Fault tolerance is the principle around which the blockchain is built,

and the consensus models offer this capability to the network, which

can differ based on the type of model being used.

The liveliness of the network is ensured by the consensus models by

making it certain that all the correct nodes are participating to

eventually produce value.

The consensus models cannot serve the purpose it has been designed for in

the blockchain as the core of trust protocol on the network is driven through

consensus models. Essential to the blockchain is the presence of a

consensus model; however, varying consensus model implementations can

be different over private or public networks. The blockchain is highly

reliant on a consensus model, but it is not necessarily the same case or the

other way round.

The consensus algorithms which have existed for a long time and have

survived the pre-blockchain era are Proof-of-Work (PoW) and Directed

Acyclic Graph (DAG). Many of these algorithms, when designed initially,

were not created taking blockchain into consideration and were created for

different applications, including military and medical. As the applicability

of these algorithms to blockchain has been very meaningful, the consensus

models have grown in popularity.

The risk of dealing with changes in consensus models across integrated,

interoperable networks, along with this dependency, is also realized by

blockchain.

The networks may face these challenges with changes in consensus models:

The forking of a blockchain, i.e., splitting it into multiple parallel

networks, is caused by a change in the consensus model, which

diminishes the total value. Because of Forking, the user community,

which forms the basis of the network’s success, would also result in

splitting.

For consensus to converge under certain conditions, it may require

more time based on the design of the consensus algorithm or on the

changes being done. These conditions being dynamic would affect

Performance that, results in other nodes turning malicious or a

network partition to have delayed messages that were exchanged