Bitcoin Forum

Hello everyone,

I was wondering, what would you all think about a blockchain to facilitate peer-to-peer computing?

Basically, people ("clients") would buy this computing token, connect with peers ("providers") onto a blockchain, and then make a "deal" for trading X computing power (e.g. AMD Ryzen 5900X or NVIDIA 3090) for Y time via the blockchain with the token as payment.
After, a client would give a review that contributes to the provider's reputation and voila, the exchange has completed. In addition, providers will undergo computations randomly as an audit to receive a reward.

I think people would be attracted to the low prices. And while proof that providers don't look at data is impossible to attain, I think the pricing benefit of perfect competition (whereas today's high-performance computing industry is more like monopolistic competition because services are all slightly differentiated) would lure non-sensitive workloads to this ecosystem.

TLDR a Filecoin but for computing.

What do you all think?

Thanks!

Hello

It's clear about the utility of the coin which will be used for payments, but can you please explain how users will be able to share/exchange computing power between them? Do you have a clear vision/architecture on how this can be reached?
Honestly i can't figure out how this is similar to filecoin.

Maybe in the Project Development (https://bitcointalk.org/index.php?board=12.0) board you can find better replies from veteran users.

It's just an idea so I'm not 100% solid on the architecture yet.

It would probably go something like this:
- Customer uploads a Docker image to a Docker registry
- Customer connects with other online peers to make a deal
- Once a deal is made, tokens on the blockchain are locked (a built-in escrow service)
- The provider begins running the Docker container until the time is up
- If the provider has stayed online this entire time, once the time is up, the token is unlocked, and the customer can flag the host if the provider didn't fulfill the request well (e.g. the provider's servers were much slower than they should be due to temperature throttling)

That's the general gist of my idea. There would be years of development needed to work everything out, but this idea has just been bugging me for a while so I just want some idea feedback :)

You're right about the project development board, thanks! Will definitely do in the future :)

There are already some centralized business that are doing this. If they can, a blockhain would effectively and efficiently do it. With the overall increase in network speed and current short supply of graphics card, a coin at the moment could be an instant success. But the execution should be near perfect as computing is a resource hungry process and the providers would want instant reward for their dedication.

If I am not mistaken even Siacoin which is one of the oldest altcoins that people like to merge mine with Bitcoin also did this, not so much on computing but on file space. And then we have all the folding projects at home like Seti and the Human Genome project that did all something familiar, giving folding coin back as a reward even I used to do it:)

as reference of algorithm might be on acceptation for a trader and investors the offers from developer might helps on extent to deliver use with difference on consequence as the dispute of funds on submission that more to keep on manage as moderating use with the release of data and hashes from the blockchain server.

Sounds good.

Meanwhile, it doesn't only require a long time to realize it but a strong genuine team. The concept sounds good as well, however i still can't see a strong enough reason to build such a sophisticated service; the market of sharing computing power is in considerable growth, however the great part of it is made by big companies to undertake their essential processes, so if those companies don't adopt cryptocurrencies, they won't rent additional power using decentralized apps.
As the development will take years, maybe by the time of the final product release the market will change.

Very true. Storj didn't take off compared to Filecoin because Storj takes 15 months for miners to fully realize their earnings. So it's a lot to worry about, someone would definitely need a lot of resources to get a project like this working, and it'd need to be near perfect.