Mindful of recent developments in the world, is there any technology to distribute hosting, so the data is shared between many small peers around the world and requests to a site are load balanced across all or some of them and split over "safe" geographies? For common good sites like Wikipedia, The Internet Archive, that sort of thing?
Surely I'm not the only person to have thought of this 
Boosts for reach gratefully received 
### ngosang/trackerslist | #trackerslist
Updated list of #public #BitTorrent trackers
QUESTION
does anybody have any tips on how to deal with stalled #bittorrent downloads from @internetarchive? these are all public domain movies from the 1910s, 20s and 30s and they stop 99.5 or 99.9%. some seem to be complete at 99.9% but most don’t work.
bittorrent can be very frustrating like that. i force checkout and download and nothing happens.
I've been running the #Kiwix seeder scripts on a small seedbox (~4 TB) for a month now. It's been a charm.
https://github.com/kiwix/seeder/blob/main/README.md
Can't wait to get more leechers!
Want to help to spread the latest ZIM files using #bittorrent? It never has been easier to create a super-seeder of (part of) our full catalog of offline archives! Everything is automated and you'll be kept in sync with the Kiwix library! You just need a GNU/#Linux, a bit of storage & bandwidth. Checkout https://github.com/kiwix/seeder #FOSS #EDTECH
There is this big glaring problem in the #bittorrent v2 spec that basically defeats the whole purpose of the spec. So v2 introduces per-file merkle trees, which rocks. That ostensibly means that you can match identical files in different torrents, and cross-seed them e.g. if someone has a single item but another person has a larger pack that includes that item. That could also be useful for mutable torrents that contain a hash tree of versions like git. The problem is that peer matching is done on the top-level infohash - which includes non-tree data - and peers are specifically required to sever connection without a matching infohash. Maybe it would be expensive for trackers to match peers by file hashes, but for DHT peers it makes perfect sense to advertise what file hashes you have for cross-seeding - that's sort of how the entire keyspace in kademlia works.
special invite to any #BitTorrent heads out there who want to contribute to our uploaders guide, seeders guide, torrent creation guide, or want to help get the tracker scraping and announcing part off the ground. if you ever wondered what it would be like if there were better social overlays to bittorrent where trackers could announce around peers based on explicit social relationships, curate collections collaboratively across tracker instances... that's the kinda sicko shit we're into
ISP Must Unmask 100 Alleged #BitTorrent Pirates in #RIAA Lawsuit
https://torrentfreak.com/isp-must-unmask-100-alleged-bittorrent-pirates-in-riaa-lawsuit-250219/
Raubkopien für KI-Training: Meta nutzt BitTorrent – haben aber "nicht geseedet"
Um an immense Datenmengen für das KI-Training zu kommen, hat Meta ganze Schattenbibliotheken heruntergeladen – per BitTorrent. Illegal sei das nicht gewesen.
@gerowen Nice! Have you considered to use https://github.com/kiwix/seeder for sync and sharing via #BitTorrent?
@skye @zacwest This one's over a decade old, and still as relevant as ever: https://xkcd.com/488/
In short, if your collection goes poof; fear not — there's a friendly ebook #bittorrent community near you.
https://www.e-flux.com/journal/151/652643/taking-ai-into-the-tunnels/
Interesting read. First a critical intro to #ML and #AI , then discusses #bittorrent , Piratbyrån and similar hacker/activist movements in Sweden (which were part of my activist education). Makes the connection between radical politics, Tiqquns zones of opacity and federated learning in #ML (distributed models and private but shared datasets).
Who wants to help out with using #bittorrent v2, which has file-based merkle trees, to be able to combine the smaller chunks of at-risk data scraped by many independent people, none of whom can store the whole thing?
It's the first time I've seen a real use case for Bittorrent v2, a cool p2p and interface design problem (also need to make the site capable of representing ragged, overlapping parent/child relationships, make it possible for people to "claim" chunks of a dataset as ones they are in the process of scraping), and all about making sure the fascists can't own the informational space and control what is allowed to be known.
To help sharing knowledge, and help our North American friends, if you have 104 GB free and some bandwidth, you can download Wikipedia
via BitTorrent and keep it seeding.
wikipedia_en_all_maxi_2024-01.zim
https://library.kiwix.org/#lang=eng&category=wikipedia
BitTorrent clients :
- qBitTorrent https://www.qbittorrent.org
- Transmission https://transmissionbt.com
Pour participer au partage de la connaissance, et aider nos amis nord-américains, si tu as 104 Go de libre et de la bande passante, tu peux télécharger Wikipédia
via BitTorrent et surtout laisser le fichier en partage.
wikipedia_en_all_maxi_2024-01.zim
https://library.kiwix.org/#lang=eng&category=wikipedia
Clients BitTorrent :
- qBitTorrent https://www.qbittorrent.org
- Transmission https://transmissionbt.com
"#Piracy bad
It hurts #artists and #writers
Well, we shaft them, and who it hurts is us really
Anyway, piracy bad"
Also corporatocracy:
"New #Meta Emails Reveal That the Company Downloaded 81.7 TB of #Copyright ed #Books via #BitTorrent to Train Its #AI Models"
@hbm Nice! We might have a new interesting tool for you https://github.com/kiwix/seeder. Still in beta testing, we need early adopters! #BitTorrent
ok galaxybrain idea number 2:
Private trackers usually work by autogenerating torrent files with unique announce urls to be able to ensure torrent files aren't shared and to assign credit for uploads/downloads.
In addition to having generic RSS feeds by tag, we could also have people sign up for tags as a part of their account, as well as indicate how much space they have available, and then we could autogenerate a unique RSS feed per account that load balances torrents by priority.
This way, each torrent has some minimum number of seeders instead of needing to sign up for a potentially unbounded amount of downloads in a tag feed or needing to manually distribute over a collection of torrents.
We would either need a very large number of seeders for this to work in the aggregate, or else make priority a real-valued number based on voting, but even partially, being able to subscribe to a feed that is sensitive to seed coverage and preferences is not something i have ever seen done in bittorrent.
https://codeberg.org/sneakers-the-rat/sciop/issues/32
i just had an extremely galaxybrained #p2p idea - so ingress bandwidth is free/cheap compared to egress on various cloud providers. for this #bittorrent archival tracker website i'm writing, how dope would it be to be able to provide a list of URLs to some dataset/website/etc., the server then streams that data, hashes it, discards it, generates a torrent file with the common domain as a web seed... and then the rest of this site kicks in where you have people who follow torrent .rss feeds with seedboxes to autoarchive them.
https://codeberg.org/sneakers-the-rat/sciop/issues/31
