4 ms·

Why Can't Crypto Mining be used to Power other things? (e.g. video rendering)

3 points by cxejohns 7y ago

I was just reading about various crypto altcoins and wondering, would it be possible to turn the processing used for bitcoin mining into something more useful? If so, what? And if not, why not?

jki275 7y ago
primecoin did this. There might have been others.
cbhl 7y ago
The last time I considered this: - Most grid/cluster-based video rendering systems assume trust. Shared filesystems, central nodes that allocate work, users submit jobs to the central node. This trust allows for lower overhead, as each unit of work only needs to be processed once. (By comparison, with blockchains, you need to wait for n copies of the work to be done and compare them to verify that a bad actor hasn't submitted garbage; these are "confirmations".) - One key property of cryptocurrencies is resilience in the face of bad actors. In particular, you want each node to independently be able to come to consensus without trusting a central node. (Otherwise, you would just devise a scheme that trusts a central node, such as a Bank.) So if you want to replace Proof of Work or Proof of Stake with a "useful" task, you have to devise a scheme that retains the property of "consensus without trust", while also providing "useful" work (that is, correct video rendering results, in a reasonable response time). I'd look up "51% attack", as well as the documents describing the design of scrypt (https://www.tarsnap.com/scrypt.html https://www.tarsnap.com/scrypt.html) and other password hashing functions to get a sense of the attack modes you have to design for. - Assuming you come up with a good scheme, you then need to convince the rest of the world to adopt it. (Note that some actors have hardware that is optimized for the task defined by Bitcoin. Also, if your user base is too small, then it's trivial to 51% attack by just buying more computers than you.) I don't think it's impossible; I just don't think it's cost-effective (versus just buying and building a grid/cluster with trusted nodes).
- cxejohns 7y ago
  Thanks for the breakdown. If I have it right, it sounds like block creation is an inherently 'useless' process. One could make it useful, but not in a truly integral way, given how normal, centralized data processing works.
  - cbhl 7y ago
    I don't know if 'useless' is the right word. My understanding is that a block is simply a page of the distributed ledger (that is, a list of transactions), plus a hash preimage problem. (What value do I append to this data, such that the hash function returns some pre-specified value?) That such a computation is expensive to compute (but cheap to verify) is core to some of the properties of cryptocurrency. The alternatives provided by normal data processing are cheaper and faster (if you have trust). If you make the computation cheaper, then financial transactions in the ledger are reversible. But if you make the cheap task expensive, then people will just pay for regular grid computing because it's cheaper.
rajnathani 7y ago
There's Gridcoin [1] which states "which securely rewards volunteer computing performed on the BOINC platform". Though I personally am not aware of how successful it has been so far. With regards to video rendering, cbhl correctly brings up the trust and latency issue. I feel the latency issue is particularly important, as even AWS recently announced their Local Zones data centers [2] for "highly-demanding applications that are particularly sensitive to latency" with an example clearly stated for the large video production studios in LA: "3D modeling & rendering, video processing (including real-time color correction), video streaming, and media production pipelines". Personally, I feel there are latency insensitive applications such as protein folding which could benefit from proof-of-work, and there does seem an attempt at a crypto coin [3] for this. However I'm not too sure as to whether it's the lack of good execution for these projects not taking off, or if there's a more intricate sociological or other explanation for it. [1] https://gridcoin.us/ https://gridcoin.us/ [2] https://aws.amazon.com/blogs/aws/aws-now-available-from-a-local-zone-in-los-angeles/ https://aws.amazon.com/blogs/aws/aws-now-available-from-a-lo... [3] https://foldingcoin.net/ https://foldingcoin.net/

3 points by cxejohns 7y ago

I was just reading about various crypto altcoins and wondering, would it be possible to turn the processing used for bitcoin mining into something more useful? If so, what? And if not, why not?

jki275 7y ago

primecoin did this. There might have been others.

cbhl 7y ago

The last time I considered this: - Most grid/cluster-based video rendering systems assume trust. Shared filesystems, central nodes that allocate work, users submit jobs to the central node. This trust allows for lower overhead, as each unit of work only needs to be processed once. (By comparison, with blockchains, you need to wait for n copies of the work to be done and compare them to verify that a bad actor hasn't submitted garbage; these are "confirmations".) - One key property of cryptocurrencies is resilience in the face of bad actors. In particular, you want each node to independently be able to come to consensus without trusting a central node. (Otherwise, you would just devise a scheme that trusts a central node, such as a Bank.) So if you want to replace Proof of Work or Proof of Stake with a "useful" task, you have to devise a scheme that retains the property of "consensus without trust", while also providing "useful" work (that is, correct video rendering results, in a reasonable response time). I'd look up "51% attack", as well as the documents describing the design of scrypt (https://www.tarsnap.com/scrypt.html https://www.tarsnap.com/scrypt.html) and other password hashing functions to get a sense of the attack modes you have to design for. - Assuming you come up with a good scheme, you then need to convince the rest of the world to adopt it. (Note that some actors have hardware that is optimized for the task defined by Bitcoin. Also, if your user base is too small, then it's trivial to 51% attack by just buying more computers than you.) I don't think it's impossible; I just don't think it's cost-effective (versus just buying and building a grid/cluster with trusted nodes).

cxejohns 7y ago
Thanks for the breakdown. If I have it right, it sounds like block creation is an inherently 'useless' process. One could make it useful, but not in a truly integral way, given how normal, centralized data processing works.
- cbhl 7y ago
  I don't know if 'useless' is the right word. My understanding is that a block is simply a page of the distributed ledger (that is, a list of transactions), plus a hash preimage problem. (What value do I append to this data, such that the hash function returns some pre-specified value?) That such a computation is expensive to compute (but cheap to verify) is core to some of the properties of cryptocurrency. The alternatives provided by normal data processing are cheaper and faster (if you have trust). If you make the computation cheaper, then financial transactions in the ledger are reversible. But if you make the cheap task expensive, then people will just pay for regular grid computing because it's cheaper.

rajnathani 7y ago

There's Gridcoin [1] which states "which securely rewards volunteer computing performed on the BOINC platform". Though I personally am not aware of how successful it has been so far. With regards to video rendering, cbhl correctly brings up the trust and latency issue. I feel the latency issue is particularly important, as even AWS recently announced their Local Zones data centers [2] for "highly-demanding applications that are particularly sensitive to latency" with an example clearly stated for the large video production studios in LA: "3D modeling & rendering, video processing (including real-time color correction), video streaming, and media production pipelines". Personally, I feel there are latency insensitive applications such as protein folding which could benefit from proof-of-work, and there does seem an attempt at a crypto coin [3] for this. However I'm not too sure as to whether it's the lack of good execution for these projects not taking off, or if there's a more intricate sociological or other explanation for it. [1] https://gridcoin.us/ https://gridcoin.us/ [2] https://aws.amazon.com/blogs/aws/aws-now-available-from-a-local-zone-in-los-angeles/ https://aws.amazon.com/blogs/aws/aws-now-available-from-a-lo... [3] https://foldingcoin.net/ https://foldingcoin.net/