Make stats files digitally signed

[RoadTrain]

Hello. I am currently developing a project that can bring a substantial computing power to F@H by rewarding users for solved units.
Reward is proposed to be calculated according to earned poits.
Though there's a problem because the system should be distributed and buit upon a network consensus.
The solution I propose for F@H is to provide a digital signature to stats files so every node can check their legitimacy.

I'd like to discuss options.
Thanks.

[Jesse_V]

Can you provide more details of the system/project you're proposing? I think it's important to know more.

I don't understand the need to digitally sign the flat files. The server are extremely careful with rewarding the promised credit, and there's numerous checks for scientific integrity.

[P5-133XL]

Are you implying that you intend to create an incentive to hack Stanford's DL site to modify their flat files? Are you intending to distribute the flat files to your nodes (creating the need to verify legitimacy) or will each node DL their own copy from Stanford?

[RoadTrain]

I'll explain shortly.
The project I am planning to develop is some sort of a cryptocurrency. If you know about Bitcoin you may also know how much computing power is actually wasted.
In my draft I'd like to add (more precisely reward) useful computations. Theoretically it could be done in different ways, but most of them are impossible due to computing software's proprietary nature (and a lot of scientific work needed to organize truly decentralized computing).

So I've come up with a solution. Users of the proposed cryptocurrency will be rewarded according to their points. If you're familiar with Bitcoin's protocol it might look like this:

1. User installs F@H and sets his username to the address (public key) he wants to receive reward to and starts folding.
2. one of the network nodes downloads user_summary from Stanford and distributes it between others like a block. It can be done like once a day or so.
3. after there's a "stats block" in the blockchain, the User is able to create his own blocks based on points he earned.

The essential thing is network consensus, in 2 and 3 any node should be able to confirm the blocks' legitimacy. And basically they should confirm stats' legitimacy. This can be done by checking singature.

This system looks very tricky to me, but I have a lot of thoughts on how to improve it. And this solution requires minimum changes on F@H side.

If Stanford wants to collaborate, a more secure and robust solution can be implemented.

[bruce]

What is "concensus" in this context?

FAH is not like BOINC

[RoadTrain]

Consensus means that all nodes must agree on block's legitimacy to maintain blockchain consistency. The flat file has to be signed in order to prevent someone from distributing modified version with modified points or so.

[P5-133XL]

The biggest issue I see is convincing Stanford to ever accept the liability of basing your new crypto-currency on their data. They are going to get sued, if they get something wrong. They've had to back-credit people before. So mistakes have happened in the past and are likely to happen in the future. Now we know that there is a potential downside, what's their monetary upside that is going to more than pay for those potential lawsuits? I really do not see them ever being willing to accept liability.

[RoadTrain]

Fair point. I'll try to rethink the whole concept.
Thanks anyway.