First, let's discuss whether Bitcoin's proof-of-work-chain can be considered a consensus mechanism.
The article proposing the Bitcoin network concludes with the idea that the “proof-of-work-chain” consists of two parts: the proof-of-work mechanism and the chaining process. This combination can be seen as a consensus mechanism.
However, Bitcoin’s proof-of-work mechanism does not fully satisfy the 'Agreement' condition required for consensus mechanisms.
In Bitcoin’s proof-of-work, temporary forks can occur, especially during the creation of new blocks. Miners might find valid solutions to the cryptographic puzzle simultaneously, resulting in competing valid blocks. This leads to different parts of the network initially supporting different blocks.
While the network eventually converges to a single chain, and the longest-chain rule determines the canonical blockchain, there are brief periods when consensus is not unanimous, and different nodes might temporarily support different blocks. This temporary fork phenomenon is inherent in the probabilistic nature of Bitcoin’s proof-of-work mechanism.
The debate on whether the combination of proof-of-work and the chaining process in Bitcoin constitutes a consensus mechanism remains unsettled. The Bitcoin network mechanism does not fulfill all conditions of a consensus mechanism, which include Termination, Uniform Integrity, Agreement, and Uniform Validity. [citation]
[citation] Kshemkalyani, Ajay D., and Mukesh Singhal. Distributed computing: principles, algorithms, and systems. Cambridge University Press, 2011.
It is important to note that the designers of the Bitcoin network proposed the 'proof-of-work-chain' as a consensus mechanism, implying the combination of proof-of-work with the chaining process, not proof-of-work alone.
Therefore, the early term 'proof-of-work-chain' might be more accurate than 'blockchain,' as it emphasizes the integration of proof-of-work and the chaining process. Separating these two elements affects the entire network, including the effectiveness of the chaining process without proof-of-work.
Without PoW, the connected blocks can be entirely replaced by an altered chain. This is true whether there is a lack of a Sybil or DoS attack prevention mechanism—such as a sufficiently difficult proof-of-work proportional to the total hashing power of the network—or in a permissioned network.
In the absence of PoW, it is possible to replace the entire blockchain with a new valid one, where all blocks are generated based on the hash of previous blocks, thereby creating and replacing the existing blockchain with a new valid chain. Therefore, yes, PoW is an indispensable component in blockchain, and a blockchain system without PoW does not make sense for preventing the alteration of transaction history. This is despite Bitcoin's proof-of-work-chain having its own problems and inefficiencies, such as consuming a large amount of energy while processing only 7 to 10 transactions per second.
Additional information and explanations:
- A Narrated PowerPoint
- A PhD Dissertation
- A Video Presentation
- Permissioned vs. Permissionless Blockchain: How and Why There Is
Only One Right Choice