Consensus mechanism is the central topic in blockchain space. After Proof of Work originated from Bitcoin, many projects have proposed alternative protocols, for example, Proof of Stake, Delegated Proof of Stake, Proof of Capacity, Proof of Elapsed Time, etc. Unfortunately, no candidate has demonstrated an expected balance among security, decentralization and scalability by the blockchain community. Let’s have a quick review on working consensus mechanisms and see how we can come up with a novel one.
Proof of Work (PoW)
Proof of Work (PoW) are globally well-known and exemplified by Bitcoin and Ethereum. It is widely acknowledged that, by solving mathematical puzzles, PoW has a perfect nature of game theory. However, PoW based blockchains face scalability bottleneck. The average number of transactions per second (TPS) on Bitcoin and Ethereum (Istanbul – the latest upgrade) is between 7 and 15, respectively. That speed is too slow for a mass transaction network. In addition, PoW is not really secure. Small networks (e.g. Verge, Bitcoin-Gold, Ethereum Classic) were vulnerated by 51% attack and lost tens of millions of US dollars while the cost for such an attack is not too expensive. Other concerns on PoW are electric consuming and mining power centralization to giant mining pools.
Proof of Stake (PoS)
The idea of Proof of Stake (PoS) was originally proposed years ago to save energy against PoW. Instead of possessing expensive and powerful machines for mining tasks, ones are required to hold or stake (at least) a certain number of coins. Then the network randomly chooses someone to be a block producer. Unfortunately, new issues arise long-range attack and nothing at stake problem. To avoid them, a modification so-called Delegated Proof of Stake (DPoS) is proposed. In fact, it is the most popular stake-based consensus. Basically, DPoS is a representative model, i.e. stakeholders vote for witnesses (delegates) to function as blockchain full nodes. Normally, DPoS is integrated with practical Byzantine Fault Tolerance (pBFT) algorithm. It can scale to thousands TPS, even much better. However, some authors criticize pseudo-decentrality of such mechanisms due to the vast wealth held by token issuers and rich guys. October 2018, EOS faced a corrupt governance – “mutual voting scandal”.
Other consensus mechanisms
A dozen of other consensus mechanisms has been proposed but not widely appreciated and used, possibly due to their complexity or low verifiability. For instances, Proof of Elapsed Time requires a specific hardware to run time clock, Tendermint is complicated to implement and operate, Proof of Capacity and Proof of Authority is difficult to verify. Many consensus proposals do not have a mainnet, even.
A remarkable path opened by a novel consensus mechanism
Vitalik Buterin, father of Ethereum, raises the Blockchain Trilemma that says about a trade-off among “security, decentralization and scalability”. DPoS based blockchains sacrifice decentralization to achieve high throughput. Algorand – the first pure Proof of Stake blockchain platform, is supposed to solve the Trilemma. It mainnet is launched, but no GitHub repository is publicly available now. Of course, it takes years and millions of transactions (even much more) to verify the Algorand developer’s claim, as Bitcoin and Ethereum have done. An important technical issue that is easily missed by most people that in a PoS system, any random algorithm is dangerous. Basically, no algorithm can guarantee a truly randomness, simply because it is man-made. Therefore, malicious nodes can deploy random algorithms to attack the network. In a PoW based blockchain, finding a required nonce is truly random without any human manipulation possibility. It is the best and the unique truly random source up to now, and possibly no equivalent alternative in the future.
Let’s return to the Blockchain Trilemma, it may be an unsolvable problem or too hard to solve efficiently. Instead, why don’t focus on a balance among three essential requirements: “placing security at first, improving scalability and offering an acceptable decentralization”. What are the fundamentals of a consensus mechanism? That is choosing honest nodes for transaction validation and confirmation. That is implementing game theory to incentivize network participants. Not any one but both those fundamentals gives us a hint to study a consensus that can satisfy the requirements.
Internet is an absolute decentralized space that has been serving millions of websites over decades. DELEGATED PROOF OF REPUTATION (DPoR) considers each blockchain as a mini Internet architecture and each node as a website. Then DPoR has come up with a new approach for consensus mechanism and solving the Blockchain Trilemma.
To be continued.