I wonder if someone can come up with some kind of distributed storage that isn’t insanely slow. Kinda like a CDN but on personal devices. I’m thinking like SETI@HOME did with distributed compute.
Edit: this is kinda like torrents but where the contents are changing frequently.
You should look up IPFS! It’s trying to be kinda like that.
It’ll always be slower than a CDN, though, partly because CDNs pay big money to be that fast, but also anything p2p is always going to have some overhead while the swarm tries to find something. It’s just a more complicated problem that necessarily has more layers.
But that doesn’t mean it’s not possible for it to be “fast enough”
And there’s a promising new IPFS-like system called Iroh, which should have a lot less overhead and in general just be faster than IPFS. It’s not quite ready to just switch to right now, but an enterprising individual could probably make something useful with it without too much work (i.e. months, not years).
I’m using it for a distributed application project right now, but the intent is a bit different than the IA use-case.
I wonder if someone can come up with some kind of distributed storage that isn’t insanely slow. Kinda like a CDN but on personal devices. I’m thinking like SETI@HOME did with distributed compute.
Edit: this is kinda like torrents but where the contents are changing frequently.
Something like torrents. Split the whole thing in small 5gb torrents.
You should look up IPFS! It’s trying to be kinda like that.
It’ll always be slower than a CDN, though, partly because CDNs pay big money to be that fast, but also anything p2p is always going to have some overhead while the swarm tries to find something. It’s just a more complicated problem that necessarily has more layers.
But that doesn’t mean it’s not possible for it to be “fast enough”
And there’s a promising new IPFS-like system called Iroh, which should have a lot less overhead and in general just be faster than IPFS. It’s not quite ready to just switch to right now, but an enterprising individual could probably make something useful with it without too much work (i.e. months, not years).
I’m using it for a distributed application project right now, but the intent is a bit different than the IA use-case.
Interesting, thanks