The problem consensus solves
Suppose a hundred computers each keep a copy of the same ledger. A new transaction arrives. Which computer decides whether it is valid and in what order it happens relative to others? If two people try to spend the same funds at nearly the same moment, the network must settle on exactly one outcome, and every honest participant must end up with the same answer.
With a single central server this is easy: the server decides. Without one, the participants need a procedure for agreeing, even though messages can be delayed, computers can fail, and some participants may be dishonest. A consensus mechanism is that procedure.
Computer scientists have studied the underlying issue for decades. A well-known framing, the Byzantine generals problem, described in the early 1980s, asks how separate parties can coordinate a plan when some of them may send false messages. Blockchain designs are practical answers to versions of this puzzle.
Proof of work
In proof of work, participants called miners compete to add the next block. To win the right, a miner must find a value that, when combined with the block’s data and run through a hash function, produces a result meeting a difficulty target. There is no shortcut; the only way is to try many guesses. Checking a winning answer, however, is quick. The idea behind the fingerprints involved is explained in the article on cryptographic hashes.
- Strengths: It has a long operating record on Bitcoin. Attacking the chain requires acquiring and running a large amount of hardware and electricity.
- Weaknesses: High energy use, specialized hardware, and a tendency for mining to concentrate among large operators. The energy debate is covered in blockchain scalability.
Competing miners sometimes produce two valid blocks at nearly the same time. Nodes follow a rule such as “build on the chain with the most accumulated work,” and the shorter branch is eventually abandoned. This is why waiting for several additional blocks makes a transaction more settled.
Proof of stake
In proof of stake, validators lock up some of their own assets as collateral. The protocol selects validators, often with some randomness weighted by the amount staked, to propose and confirm blocks. Validators who break the rules may lose part of their stake.
- Strengths: Much lower electricity use than proof of work, and penalties can be applied directly to dishonest participants.
- Weaknesses: Critics worry that influence follows wealth, and the designs are more complex and newer, so they have a shorter track record.
Ethereum moved from proof of work to proof of stake in 2022. Designs differ widely among networks, so “proof of stake” is a family of approaches rather than one exact recipe.
Voting-based and permissioned approaches
Some networks use a known group of validators who vote. Classical fault-tolerant protocols, often grouped under names like Byzantine fault tolerance, let a group reach agreement as long as no more than a certain fraction of members misbehave, commonly stated as fewer than one third for many designs.
- Strengths: Fast agreement and no heavy computing race, so they suit organizations that already know their participants.
- Weaknesses: Participation is limited to approved members, and communication between validators grows as the group gets larger, which makes very large open groups hard.
These are the usual choice in private blockchains used inside companies or consortia, where speed and accountability matter more than open membership.
Comparing the main families
| Feature | Proof of work | Proof of stake | Voting among known validators |
|---|---|---|---|
| What you risk | Hardware and electricity spent | Assets locked as collateral | Reputation and membership |
| Who can join | Generally anyone | Generally anyone with assets to stake | Approved members |
| Energy use | High | Low | Low |
| Speed | Slower | Often faster | Often fastest |
| Main concern | Cost and concentration | Wealth concentration, newer designs | Less open, more central control |
The table simplifies. Real networks mix features, and the details differ.
What agreement does not guarantee
Consensus ensures participants agree on the order and validity of transactions according to the rules. It does not guarantee that the data entered is truthful, a theme also raised in the discussion of blockchain myths. It also does not make a network immune to attack. A common worry in open networks is a majority attack, where one party gains control of more than half of the relevant resource, such as mining power or stake, and uses it to reorder or block recent transactions. Large networks make this expensive; small ones can be much more vulnerable.
Why different networks choose differently
There is no universally best mechanism. Designers weigh security, openness, speed, cost and complexity, and the right balance depends on the purpose. A public network meant to resist censorship by anyone may accept slower speeds. A group of companies sharing a ledger may prefer quick, accountable agreement among known parties.
The same pattern appears elsewhere in computing. Distributed databases, which are a topic in database storage, also need ways to keep copies in sync, though they typically assume participants are cooperative rather than adversarial. What makes blockchain consensus distinctive is that it tries to work even when some participants are actively trying to cheat.