Consensus Mechanisms
Every blockchain faces the same problem: thousands of independent nodes must agree on which transactions happened and in what order, without trusting each other or a coordinator. Consensus mechanisms solve this.
The solution in every case: make disagreement expensive. In proof-of-work, forking requires reproducing the accumulated hash work. In proof-of-stake, misbehavior is punished by destroying the validator’s bonded capital. The details determine the tradeoffs: energy use, decentralization, finality speed, and throughput.
Subtopics
- Proof-of-Work, computational puzzles as the cost of participation: SHA-256, Ethash, RandomX, mining pools, and 51% attacks
- Proof-of-Stake, bonded capital as the cost of participation: Ethereum Casper, validator duties, slashing, and alternative PoS designs
- Delegated Proof-of-Stake, token holders elect a small set of block producers who take turns proposing blocks
- Byzantine Fault Tolerance, explicit multi-round voting for deterministic single-block finality: PBFT, Tendermint, HotStuff
- Proof-of-Authority, only whitelisted, identified validators can sign blocks: Clique, Aura, enterprise chains
- Proof-of-Space, pre-committed disk storage replaces hash computation: Chia’s plotting, farming, and Proof-of-Time
- Other Consensus Mechanisms, Proof-of-History and broader comparisons
The key tradeoffs
| Mechanism | Energy | Decentralization | Finality | Throughput |
|---|---|---|---|---|
| Proof-of-Work: nodes compete to solve a cryptographic hash puzzle; the winner adds the next block and earns the block reward. Difficulty adjusts so blocks arrive at a steady rate regardless of total hash power. | High | High (anyone can mine) | Probabilistic (~1 hr) | Low (~7-15 TPS) |
| Proof-of-Stake: validators lock tokens as collateral to earn the right to propose and attest to blocks. Misbehavior (double-signing) is punished by automatically destroying a portion of the stake (slashing). | Low | Medium (stake required) | Fast (~12 min on Ethereum) | Medium (~15 TPS + L2) |
| Delegated Proof-of-Stake: token holders vote to elect a small fixed set of block producers (typically 21-27). Elected producers take turns proposing blocks in a round-robin schedule. | Very low | Low (elected delegates) | Fast | High |
| Byzantine Fault Tolerance: a known validator set runs a multi-round voting protocol. A block is committed only when 2/3+ of validators explicitly pre-vote and pre-commit to it. One bad round restarts with a new proposer. | Very low | Medium | Instant (1 block) | Medium |
| Proof-of-Authority: only pre-approved, identified validators can sign blocks. Validators take turns in a round-robin; a block is valid if signed by an authorized address. Used in enterprise chains and testnets. | Minimal | Very low (known validators) | Instant | High |
| Proof-of-Space: nodes pre-compute large lookup tables (plots) stored on disk. To win a block, a farmer finds the best matching value in their plots for a network-broadcast challenge. Storage capacity replaces hash power. | Low | Medium | Probabilistic | Low |
How to read the table
Energy reflects the cost of participation. PoW requires continuous computation. PoS requires capital lockup but minimal electricity. BFT and PoA require almost none.
Decentralization reflects how easy it is to become a validator. Anyone with hardware can mine PoW. PoS requires capital. DPoS requires winning an election. PoA requires being on a whitelist.
Finality reflects when a block can be considered irreversible. Probabilistic finality means earlier blocks are safer but never absolutely final. Deterministic finality means a committed block cannot be reverted.
Throughput is the base layer. All mechanisms can be scaled with Layer 2 solutions; the base layer number reflects on-chain capacity.
Related topics
- Bitcoin, the original PoW chain
- Ethereum, the largest PoS chain
- Staking, the economics of PoS participation
- Distributed Cryptography, the cryptographic building blocks behind PoS signing