Never heard of it. How can someone quickly integrate this into their CD process? Looks like I can build a magnet link that points to a Web seed, but I guess I’d need to submit to some public trackers?
Afaiu if you have a web server, you can web-seed from it, no need for trackers. Archive.org has that for every upload, although they also have a tracker of their own.
However, specifically on Archive.org there’s not much point in torrents because they download from the web seeds before seeing any other peers. Though there’s the convenience of selecting the files to download in the torrent client.
I think you still need a tracker to discover peers. So unless it’s only a Web download (which would make this whole effort stupid), peers still need to be able to find each other.
Though my understanding of BT’s P2P protocol is pretty weak so I might be running on false assumptions.
Afaiu DHT, the distributed peer tracking, works for any torrent. It’s not as quick as a tracker, but it does its thing. I download some torrents by exclusively using the DHT.
Well, this is why people don’t use BT more for their FLOSS downloads.
I’ll do some more digging just because I’m curious, but without resources to show how people can leverage this for Web downloads most are not going to make the effort for little to no benefit.
Archive.org provide web seeding, and funnily the files typically quickly download from that before even finding any other seeds, defeating the point of torrents.
Meanwhile the couple times I tried downloading Linux via torrents, the client seemed to crash itself and the system by trying to talk to the thousand peers.
Also, it adds a whole 'nother layer of complexity. Now instead of just a Web server, you need a torrent client and all the CI/CD built up around it.
Not really though web seeding is a thing
Never heard of it. How can someone quickly integrate this into their CD process? Looks like I can build a magnet link that points to a Web seed, but I guess I’d need to submit to some public trackers?
I didn’t find a whole lot of resources on this.
Afaiu if you have a web server, you can web-seed from it, no need for trackers. Archive.org has that for every upload, although they also have a tracker of their own.
However, specifically on Archive.org there’s not much point in torrents because they download from the web seeds before seeing any other peers. Though there’s the convenience of selecting the files to download in the torrent client.
I think you still need a tracker to discover peers. So unless it’s only a Web download (which would make this whole effort stupid), peers still need to be able to find each other.
Though my understanding of BT’s P2P protocol is pretty weak so I might be running on false assumptions.
Afaiu DHT, the distributed peer tracking, works for any torrent. It’s not as quick as a tracker, but it does its thing. I download some torrents by exclusively using the DHT.
Well, this is why people don’t use BT more for their FLOSS downloads.
I’ll do some more digging just because I’m curious, but without resources to show how people can leverage this for Web downloads most are not going to make the effort for little to no benefit.
Archive.org provide web seeding, and funnily the files typically quickly download from that before even finding any other seeds, defeating the point of torrents.
Meanwhile the couple times I tried downloading Linux via torrents, the client seemed to crash itself and the system by trying to talk to the thousand peers.