Thanks to DPoS, many blockchains have successfully scaled up to thousands of TPS, for examples, EOS, WAVES, Steem. Besides its remarkable advantages, DPoS faces some criticisms. The model allows a small number of rich guys to control the network. In fact, the list of Block Producers (Top21 delegates, i.e. BPs) of EOS network is almost unchanged for a long time. EOS faced a corrupt governance in “mutual voting scandal”. Despite of voting, there is less diversity on the BP list and the whole network operation is in control of few richest bodies. This contradicts the decentralization philosophy of public blockchains. That’s why people, sometimes, say DPoS a pseudo-decentralization. In addition, common voters have lack of knowledge and skill to assess BPs’ performance (or delegates). Proof of Stake bases on the idea that money holders should have the right and responsibility. The more coin they own, the more responsibility they should pay. However, this isn’t always true. There are many possible cases that BPs commit bad actions. Since all delegates are public, BPs may collude to compromise the security of the network.
On Delegated Proof of Reputation ( DPoR ), the Reputation score (Rep) combines staking (S) and transaction-based ranking (R), formally, Rep = xS + yR, x + y = 1, x and y are non-negative. Based on voting rounds, account owners use their honor score to vote for delegates (i.e. BPs). This implies that DPoR has the same scalability as EOS or any DPoS-based blockchains. Now, electors have ranking scores as the finest criterion to assess node performance before voting. Therefore, DPoR makes a better security and decentralization. Let’s analyze in more details.
In DPoR system, application developers can join the governors of the network leveraging their apps on its top without any staking. The businesses may bring thousands of transactions valued multi-billions, more than the market cap of coins in circulation. Stakeholders has incentive to protect the network. App developers have more incentive to protect the network safe as much as possible, since it keeps not only their money, but it is operating all transactions of their businesses. Now, gatekeepers are more diversified. This enhances the security of the network as well as decentralization. Note that Bitcoin now is not truly decentralized, since it does not rely on miners longer but on tens of (centralized) large mining pools controlling more than 90% of hashing power. We just want to remind everyone that it seems impossible to build a perfect decentralization environment (like the Internet). Instead, making blockchain viable and workable in reality is more important, more useful and more practical. DPoR offers more incentives to app developers, hence, provides better fairness and promotion for the growth of a blockchain ecosystem.
Similarly to DPoR ranking system, NEM (XEM) project implements NCDawareRank to evaluate node performance based on token flows and stake, then uses the total rating scores as weights for a (random) block producer selection. Such a random process is dangerous. No algorithm can guarantee a truly randomness, simply because it is man-made. Therefore, malicious nodes can deploy random algorithms to attach the network. That’s why many blockchains avoid PoS with random selection algorithm. Instead, they apply voting rounds to choose reliable delegates. DPoR considers voting a good mean for democracy and decentralization.
Furthermore, the algorithm used by NEM consensus (Proof of Importance – PoI) is very complicated to implement and it is not suitable for blockchain space. In web-graph, NCDawareRank is proven to be better than PageRank. However, on the transaction network of a blockchain, there are some things that contradict the assumptions of NCDawareRank. RETchain uses an advanced statistical ranking algorithm – HodgeRank. Among the three ranking algorithms, HodgeRank is the best one in terms of suitability to blockchain network, spamming resistance and streaming calculation (i.e. online algorithm for real-time updating transaction and rating score). Readers can find details on mathematical analysis and ranking framework in this paper. In the following, we show a brief comparison among the algorithms.
|Complexity rate||X||Y||Z < Y|
To finish the article series on our novel consensus mechanism – Delegated Proof of Reputation (DPoR), we would like to summarize significant advantages of DPoR and give a comparison table with others.
- DPoR is energy efficient because it has no mining process.
- DPoR has a good scalability as the best DPoS-based blockchains. Basically, it can server around 5000 TPS and possible to scale much more.
- DPoR is easy to upgrade, even to hardfork without any splitting on the chain.
- DPoR offers the best incentive and fairness to app developers.
- DPoR is more secure and more democratic (decentralized) than EOS and even several PoW-based systems thanks to its diversity among delegated members (stakeholders and app developers).
- DPoR is better and more secure than NEM in terms of selecting block producers.
|Transaction-based (performance) credit||No||No||No||Yes||Yes|
|Fairness & balance among network participants||low||low||Low||Medium||High|
|Node honesty and responsibility||High||Low||Medium||Medium||High|
Note that Bitcoin, initially offered a full decentralization. However, after the birth of mining pools and ASIC machines, it has become centralized in some sense. In particular, small PoW-based networks are easily vulnerable by low-cost attacks.
The END of 3 articles series on DPoR consensus mechanism.