Skip to main content

Introduction to other consensus algorithms

In this chapter, we will briefly introduce some other public consensus mechanisms.
Remember we defined what a consensus is, in IT, in the Main properties of the first "blockchain" chapter here:

"In IT, a consensus algorithm is a computer program allowing users to reach common agreements on the states of data in a distributed network."

Proof-of-Stake (PoS)

The key idea behind this mechanism is for blocks to be validated by validators that have invested coins in the system. The more coins a validator has in an escrow, the more chances he has to validate a block and earn rewards. This requires that coins are mined early enough so blocks can be validated from the launch.

PoS is a system where validators are not the creators of coins, so that transaction fees provide their only rewards (no coinbase).

Blocks deemed valid and pushed by validators are always verified afterward by the network. If a validator has tried to cheat, he loses all of his funds held in collateral.

Examples: Ethereum 2.0, Peercoin, Blackcoin, NXT...

We'll go into more details about PoS in the Tezos Basics Module as Tezos uses a variation of PoS called Liquid Proof-of-Stake.

Delegated Proof-of-Stake (DPoS)

DPoS has an added Delegation phase. During this phase, decentralized votes are made by the witnesses (witnesses are users that ask other users to transfer voting rights to them, thereby maximising their chance of being elected validators). The witness group has the same function as the parliament in a parliamentary democracy.
Another special committee group decides on the value of parameters such as fees, rewards, or how many witnesses compose the witness group. This committee does not receive any reward and these parameters can only change by vote during another maintenance phase.
We'll go back to DPoS in the "Delegated Proof-of-Stake (DPoS) section" of the "Liquid Proof-of-Stake" chapter in the "Tezos Basics" module.

Examples: Steem, Graphene, BitShares...

Liquid Proof-of-Stake (LPoS)

Tezos has developed LPoS, an evolution of DPoS. Though delegation is optional. As opposed to DPoS, any user can become a validator if she has enough coins. If she doesn't, then she has the choice to delegate. The idea is to dilute even more the mining activity and to increase inclusion.

In LPoS, a validator is called a baker. Any user owning enough coins can become a baker. Suppose a user wants to benefit from baking but doesn't own enough coins or doesn't have enough technical knowledge. In that case, he can delegate his coins to bakers, thereby benefiting from a portion of the transaction fees.

The probability of winning a bake is proportional to the amount invested. The baking time is organized in cycles, and coins are locked during this time.

We'll go into more details about LPoS in the Tezos Basics Module.

Proof-of-Burn (PoB)

Instead of investing computational power and electricity like in PoW, users in PoB must burn some coins to gain a chance of becoming the next validator. Burning coins means sending them to a locking address where they are irretrievable, thereby destroying them.

The more coins a miner burns, the more likely he is to validate a block and win rewards.

This mimics the process of mining where miners have to invest some value, except, instead of electricity, they use the coins themselves.

Examples: Slimcoin

Proof-of-Capacity (PoC)

This consensus first generates large data sets called "Plots". The more plots a miner records, the more lottery tickets she owns. Hence, the more hard-drive space (capacity) she has, the more chances of winning the rewards. This also mimics PoW mining: instead of accumulating hash power, you accumulate hard-drive capacity.

Variants: Proof-of-Storage, Proof-of-Space

Examples: Burstcoin, Storj, Filecoin, BitcoinHD...

Leader-Based Consensus (LBC)

Nodes elect a temporary leader node. This leader is then responsible for validating transactions and for ordering them. They are not necessarily made into blocks (i.e., having more than one transaction).

Examples: BigChainDB

Practical Byzantine Fault Tolerance (PBFT)

In this consensus, each node of the network awaits messages from the others. They are continuously broadcasting messages. Each node receiving a message runs a program on it using its data. Each node's result is then transformed into a new broadcasted message to all other nodes. With time, according to program rules, enough nodes will transmit the same result in the messages. The total sum of these broadcast results determines the consensus.

Examples: Evernym, Chain...

Federated​ ​Byzantine​ ​Agreement​ ​(FBA)

In this distributed network, sets of enough nodes to validate data are established. It means the sufficient number of nodes in a group is determined to best reach an agreement for that group.

A single set of nodes is called a "quorum".

With FBA, a subset of a quorum can aggregate another node, which in turn approves data to form a complete quorum. This subset is called a "quorum slice".

Each node chooses its multiple quorum slices to trust.

The aim of this is to avoid separated divergent quorums, and to ensure that all quorums converge.
Quorum slices must be large enough to have intersections (shared nodes).

To summarize, validator groups with sufficient base size can aggregate single nodes from other groups. In the end, agreed quorums would cover the majority of the network, reaching a consensus.

Examples: Stellar, Ripple.


Each validator randomly selects NN nodes among all the other validators. The more coins a validator has, the more chances it has to be selected (see also PoS above).

Each queried validator responds with its decision.

If most responses differ from the asking node's decision, it will change its answer to that of the others. In case of disagreement between nodes, more nodes are brought into the decision, and more, and more until convergence (thus the name avalanche as it resembles this form of consensus).

Examples: Avalanche

Proof-of-Activity (PoA)

Proof-of-Activity is split into two ordered phases:

  1. PoW phase

    Miners race to find enough zeros in the result of the hash function. But this time, the block they validate is usually empty of transactions. Only the address to receive a reward and the hash and nonce is present.

  2. PoS phase

    A random group of validators is chosen to sign the new block. The more coins a validator owns, the more likely he is to be selected. When all the validators have signed, the block is filled with transactions.

    If there are not enough validators for a valid block signing, another valid block and another group of validators are chosen.

Fees-rewards are then split between miners and validators.

Examples: Decred.

Pros & Cons

#Consensus AlgorithmAdvantagesDisadvantages
0Proof-of-WorkExtreme robustness (MAD + SHA256 + CBC)High Energy consumption
1Proof-of-BurnRobustness similar to PoWStill wastes ressources needlessly; Mining power to money burners
2Proof-of-CapacityLess energy consumption; mimic miningNothing at Stake (no MAD); tends to destroy hard-drives
3Leader-Based ConsensusLess energy consumptionLeader centralized; Sybil Attack[1] weakness
4Practical Byzantine Fault ToleranceVery easy to implementAbsence of privacy; Sybil Attack[1] weakness
5Federated Byzantine AgreementModest computing & financial requirementsQuorum slices centralization; Sybil Attack[1] weakness
6AvalancheFast; highly distributed (lots of nodes)Groups of nodes centralization; Sybil Attack[1] weakness
7Proof-of-StakeLess energy consumptionCoins owners centralization; other problems (see "Liquid Proof-of-Stake" chapter in "Tezos Basics" module)
8Proof-of-ActivityMore distribution on transaction's selection and fees rewardsCombines PoW & PoS disadvantages
9Delegated Proof-of-StakeLess energy consumption; Mitigating PoS centralizationCoins owners & witnesses centralization
10Liquid Proof-of-StakeLess energy consumption; Mitigating even more PoS centralizationStill some coins owners centralization

What have we learned so far?

In this chapter, we briefly described ten consensus algorithms that can be used for public blockchains. There are many others.

In the next chapter, "Smart Contracts", we'll define what they are, what are some of Bitcoin's limitations in that matter, and how Ethereum first proposed to lift those limitations.