A Proof-of-Work or PoW is a consensus algorithm in a blockchain network which is used to confirm transactions and produce new blocks to the chain.
Users on the network send each other digital tokens where the decentralized ledger consolidates all the transactions into blocks. Confirming transactions and arranging blocks is a responsibility borne by special nodes called miners, and the process is called mining.
Miners compete with each other to complete transactions on the network, for rewards. The main idea is to solve a complex mathematical puzzle that requires a lot of computational power to solve. The answer to the proof-of-work problem is called a hash.
How does a free blockchain work? In Bitcoin, the miners are rewarded, but in a blockchain to keep contracts, who will do the validation? Can a blockchain exist without miners?
A blockchain can exist without miners, proof-of-work, or a game theoretical competitive consensus algorithm. However, it won’t be a decentralized blockchain. It will require trusted third parties who will be the ones doing the validation.
You can consider a blockchain whereby the validation is done based on an alternative consensus algorithm such as proof-of-stake, but in that case, you still have a reward in the form of fees, for the risk taken in validating the rules and the cost of money locked up in that validation process.
You either have a reward system based on competition – proof-of-work, proof-of-stake, and variants of them – or you have a system without reward, but in that case, the validators must be trusted.
You re-introduce centralization and trusted third parties. That kind of blockchain can exist, but it is not really a blockchain. It is a database with digital signatures. A lot of banks are trying to do exactly that, create a centralized database and call it a “blockchain.”
It sounds very innovative, even though it isn’t.
Are blocks really an advantage in the blockchain? Do blocks formed in the blockchain need to exist? If every single transaction was validated by itself, wouldn’t that solve the block size problem? Wouldn’t it be faster to validate each transaction? Is it possible to create cryptocurrency without blocks?
Many of the systems that use a decentralized signing algorithm instead of a proof-of-work mining algorithm don’t really need blocks. Blocks are needed for proof-of-work because calculating the proof-of-work on the granularity of more than one transaction per second would mean increasing the rate at which blocks are found. Let’s say you had a blockchain where each transaction was its own block, had its own proof-of-work, and was chained to the previous transaction.
Miners would select a single transaction, calculate the proof-of-work, and broadcast that transaction. That is perfectly possible, but the problem is, the rate of orphaned transactions/blocks increases dramatically. Right now, Bitcoin sees a small fork that regularly occurs in the chain when a block is orphaned because two blocks were found more or less simultaneously within the ten-minute window.
If you make that ten-minute window a five-minute window, orphaned blocks increase significantly. You would need to account for that in the algorithm. There have been some attempts to shrink that window all the way down to fifteen seconds, which is what Ethereum does.
But in order to do that, Ethereum has a special mechanism for accounting for orphaned blocks, sharing the reward between multiple winners finding the proof-of-work. If you calculated the proof-of-work for singular transactions, and then increase the throughput rates of transactions, that would involve a thousand blocks per second, each one being just one transaction.
The problems with forking the network and synchronizing it becomes insurmountable. The only way you can do that is if you no longer care about competition between miners and proof-of-work. If the consensus algorithm is fundamentally different – we see that in distributed ledger technology where consensus is done through signing instead of mining, and the energy requirement is zero.
With that model, you could issue blocks as fast as you want, at which point you don’t need blocks anymore. You can just chain transactions together. This is why a lot of people don’t consider distributed ledger technologies (DLTs) to be blockchains. They are not blockchains because they don’t require blocks or chains.
Can the cost of energy be minimized by green energy? What if the mining economy throughout the world was mining with green energy?
Green energy plays a very important role in mining. One of the characteristics of mining is that it can technically occur anywhere, especially now that we have the ability to receive blocks by satellite. You could pretty much mine anywhere in the world where you have an available energy source.
It can come from wind, solar, hydroelectric, dirty coal, or nuclear energy. It can come from any of these sources. Energy is fungible, but the generation of energy and energy production facilities have different properties. For example, solar produces the most amount of energy during the midday sun, and the least amount at night, which will be zero.
Often, the amount of energy produced does not match the amount of energy consumed at those times. In fact, the most energy consumed in a city is usually around 8 or 9 pm in the evening, whereas the production of solar energy is not peaking at that time, but near zero.
How do you reconcile these mismatches?
You have energy being produced for free once the capital costs and capacity are in place. There are low operating costs for solar energy other than maintenance costs; therefore, if you are producing it in the middle of the day and nobody needs it, that energy is wasted.
Mining can take energy that would normally be wasted and use it to produce income, which allows you to very rapidly depreciate the capital expenses of building a solar plant in the first place. You can take a solar plant that would otherwise have depreciation over a period of five or ten years, and instead depreciate it over a period of one or two years.
This leads to an enormous investment in solar.
If you can depreciate the cost of the underlying capital, that makes the deployment of solar much cheaper. Of course, the miners who are using this energy for mining, the cost of that energy can be much lower so the profitability can be much higher.
This is because mining can happen anywhere where there is an energy source, mining will happen in the places where electricity is the cheapest, where the difference between capacity and demand is the greatest, and the distribution networks for other use cases exist poorly or don’t exist at all. All of those factors represent opportunities.
It means that the greatest opportunities for mining are from sources of alternative energy: wind, solar, hydroelectric, geothermal. Those energy sources are often remote and difficult to distribute due to the cost of building distribution networks, and where there is a large mismatch between capacity and demand.
Bitcoin is currently underwriting massive investments in alternative energy around the world.
The computing power used to generate proof-of-work derives value from the fact that it secures the Bitcoin blockchain. However, are there any workable solutions for diverting that computing power for bonus value creation and tackling other problems that require a lot of computing power?
Are applications like protein folding or finding sets of primes workable in this space?
There have been a number of systems that have attempted to use the mining consensus algorithm to do things that people consider “more useful”, such as protein folding or finding primes.
Bitcoin doesn’t do that. Bitcoin uses an algorithm which has no other application other than proving to the rest of the world that miners have committed/expended a certain amount of energy to back the security of the network.
That is essentially the promise each miner makes.
‘Here is a block. I have validated all the transactions according to the consensus rules.’
‘I swear this under penalty of losing the reward for the energy I have expended, and here is the proof-of-work, demonstrating my commitments to validating these transactions and obeying the consensus rules.’
That promise comes in the form of energy consumed for the proof-of-work algorithm. The problem with doing something else in addition, that some other people might consider useful, is that it splits the reward, meaning that the miners have two reasons for which they are mining.
One is to secure the network, and the other is to produce protein folding signatures or large primes.
What happens if the production of primes becomes more valuable than securing Bitcoin? What happens if there is a disruption in that network? For example, we find a new application for specific types of protein folding or prime numbers that make it extremely valuable to produce them instead.
Well, at that point, it would not be worthwhile doing the security of the Bitcoin blockchain.
Using that equipment for doing the protein folding would be more worthwhile, even as a secondary function and Bitcoin security ceases being the primary function, at which point you can no longer trust the promise.
The promise being, ‘I committed this energy to prove that I backed the security of the network by validating transactions.’ If you can’t trust the promise, then Bitcoin’s security fails. The purpose of proof-of-work is to secure Bitcoin.
You can trust the promise because it has no other incentive or reason you would be doing this calculation other than proving your commitment to the security of Bitcoin by validating transactions correctly.
Bitcoin’s mining is useful in backing the security of the network.
The more energy that is committed to that security, the more secure the network is, and the harder it is to attack or compromise the network.
Bitcoin’s proof-of-work algorithm produces useful work for securing a global currency worth almost $15 billion, that ensures it cannot be attacked by any actor, however large they may be, without colluding and investing a lot of money in hardware and energy resources. That becomes very difficult, the more energy and hardware there is backing Bitcoin’s security.
Security is the useful application of Bitcoin’s mining; it is the only application because we want it to always be the dominant use of Bitcoin’s mining, and never become secondary to something else.
Can you explain the most important differences between proof-of-work and proof-of-stake?
Very simply, proof-of-work requires the investment of energy, which is a property outside of the system. Energy is a scarce resource that has costs to produce and distribute, everywhere, in any form.
Proof-of-work incentivizes miners to deposit that energy as fuel for computation; to prove that they have deposited that energy, they produce a proof-of-work in order to claim a reward. Proof-of-stake has a similar function, but instead of requiring validators to deposit energy, it requires a deposit of cryptocurrency, which is within the system, in the form of a stake.
They basically bet an amount of cryptocurrency that is locked up for a certain number of blocks. They put that deposit of cryptocurrency behind their claim, in order to validate the rules.
If everybody also agrees with their validation of the rules, they remain on the majority chain and receive a reward in return, and in proportion to, the stake that they put in.
If they find themselves on the wrong chain, depending on how the proof-of-stake algorithm is implemented, they may lose some/all of their stake, or have their stake locked up for a while, without being able to gain anything from it.
Certain kinds of proof-of-work are really valuable and also useful for ecosystem’s development. Long term investment can be useful in securing the network, building loyalty and creating a profitable system.
This article is a transcription of Andreas Antonopoulos’ explanation on the value of Proof-of-Work.