Jump to content

Webseeding help with minimum download rate (QoS)


hermanm

Recommended Posts

Hi, I realize as soon as the pieces are available from Bittorrent peers, the µTorrent client stops asking for data from the web server. My download was not fast from the web server, but it was steady. Now, my speed is only 5 kB because of my lousy peers. I would be nice if the web server could provide a level of download speed to match up my upload speed. Not a 12:1 ratio mind you, but a 1:1. This would not hammer the web server but still provide a steady way transfer rate.

The alternative would be to block all peers manually via ipfilter.dat and make it seem like there are no other eligible peers.

Link to comment
Share on other sites

... HTTP seeds are valid peers. I guess you don't know what behaviour you're experiencing >< What I got out of http://getright.com/seedtorrent.html and http://www.bittornado.com/docs/webseed-spec.txt was the server had to control how and when it connected to the swarm.

When webseeds are added into a torrent, uT treats them like an always-on peer... with the same benefits and limitations such as ipfilter as you mention above.

Link to comment
Share on other sites

Hi, I realize as soon as the pieces are available from Bittorrent peers, the µTorrent client stops asking for data from the web server. My download was not fast from the web server, but it was steady. Now, my speed is only 5 kB because of my lousy peers. I would be nice if the web server could provide a level of download speed to match up my upload speed. Not a 12:1 ratio mind you, but a 1:1. This would not hammer the web server but still provide a steady way transfer rate.

The alternative would be to block all peers manually via ipfilter.dat and make it seem like there are no other eligible peers.

Can you provide us with the torrent where this occurs?

Link to comment
Share on other sites

@jewelisheaven: Web-seeds as implemented in µTorrent (and as described by the GetRight spec) are not true peers. The server has absolutely no control. No, the server isn't banned when there are other peers. It's simply not queried/used when µTorrent can see another source in the swarm for the piece it wants.

When µTorrent is downloading from a web-seed, it's using basic HTTP/FTP transfers (segmented downloads to get specific pieces of the file).

Link to comment
Share on other sites

µTorrent will request a whole piece (e.g. 2MB, 4MB, etc). In my case, I was trying to retrieve files my file server via HTTP. I alread had pieces from a my torrent, so it was best to piece it togethre via web seed. I just had to block all public peers since the speeds were dismal. To give you an idea, my file server ahd to send 80GB just to receive 4GB.

I guess what I really want is a private form field in µTorrent where I can retrieve my own files from a HTTP or FTP servers that will piece together my torrent faster. That field wouldn't observe any webseeding rules currently in place.

Link to comment
Share on other sites

jewel - See post #6 by Ultima - I think he explains it perfectly.

@jewelisheaven: Web-seeds as implemented in µTorrent (and as described by the GetRight spec) are not true peers. The server has absolutely no control. No, the server isn't banned when there are other peers. It's simply not queried/used when µTorrent can see another source in the swarm for the piece it wants.

When µTorrent is downloading from a web-seed, it's using basic HTTP/FTP transfers (segmented downloads to get specific pieces of the file).

Alus, I sent you the .torrent file via e-mail. The .torrent file does have the webseed embedded, but I can send you a link to a private server if you want.

Link to comment
Share on other sites

Wouldn't it be better if µT took advantage of web seeds automatically whenever the swarm sucks? If it were an option, people might just prefer web seeds even with a healthy, speed maxing swarm

Yeah, if it were an option, then it would mean people would unnecessarily hammer servers for data when the piece can already be found in the swarm. Web-seeds should be treated as backup seeds, not as true seeds.

The most accurate measurement of swarm suckage is availability (or lack thereof -- availability < 1.0). Web-seeds already take care of that. That a swarm is "slow" would not be accurate, as "slow" is subjective.

Link to comment
Share on other sites

Archived

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

×
×
  • Create New...