Hacker News new | past | comments | ask | show | jobs | submit login

There was a discussion about the world's oldest torrent, where others and I mused about solutions to the problem of ephemeral seeds: https://news.ycombinator.com/item?id=10962305

I suggested that one could maybe bridge Bittorrent and IPFS, then someone suggested that IPFS would be ideal for being the 'canonical' source of the file while the other protocols (HTTP, BT, WT) are merely ways to access the file.




The Internet Archive seeds all of our items' torrents since 2012. IPFS is a more straightforward way for us to be sure we've got independent seeds of all of our content. Got 25 petabytes of disk?


I've always wanted to write a distributed ArchiveTeam torrent server that, instead of distributing crawlers for ingestion, equally distributed IA torrents to distributed end-user storage (in a VM most likely). If seeding for a torrent dropped below a threshold, that torrent would be advertised to other client nodes to download and then serve themselves. Client VMs could come and go, and the announcer would always ensure remote storage was handled in an orderly fashion.

Perhaps this year is the year!


Uhhh, 20 GB free...

I'm just a third-party, but if you were to get in touch with the IPFS team to host a limited amount of IA content as a proof-of-concept, I'm sure it would be a great publicity for both of your organizations.


We're in touch with them, yes, and we have a proof-of-concept node.


> then someone suggested that IPFS would be ideal for being the 'canonical' source of the file while the other protocols (HTTP, BT, WT) are merely ways to access the file.

Hey! That was me! I still believe its a great idea ;)




Consider applying for YC's Spring batch! Applications are open till Feb 11.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: