Difference between Proof of Useful Work and Internet Computer Consensus?

https://wiki.internetcomputer.org/wiki/Proof_of_Useful_Work
Above is an explanation of Proof of Useful Work compared to PoW. Does the Internet Computer Consensus include Proof of Useful Work? Are Internet Computer Consensus and Proof of Useful Work different? Do Internet Computer Consensus and Proof of Useful Work complement each other?

The following is my understanding.
Proof of Useful Work refers to dedicated hardware and NNS. In other words, Internet Computer Consensus and Proof of Useful Work are different and complement each other to ensure security.

2 Likes

PoUW is described as an alternative consensus algorithm that involves a blockchain being produced by dedicated hardware called “node machines” that are of very similar, standardized specifications. These machines don’t do hashing but rather produce and process blocks of transactions that represent smart contract computations. The Network Nervous System is described as a key ingredient of the scheme, which creates a highly secure network in which the Internet Computer runs on a sovereign network of dedicated hardware formed from node machines. In PoUW, the repetitive hashing work of PoW has been replaced by useful smart contract computation.

4 Likes

@Forreal
Thank you. I understood more about Proof of Useful Work.

The 2021 Internet Computer Consensus (ICC) has been written. But there is also Proof of Useful Work. Are Internet Computer Consensus and Proof of Useful Work different? If someone asks me what is the consensus algorithm for Internet Computer, should I say PoUW?

2 Likes

Here’s a crack at POUW and Internet Computer Consensus.

Proof of Useful Work is the ability of a distributed network to execute smart contract computations on multiple nodes and prove that the computations were correctly calculated.

Internet Computer Consensus is the mechanism to ensure that multiple nodes agree that the smart contract computations were correctly carried on honest nodes.

4 Likes

Or in very laymen terms useful work is the servers executing code and the consensus is the part where others agree it was done correctly. Not to be confused with mining PouW that being apples and cars not even oranges.

3 Likes

@mparikh @ZackDS
Thank you for your response.

I have organized my thoughts. Do you all think the following is correct?

PoS Issues

  • Most of the nodes are running in the cloud
  • Easy to attack, as long as you can prepare enough cryptocurrency for staking.
  • There are recommended specifications for nodes, but they are not enforceable.Therefore, it takes a long time to finalize the finality of a block.

PoUW is a way of thinking that directly solves the above issues. NNS realizes PoUW. Internet Computer Consensus (ICC) by itself cannot directly solve the above problems. The ICC is responsible for consensus of blocks and consensus of the status of smart contracts after they are executed. PoUW and ICC work complementary to each other.
Also, the resolution of issues improved security and TPS and scalability, etc.

2 Likes

I think you should separate out implementation from design.

  1. There’s nothing to prevent POS nodes to be run in-premise/dedicated. The economics of ETH validators nodes just make economical sense to run in cloud.

  2. The ease of attack is not a clear differentiating factor. You could spend 350 billion USD on BTC as well. You could control the NNS in theory through the same logic.

  3. Again some POS implementations do require stringent responses in specific time frame. So if your node is not upto stuff, you would get slashed if you couldn’t produce a response in time.

HTH

2 Likes

In the case of IC PoS stuff you mentioned do not apply. Subnets takes it to a whole new level. You and others need to take a step back from what you got used to in the so called “Blockchain technology” . One just can not compare the way thigs works with ETH chain for e.g POS also no even close to FLUX who advertise PoUW and is a whole different approach with half of the network mining with graphics cards and the other half staking just to be able to deploy stuff with Docker (that in it’s own is not a bad thing just don’t see it fit for blockchain).

1 Like

@mparikh @ZackDS
Thank you for your response.

I have organized my thoughts again. Do you all think the following is correct?

Operation
PoS operational issues

  • Most of the nodes are running in the cloud
  • There are recommended hardware specifications for nodes, but they are not enforceable. Therefore, it takes a long time to finalize the finality of a block.
  • Easy to attack, as long as you can prepare enough cryptocurrency for staking.
    The above are operational issues and cannot be resolved by the ICC.
    PoUW proposes a way to solve the above problems.
    However, PoUW only shows a way of thinking (operation/design) and does not guarantee a solution; it depends on implementation.

Implementation
NNS is an implementation of PoUW; It selects nodes based on node provider, data center location, etc. This solves the “PoS operational issues” mentioned above.

Supplemental Information

  • The ICC is responsible for block consensus and status consensus after executing a smart contract.
  • NNS and ICC work in a complementary manner to each other, improving security and TPS, scalability, etc.
  • NNS improves on “Easy to attack, as long as you can prepare enough cryptocurrency for staking”. NNS chooses its nodes based on node providers, data center geography, and other factors. However, it is theoretically possible to control NNS itself if only it can prepare enough cryptocurrency.
  • Some PoS implementations slash if no response is received within a specific time frame.
  • Internet Computer have own technology and architecture, which differs in many ways from common PoS and PoUW, making comparisons difficult.

Questions

  • Is NNS an implementation of PoUW, or is NNS a partial implementation of PoUW, or is PoUW implemented by NNS and ICC?
  • What is the purpose of Proof of Useful Work? Is it to reduce wasteful resource consumption like PoW?
  • Does NNS completely prohibit nodes from running in the cloud?
  • Who in InternetComputer removes or thrashes nodes that deviate statistically (too fast/slow nodes)?
1 Like

https://twitter.com/i/spaces/1OdJrzbRkjvJX

1 Like

PoUW is how ICC works to create a blockchain. The NNS is built on top of that to coordinate updates to the protocol. The NNS is built on top of the consensus mechanism and is not a blockchain/consensus mechanism itself

It’s to waste as little computation as possible while still getting security guarantees as close to PoW as possible

On a technical level, the NNS itself does no such thing. It’s just a mechanism to mange the network. What disallows nodes from running in the cloud is how people vote on proposals, and the processes that create proposals to approve/remove nodes.

Right now it’s all manually done based on alerts/metrics, but the goal is to have all of that automated eventually

7 Likes

@JaMarco
Thank you.

@Severin
Thank you. I understood it better.

You may like reading this article, please check