800% improvement. This is also what the Azureus plugin Ono strives to do and they claim huge speed improvements, too.
But you've already misunderstood the proposal, and your theory was based on that misunderstanding. Perhaps you'll see more clearly now.
For more reduction of peering costs, the ISP would have to cache content, or in Bittorrent terms, provide seeds. How much reduction is needed before it is enough? Answer: we don't know. The ISPs are all complaining that P2P makers must reduce the impact but they've never said by how much, or when, or even where in the network (the last mile, the boundary gateway, upload, download). They've given absolutely no data. Oh, but they do call us Bandwidth Hogs -- so do they want to make piglets or do they really want to make bacon?
Yeah and no. Once a file is unpopular, the once-in-a-while normal download of that file isn't all that impactful. But when something is newly released, and several hundreds of ACTIVE connections are crossing the network boundary to get it, then having a cached copy and internal peers as data sources really can cut the demand on the boundary gateways.
Caching may be of concern to those who are using private systems, because caching would reduce the demand for their upload bytes and their ratio will suffer. That said, if this model becomes widespread, then administrators should reset their expectations about ratios anyway.
This would make ISPs the infringers, not merely conduits, in the case of infringing files. No more than YouTube is an infringer. DMCA take-down notice. Done, it's off the local cache. If my downloading of the content is within my rights, I can contest the DMCA notice (probably a waste of time because most uses are not fair use) or get the file without the benefit of the cache.
This does not add to or take away from a downloader's rights or a copyrightholder's rights in any way, nor does it help nor hinder a copyrightholder in protecting his rights.
It would still leave customers going off the network whenever they want something that's not on the ISP's tracker, or when they don't trust the ISP.
You're right! You can't make P2P a zero-impact technology, but that isn't the goal. The goal is to reduce its impact in a meaningful way -- and my proposal would do that.
--
Robb Topolski -= funchords.com =- Hillsboro, Oregon
HTTP is the new Bandwidth Hog...
|
