Proof-of-work is a common basis for DLT, or distributed ledger technology, the technology that makes blockchain possible.
Proof-of-work, or PoW, is powerful and versatile enough to enable Bitcoin transactions to be processed in a decentralized, secure, peer-to-peer manner. As with any DLT, a consensus mechanism eliminates the need for a central authority. The proof-of-work algorithm relies on sheer computational power to accomplish this. Other consensus algorithm types, such as leader-based, economy-based, voting-based, and virtual-voting based (hashgraph), use different methods to reach the same goals.
Most blockchain activity, including that of Bitcoin and Ethereum, the two biggest players in that space, rely on proof-of-work. Even so, it has flaws as a consensus mechanism and as a method of block generation.
How does proof-of-work work?
Proof-of-work requires the use of a computer, which you might call a mining rig, to perform the computational work that represents the mining. It burns a lot of electricity, which is expensive. And all of this mining power is applied to solving math problems that are totally useless to humanity, except that they make it possible to form new blocks and confirm transactions on the ledger.
One reason for the use of proof-of-work is to slow things down, so that block generation occurs at a manageable pace. PoW presents problems that only miners (computers) can solve roughly every 10 minutes. Thus, valid blocks come in relatively slowly.
Occasionally two miners add a block at the same time, which is called a fork. That presents a problem that the community as a whole must come to an agreement on: Which of these two do we continue with and which one do we, basically, chop off? It takes a little while to make that decision, and PoW allows for that.
We need to make sure that new blocks aren't added too quickly. The purpose of proof-of-work is to make you run more slowly so that the community will have time to deal with each fork before the next fork happens.
Problems with proof-of-work
Need for more mining power
Unfortunately, that methodical pacing results in a waste of resources and a very high level of energy consumption. Because proof-of-work mining requires so much computing power, it tends to consolidate miners down to the few people who can afford the equipment. It also tends to pull computing resources into locations where electricity is cheap. So we have a consolidation of miners and a geographic consolidation.
There are also security problems with PoW. What if there is a firewall around a good portion of network participants, and someone with malicious intent is running the firewall? That person might decide to turn off the firewall, partitioning the network in two different chains.
That bad actor also could manipulate the firewall by letting some packets through and not others, resulting in different consensus on each side.
Consider the use of proof-of-work for something like a stock market, where the ordering of transactions matters. With PoW, whatever miner puts the transactions in a block can place them in any order they wish. They also could leave some out if they want – perhaps because they were bribed to leave out a rivals' transactions. This results in a lack of fairness in ordering and fairness in access.
There also can be issues with fairness of timestamps, as it’s the miner who puts a timestamp on the block. However, the miner could lie about the timestamp or their clock could be wrong. Because the whole community of network participants does not weigh in on what time the timestamp should be, PoW is vulnerable to issues of fairness.
Room for improvement
Proof-of-work has gained prominence among consensus mechanisms because it does enable the propagation of a single chain. Furthermore, it does so at a pace that gives network participants time to quickly chop off any fork that threatens the integrity of the chain.
However, this form of distributed ledger technology presents room for improvement in terms of fairness, security, expense and efficiency.