Proof of Work is a term frequently mentioned in relation to cryptocurrency and blockchains. This term can be slightly confusing for those who are new to the space, especially since fully understanding PoW requires an understanding of other concepts that are foundational to blockchains. In this article, I will attempt to provide a comprehensive yet easy-to-understand explanation of these foundational concepts and how they relate to PoW.
When considering traditional financial institutions, for example, it is clear as to who maintains a ledger; the centralized institution does. An institution such as a bank uses its own secure systems to appropriately update its internal ledger so that no individual can spend money or sell assets they do not have. Banks are incentivized to accurately maintain this ledger because any inaccuracy could incur financial loss and would lose the trust of depositors.
Similar to financial institutions, the purpose of a blockchain is to maintain an accurate transaction ledger, although, blockchains aim to do this in a publicly verifiable way. More important, the real value proposition of a blockchain is that it can maintain this ledger in a trustless and permissionless way. These qualities further contribute to other desirable aspects of a blockchain, such as censorship resistance and immutability. Features such as these are enabled by transitioning ultimate power over the ledger from a single centralized entity to many distributed individuals.
While a more distributed system is inherently a more decentralized system that exhibits the positive characteristics of decentralization, this kind of system requires a carefully developed framework that allows for the coordination of good actors and severely limits the potential for-and influence of-bad actors. A decentralizing transition of power that puts responsibility in the hands of a large number of distributed individuals requires a multitude of different mechanisms in place for the system to function as intended.
These effective mechanisms are visible in the transaction processes where distributed individuals collectively contribute to maintaining a ledger by accepting the responsibilities of validating and ordering transactions. This process begins with those who wish to transact on the network submitting their transactions to a validator. Validators then check these transactions for correctness and broadcast them to the entire network. Once broadcast to the network, the transactions end up in a pool of pending transactions that await addition to the blockchain. The process of officially adding these transactions to the blockchain involves individuals organizing a specified number of transactions from the pool into a block of transactions that will eventually be added to the chain of previous blocks.
“The process of officially adding transactions to the blockchain involves individuals organizing transactions into a block that will eventually be added to the chain of previous blocks”
An important and subtle element in this description of how transactions are officially added to the blockchain is that the block will eventually be added to the chain. The reason this process must be described in this way is because transactions are not added to the blockchain as soon as they are organized into a block. As a matter of fact, many people will have organized transactions into what they suggest should be the next block-and they do this all at the same time. This means that each person’s suggested block could include exactly the same transactions or include entirely different transactions. The series of transactions included in the suggested next block is at the sole discretion of the individual assembling the block. Although anyone can suggest contents for the next block, they will not know and will not have control over when they are chosen to actually add a new block to the chain. Unless an individual’s block is selected, the transactions they organized into a block will have no effect on the state of the blockchain.
Adding a block to the blockchain is the same as ordering those transactions contained within the block. Choosing whose block to add to the chain, thereby ordering the transactions, requires agreement to be reached through consensus. An important role consensus plays in the security of a network is that it ensures the same person is not always responsible for the ordering of transactions. Several different consensus mechanisms exist to achieve these goals.
One of the more popular consensus mechanisms is Proof of Work. The PoW consensus mechanism functions by requiring individuals to compete for their block to be added to the chain by way of solving cryptographic problems. The first individual to solve a cryptographic problem gets to add their block to the chain and is then rewarded with a block reward. A block reward is paid out in a set amount denominated in the network’s native token and its payment is typically a result of either an inflation of supply, collected transaction fees, or in some cases, a combination of both. This relationship between earning payment and the combination of exerted effort and luck is why competing in PoW consensus is referred to as mining cryptocurrency.
“The PoW consensus mechanism functions by requiring individuals to compete for their block to be added to the chain by way of solving cryptographic problems”
Mining for gold means exerting effort by sifting for gold that might be hidden within silt. When mining for gold, there is an element of luck at play. One person might mine for a long time before they find any gold at all, while another person might mine for a short amount of time and strike gold right away. Although mining for cryptocurrency does not imply any sort of hidden bitcoin, or any other cryptocurrency for that matter, there is still an element of luck at play when it comes to solving cryptographic problems. This luck element of PoW similarly means that one crypto miner could work on solving cryptographic problems for a long duration of time before they finally come up with a solution, while another crypto miner could work on solving these problems for a very short duration of time before coming up with a solution. Another parallel that can be drawn between gold mining and PoW mining is that they are both scalable. An entity that increases their mining power for either commodity-gold or crypto-could statistically expect increased average returns as they scale their operation.
There are numerous characteristics of PoW consensus that can impact a network in a multitude of different ways. A couple of these characteristics are the effect of total network computing power on the difficulty of cryptographic problems and the impact of electricity costs on scalability.
A self-regulating element of the PoW consensus mechanism is that as the total computing power on a network increases, the difficulty of solving cryptographic problems also increases. The total computing power on a network refers to the sum of a mining-specific measurement called hash rate among all of the computers participating in a specific network. The total network hash rate is increased when more people contribute computing resources to the network or when the same people contribute more computing resources to the network. Additionally, as technology advances, a similar number of hardware units produce more computing power. In any of these cases where total network computing power is increased, the difficulty of solving cryptographic problems also increases. This means that a greater amount of computing power is required to solve the same number of cryptographic problems previously solved with less computing power.
Electricity costs have the potential to impact the scalability of PoW mining operations. Since the cost of electricity is a necessary resource for PoW, these costs can quickly add up to a fairly large percentage of mining revenues. This overhead can act as a limiting factor in how large PoW mining operations can scale and where they can thrive.
These characteristics, among others, can be used to justify why PoW consensus might ultimately contribute to either the decentralization or centralization of networks using this consensus mechanism.
The PoW characteristic of increasing difficulty associated with increasing total network computing power implies that participating in PoW consensus requires continued investment in order to maintain the same amount of relevance within a network. This is especially visible on the Bitcoin network as the computers required to participate in Bitcoin PoW are highly specialized. Continual need for hardware investment means that no one can easily reach a point of significant influence over a network and maintain it indefinitely without any additional effort. This could potentially act as a decentralizing mechanism since not only do new entrants need to purchase hardware, but so do established entities who wish to stay relevant in the business.
The requirement for continued investment could also positively contribute to the trustlessness of a network. This is because as hardware investment increases, so does technological advancement, and therefore so does hardware specialization. If entities make large capital investments in the very specialized hardware utilized by mature PoW networks then these entities would be unlikely to act maliciously since doing so would render their investment useless. Malicious behaviors would compromise the network, causing people to lose trust, and therefore any associated assets would become less valuable. Since specialized hardware is algorithm specific, which frequently makes it network-specific, this specialized hardware is useful solely for competing in PoW consensus on a specific network. If the network becomes less valuable, then the rewards generated by the hardware also become less valuable.
A potential decentralizing effect of electricity costs is that the scalability of PoW mining operations is constrained by the amount of this resource that is available in general or at a price that maximizes returns. For example, if there exists a location with a surplus of electricity at a cheap price, someone may equalize the surplus of energy with a PoW mining operation. This operation may be limited to how large it can be scaled in the particular location because of how much energy is available at a low price. Scenarios such as this could potentially lead to a more distributed network of miners.
Some of these same effects could also potentially contribute to the centralization of networks using PoW consensus. If more and more people are competing in PoW consensus, their increased investment would increase total network computing power, which would increase network difficulty. If this happens at the same time as hardware specialization increases and this specialized hardware becomes more expensive, some individuals could get priced out of upgrading their equipment. The inability to add or upgrade hardware would result in the individual’s existing hardware slowly over time generating a smaller and smaller return until it would be unprofitable for them to continue their operation. Individuals in this category would therefore eventually stop competing in PoW consensus and the network could become more concentrated with bigger players.
If these effects don’t contribute to a highly centralized state; they may just potentially contribute to a less decentralized state consisting of many large stakeholders instead of many large and many small stakeholders.