The two most common types of consensus mechanisms used in decentralised networks are proof of work and proof of stake, or more commonly referred to as PoW and PoS.
The PoW concept was originally published by Cynthia Dwork and Moni Naor in 1993, but later coined “Proof of Work” in a 1999 document published by Markus Jakobsson and Ari Juels. The PoW protocol was intended to deter cyber-attacks, primarily distributed denial-of-service attacks (DDoS), which exhaust the resources of a computer system by sending countless fake requests. The creator of Bitcoin, Satoshi Nakamoto, applied the PoW protocol to his/her cryptocurrency, which revolutionised the traditional transaction mechanism (Beller, 2018). In fact, the PoW protocol is the foundational idea behind Nakamoto’s 2008 Bitcoin white paper because it allows for trustless and distributed consensus.
Every new transaction that is published to the Bitcoin network is grouped together in a “block” of recent transactions. The block is then compared to the most recently published block in order to verify no unauthorised transactions were inserted. With a newly verified block added to the network nearly every 10 minutes, this continuous process creates a sequence popularly known as a “blockchain”.
Trustless and distributed consensus relies on the idea of a public ledger, in which all users have the ability to check and agree on any block of transactions. Traditionally, a person making a transaction relies on a third-party to validate the transaction (e.g. Paypal, Visa, American Express). These third-party providers maintain their own private registers that store transaction histories and the balances of each participating account. By providing all users on the decentralised network a copy of the public ledger (the blockchain), PoW makes third-party validation obsolete. The protocol creates a trustless and distributed consensus without the need for third-party providers.
Although the PoW protocol and mining mechanism applies to a variety of cryptocurrencies, Bitcoin is the original, largest, and remains the best example.
So how does the decentralised network encourage users to verify and continuously update the ledger as new blocks are produced every 10 minutes? The concept sounds like a public service, which could lead to a free rider problem, so it is necessary for the network to reward users who take on the task of verifying newly added transactions. The decentralised network encourages users to assist in keeping the transaction record operational and updated by awarding newly minted Bitcoins to the user(s) who solve a mathematical puzzle that is based on the pre-existing contents of the block (Bohme, 2015).
These mathematical puzzles can only be solved by computationally intensive methods that include a random component. Faster computing is more likely to solve a given puzzle and will subsequently solve a greater. Once a puzzle is solved, the user(s) publish a block which contains the proof of work, a solution that references the completed block. Other users are then able to verify the solution and start working on a new block containing outstanding transactions.
The whole process is referred to as mining, because the user is essentially working for newly minted Bitcoins. The process recursively ensures that the historical ordering on all blocks is agreed upon by the entire network. Miners effectively vote on the correct record of transactions, and in that way verify the record. Sources recommend only considering a Bitcoin transaction complete after at least 6 network confirmations of the block. This ensures the transaction is recorded as a permanent part of the distributed ledger. Waiting for further confirmations will entail a longer delay (currently this is approximately 1 hour).
The computerised PoW calculations are extremely power-intensive, continuously consuming more than 173 megawatts of electricity. This is roughly 20% of the annual average nuclear power plant output (World Nuclear Association, 2015), or USD 178m per year at average US residential electricity rates.
Computational costs have grown sharply and may rise further because Bitcoin automatically adjusts puzzle difficulty so that the time interval between two blocks remains roughly 10 minutes. This means that as more computing power joins the Bitcoin system, the puzzles automatically become more difficult, increasing computing and electricity requirements. Taylor (2013) has found that spikes in the Bitcoin-USD exchange rate have been followed by increases in computational difficulty.
PoS is an alternative way of verifying and validating a block of transactions that uses significantly less computational power than its PoW counterpart. The PoS system picks a validator, known as a forger (the PoW equivalent of a miner), by the amount of stake (coins) the forger has and the length of time the stake has been made available. For example, if a user has 1,000 DASH coins in a wallet, there will also be an age associated to how long the coins have been in the user’s possession. The 1,000 DASH coins is considered the stake. The amount is basically a security deposit, which means that the validator is holding a significant stake in the DASH coin and with long standing ownership is considered to be more committed to validating the block. This allows the construction of a trusted and distributed network with loyal validators. Instead of receiving a newly minted cryptocurrency for their service, the validator earns either part or all of the transaction fee. This process is called forging instead of mining because the validator is essentially forging a block to the chain without acquiring a newly minted coin/token.
Although the PoS system has been accused of being more of a conceptual rather than working mechanism, it can still claim several benefits over other methods such as PoW. In either case, it is worth understanding some of the PoS system nuances in more detail.
One of the most significant practical benefits to the PoS system is that it eliminates the need for expensive hardware. The process is supposed to be so energy efficient that a normal laptop or computer running the blockchain’s validator client will be sufficient as long as the laptop or computer is online. PoS also creates more loyal validators, meaning the longer a validator holds a high stake for, the more likely they’ll be chosen for forging and earn transaction fees. The efficiency of the PoS system subsequently leads to faster validations, which is a considerable benefit given the explosive rate at which consumers are adding transaction records to blockchains.
In the PoS system, each validator owns a certain amount of stake in the network that they bond. Bonding stake means that a user deposits some money into the network, and effectively uses it as collateral to vouch for a block. Therefore, under a PoS mechanism the users will trust the chain with the highest collateral, while in PoW networks the users know a chain is valid because there is an abundance of computational proofs behind it. That said, one significant issue on the horizon for PoS is the potential for a small group of users to own a majority of tokens/coins and become the permanent validation pool, effectively dismantling the idea of a decentralized system.
The PoS consensus mechanism is still being developed, with too many differences between implementations to name here. Time will tell which has the most potential to evolve into a solid and robust mechanism which can be scaled up for mass adoption.