News for hackers. Stuff that matters.
Re:An engineer's observations on Web3 and its possibilities by ahmedfromtunis on Tuesday November 23, 2021 @09:02PM (#29326192)
I totally agree with you. The only Blockchain-first technology I got excited about was a system where people can put their free disk space for rent. They get paid in some coin, and that is how the coin gets generated. Sure, you can do it without the Blockchain, but the tech has some real added value.

Alas, for whatever reason, it never got deployed.

Re:An engineer's observations on Web3 and its possibilities by phire on Wednesday November 24, 2021 @12:29AM (#29327176)
The biggest problem with this idea is proving data replication and guaranteed retrievability to the users in a trustless manner.

With a traditional cloud provider, you trust them, upload your data and they deal with the rest, backing it up internally and guaranteeing retrievability with their reputation.

But with blockchain p2p storage, it's really hard to provide the same guarantees. If you only upload the file once and let the blockchain replicate it, then a malicious (or greedy) entity can manipulate the blockchain and control all replicas (backed by a single disk copy). So your software must encrypt multiple copies of your data and upload seperate replicas with different encryption keys.

But that still leaves you open to a very malicious entity going out of it's way to track all uploads from a node, and monopolising control over them, then deleting them when you try to download them. To actually provide the same guarantees that a cloud provider does in a trustless way, your local software will also have to continually download chunks to make sure they are still retrievable, and you might also need to use the tor network for all accesses.

With all that extra complexity, there is a lot to go wrong, and a traditional cloud provider might end up being the cheapest option. I used to agree with you, but after trying to design such a system with these trustless guarantees myself, I've come to the conclusion that there is no good use case for blockchain, outside of exchanging value.

Re:An engineer's observations on Web3 and its possibilities by aspenmayer on Tuesday November 23, 2021 @09:25PM (#29326332)
It’s possible you’re thinking of Chia, IPFS, or FileCoin, all of which are “live” and “deployed” currently.
Re:An engineer's observations on Web3 and its possibilities by uncomputation on Tuesday November 23, 2021 @09:47PM (#29326450)
Filecoin is “live” only technically. There is no way for individuals to contribute, which seems to be intentional as the leadership is looking for corporate support first. It currently has no real use cases. I say this as someone who is also excited by the idea.
Re:An engineer's observations on Web3 and its possibilities by aspenmayer on Tuesday November 23, 2021 @10:12PM (#29326580)
Perhaps you can clarify the relations between IPFS and FileCoin? I was under the impression that they were related projects of Protocol Labs, with IPFS being a dependency of FileCoin.
Re:An engineer's observations on Web3 and its possibilities by shepardrtc on Tuesday November 23, 2021 @10:27PM (#29326649)
If you want something similar to Filecoin that is live and in use, check out http://siasky.net/

Siacoin has been around for a while and approaches things differently. Plus its already being used by people.

Re:An engineer's observations on Web3 and its possibilities by phire on Wednesday November 24, 2021 @12:16AM (#29327126)
IPFS isn't blockchain. and Chia doesn't allow people to rent out storage space, it just lets them mine with HDDs instead of GPUs.

The other one I'm aware of is Sia.

Re:An engineer's observations on Web3 and its possibilities by pcthrowaway on Wednesday November 24, 2021 @07:51AM (#29329488)
Arweave is the big one I know of