Blockchain Crash Course: Consensus Mechanisms
Cryptocurrencies have rammed into this world without too much notice. However, a key word has stuck with us as we are riding along the emotional rollercoaster that these digital assets are making us experience: decentralization.
Decentralization is achieved when decisions are made and finalized in a distributed network through an appropriate consensus mechanism. What that means is that all participants that generate transactions can check that these very transactions have been successfully included in the immutable ledger worldwide. And that any participant can check this fact by simply using a blockchain "scanner".
How all the above falls into place to enable a trust-less distributed network, where there is fundamentally no need to look up to some central authority or a proxy of it for the network to be fully functional, is to understand the role that consensus mechanisms play in the grand scheme of things.
Let’s give it a simple definition first.
A consensus mechanism is a set of rules that govern how the nodes maintain and to some extent process the transactions being pushed to the nodes.
Nodes are any device able to confirm the authenticity and singularity of the transactions that it is being sent to (miner in POW or validator in POS). A node is not limited to the usual hardware represented by the computers or supercomputers, but basically by any device that has computational capabilities, of course the contribution from less computationally capable devices will be less obvious.
Now, for simplicity let us compare distributed systems to a group of people. If the group of people is left to govern itself with no apparent rules, it is easy for other groups of people to attack it and establish themselves as the leading group of people. We have intentionally used the rudimentary word “group of people” because if we were to use the word “society” instead we would be assuming that this group of people is following some sort of social etiquette in order for the single individual to continue being part of the group.
Well then, what social standards are to groups of people is exactly what consensus mechanisms are to distributed systems.
Social standards: groups of people = Consensus Mechanisms: Distributed systems
In fact, once the social standard is being set, the groups are self-aware, and all future developments are shaped by these standards. Consensus mechanisms work similarly. When a consensus algorithm is set, the distributed network will be self-aware and govern itself to find more or less fairly who is entitled to create the next block which will contain all the transactions to be validated.
Consensus mechanisms are self-policing protocols without which any blockchain-based network will be vulnerable to various types of attacks.
As to the number of consensus mechanisms in existence, we are unable to accurately enumerate it, as we believe that similarly to social standards, there can be only as many as people are comfortable with. Within the top 20 digital assets according to coinmarketcap.com we can easily observe that any consensus mechanism has equal chance of succeeding (which by the way, has nothing to do with how environmentally friendly an algorithm is).
Here we will attempt to summarize in the clearest way possible the differences between consensus mechanisms:
Proof of Work (POW)
Miners (nodes) need to solve a mathematically intensive puzzle with the following features:
- Asymmetric (difficult to answer but easy to check)
- No shortcut to solution exists but only 1 strategy is available: trial-and-error. If a miner would like to increase its chances, then he or she can increase the available computational power by purchasing hardware (in exchange of value, in most cases fiat)
- Difficulty is dynamically adjusted just like in GMAT standardized tests to account for how fast or slow is the response to the difficulty of the question
- Environmentally damaging for the "work" being done to keep the network safe and running
Proof of Stake (POS)
While in POW whoever is more geared up wins, in POS the efforts are more virtualized and practical. In other words, the miners become validators simply by staking (locking up) their tokens for a certain period of time.
In POS, a randomized process will select who gets to generate the next block and therefore receive the reward for the “effort” (the effort is more virtualized, as staking the tokens have a cost opportunity equal to the outperformance of the market with respect to the token held). In that, it is more fair because any token holder is allowed to participate and take an interest in the improvement of the network.
The rewards received for generating the next block greatly depends on how the reward system has been designed. Particularly, rules can be set for example on the to date duration of the staking or simply on the amount staked.
A more extreme case of POS if you will, is represented by the DPOS (Delegated Proof of Stake), which really works like the US government democratically-elected government representatives of the American people. Congressmen and senators are the representatives in government of the wider electoral base, i.e. anyone having the right to vote (token holders). However, a big and perhaps controversial difference in the context of voting, is that the weight of each vote changes according to the token holders' net worth. But we could argue that 2 senators per state is equally unfair, and the weight is unbalanced when it comes to final voting on government bills and major political decisions. As the delegates in the distributed network are looking to receive more votes with the objective of being elected, they also organize campaigns and lead efforts by spending their budget in improving the network’s value, reputation and usability.
Delegated Byzantine Fault Tolerance (dBFT)
The Delegated Byzantine Fault Tolerance (dBFT) is a consensus mechanism which overcomes the Byzantine Generals Problem. What essentially this algorithm does is that it attempts to solve a situation between nodes (delegates) where untrustworthiness is presumed. Similarly, to the dPOS, in dBFT we have the users who will vote the delegates that they deem more honest. Not an issue if one of the delegates chosen turned out to be dishonest since it can be easily replaced at the next voting round.
The fundamental difference to dPOS is that among the delegates a Speaker will be chosen, which will act as a master delegate in charge of proposing a new block. The block’s validity will be vetted by the delegate pool and for it be considered valid, at least 2/3 of the delegates must support it, otherwise it will be scrapped together with the Speaker.
In summary, the distributed network governed by dBFT is comprised of nodes, delegates (who can approve the blocks), and a speaker (who proposes the next block). Various scenarios illustrate how the dBFT protocol is robust enough to protect against malicious actors (both delegates and Speakers) within the network. The basic assumption is that only a minority of delegates will act dishonestly. Therefore, if a Speaker proposes a faulty block, the delegates will rectify it by voting.
About DEXTF Protocol
DEXTF’s on-chain asset management protocol will allow the creation, management and trading of digital-native funds units, connecting investors, arbitrageurs and portfolio managers in a way that has never been done before. Working to bring the biggest crypto investment funds on DEXTF protocol and to eliminate the need of a third party custody service.
Follow us on: