Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The Pirate Bay's 20th Anniversary

torrentfreak.com

Post reply on HN
Print view View original

121–124 of 124 posts

Re: The Pirate Bay's 20th Anniversary

#121

Post by indigochill »

didntcheck wrote: post #97
indigochill wrote: post #95

Earlier quoted context omitted.

> "Distributed cloud backup" Relatedly, when Docker started charging open source projects for Dockerhub storage, a friend and I started brainstorming a torrent registry driver (although we didn't find a way around the project still needing to provide a signature for verification at least, even if it didn't need to maintain hosting for the image itself) Nothing to show for it just yet, but point being I think this "di…

It sounds like IPFS (or other content-addressed distributed data stores) may be more what you're after there than Bittorrent

IPFS would be another potential solution. What drew us to Bittorrent for this use case is the (at least original) ethos of clients both downloading and uploading cooperatively such that as long as there's any community that values the image, it will continue to be available (which would basically be true for an IPFS implementation as well).

But the registry implementation in the "distribution" project takes storage drivers, so an IPFS storage driver is certainly a possibility. In part for me going with Bittorrent was also just an excuse to get more familiar with the technical details of Bittorrent.

Re: The Pirate Bay's 20th Anniversary

#122

Post by rakoo »

indigochill wrote: post #120
londons_explore wrote: post #96

Earlier quoted context omitted.

What's the signature problem? For a read-only registry, signatures can all be included in the torrent. For a versioned registry, all the info can be stored in any cloud key-value store - eg. The bittorrent DHT

It could just be a lack of understanding on our part, but here's the scenario we're thinking about: Say the Redis team seeds a redis:1.0 image. So far, so good. But what if a nefarious hacker seeds a redis:2.0 image? In a traditional registry, the auth system for the registry ensures that only those who are authorized can write to a particular registry. But in (at least our understanding of) a swarm, there is no auth…

It has always been possible to store immutable data in the DHT, that's partly what it was built for. "This is a sha1 of a content, I want to know what content is behind"

There is a BEP (Bittorrent Extension Protocol) for storing mutable data: like all good things in computing, it's just an indirection. Create a pair of asymetrical cryptographic keys for a specific (mutable) content and store the hash of the public key. People will search for that hash, get to some content that is any bytes you want but signed with the private key.

What you want is not exactly that, but more of a naming system. The easiest would be to let each image owner be their own source of truth: the redis team puts up an HTTPS server with a well known path, say "/.well-known/docker?identity=redis:2.0" and the reply includes the magnet link. My node would cache the association under the assumption that it never changes (it shouldn't; a change in the image implies a change in the name).

As you yourselves discovered, the swarm doesn't solve the naming issue (ie Zooko's triangle), there needs to be a third-party here

Re: The Pirate Bay's 20th Anniversary

#123

Post by soulofmischief »

daemoens wrote: post #94
soulofmischief wrote: post #88

Earlier quoted context omitted.

It's only fair to then reflect on the almighty dollar and the scams, criminal deeds and atrocities which it has also enabled. If you still find utility in the dollar bill after such reflection, then the argument doesn't hold up to cryptocurrency.

The difference is that crypto has only had a negative impact on my life. It's more of a mixed bag with the dollar.

What is the negative impact?

Re: The Pirate Bay's 20th Anniversary

#124

Post by anacrolix »

rakoo wrote: post #122
indigochill wrote: post #120

Earlier quoted context omitted.

It could just be a lack of understanding on our part, but here's the scenario we're thinking about: Say the Redis team seeds a redis:1.0 image. So far, so good. But what if a nefarious hacker seeds a redis:2.0 image? In a traditional registry, the auth system for the registry ensures that only those who are authorized can write to a particular registry. But in (at least our understanding of) a swarm, there is no auth…

It has always been possible to store immutable data in the DHT, that's partly what it was built for. "This is a sha1 of a content, I want to know what content is behind" There is a BEP (Bittorrent Extension Protocol) for storing mutable data: like all good things in computing, it's just an indirection. Create a pair of asymetrical cryptographic keys for a specific (mutable) content and store the hash of the public ke…

https://github.com/anacrolix/btlink
Post reply on HN