PoW Or DPoS? Neither! Points of Prof. Cao On Current Consensus Algorithms | An Interview By Mars Finance Part II
Huining Cao:
Now let’s take a look at some common census in the following two categories:
PBFT, POW and its derivatives.
PBFT demands interactions, so it does not fit into public blockchains in large scale. Since we focus on public blockchains, we shall not go into details. Briefly speaking, Neo uses dBFT which is a derivative of PBFT, so NEO as a public blockchain will possess a bad performance and will not get far. Though still not satisfying enough, POW is currently the safest choice. It is quite simple: a miner puts a transaction into a block, verifies the matching of the public key and the private key, adds a nonce, and finds the corresponding hash value. He will be able to get the block done when the hash is lower than a given difficulty, to broadcast it to all nodes, and to be rewarded. Now the average time for finding a block is designed to be 10 min. If the computing power increases, the difficulty will alter automatically so that the average time remains.
It is 10 min for a reason. The system will be too slow with a longer time and have too many branches with a shorter one. Also, it has a potential to be centralized when malicious nodes keep mining and take control over the system.
Host:
What are the disadvantages of centralized tokens? Are they useful for MLM or pyramid schemes?
Huining Cao:
As for MLM, MLM always possible when there are tokens. There is no way to get over it, and so we need proper regulation. By being distributed, every node can have a complete ledger. By branches I mean two miners complete their blocks at the same time, and it will be confusing for others if there are too many branches.
Currently, we choose the longest and with highest difficulty chain in cases of branches.
Host:
Do current tokens obey current economical laws?
Huining Cao:
There are not many new laws in economy. Let’s move on to the disadvantages of POW.
POW is smartly designed, but it has a weakness too — monopoly is possible through computing power.
Three traps in blockchain world: First is the monopoly by large mines; second is financial shuffle by centralized exchanges, deceiving capitalists with high coin prices, manipulating the market, and doing those inside jobs; third is that some unscrupulous coin issuers use pseudo-concept of blockchain to gather money and then abscond.
First: in the smart design by Satoshi Nakamoto, computing power is centralized by Kaomao with ASIC. Kaomao is a student from University of Science and Technology of China School for the Gifted Young, who turned to design mining machines for calculating hash values in middle of his Phd at Yale. His machine is a thousand times faster than others and soon occupies a large proportion of the market, forming a competition in computing power and letting the energy consumption go as high as that of Denmark.
Kaomao is missing now, possibly kidnapped and killed. What he has left results in the huge influence blockchain has on the environment and restricts blockchain in a rather small scale.
POW’s weaknesses include low transaction process speed, high energy consumption, and small scale, so it needs improvement before being applied in a large scale.
The first derivative of POW is POS, invented by my friend SUNNY King and used in Peercoin in 2011. What is different about POS? POS links the difficulty of mining to the coins held by the miner and mortgage those coins to a public account. The higher fraction your coins posses in total mortgages, the higher the hash is, the lower the difficulty is, and thus the more likely you make it in mining. Sometimes the hash difficulty also depends on the coin age — the longer you have held the coins, the easier you mine. Then some nodes are more likely to mine the blocks, and thereby the transaction speed is improved and the energy used is lowered.
But other problems come to the stage: those who have coins already keep mining and get more and more like a Matthew effect. Then there is a centralizing tendency, because whoever holds more coins is more likely to get the next one and more likely to be attacked by hackers. ETH is in its transition from POW to POS. But it is still said to be slow after POS’ birth in 2011. So don’t be fickle-minded. If you are already getting away from decentralization, why not just be a centralized bitch? Here then come POA and DPOS.
In 2013, Daniel Larimer claims DPOS to be another type of POA. While with POA a miner chooses the next miner to do the mining, with DPOS a committee is elected by the community and the members mine in turns. BTS has 11 super nodes, and EOS has 21. DPOS gains high speed and low energy consumption by allowing one miner each time, but the problems is that DPOS loses the essence of blockchain by being too centralized. That means the 21 super nodes can easily corrupt or be under DDOS attack by hackers. And it cannot be applied, for a central organization will suffice if everyone has trust in others.
DPOS is unreliable at all without sufficient trust as prerequisite. Do you think Xiao Li is more reliable than Xiaochuan Zhou? So you see, DPOS is a big idiot with a bunch of smaller idiots mastebating together. It has no future to me.
Second: the malicious centralized exchanges will be challenged by the decentralized exchanges. When blockchain performs better, we can randomize the centralized pairing and decentralized settlement, enhance transaction speed and user experience, and make the decentralized exchange a competitive platform, an exchange for every single customer.
Third: malicious coin issuers do regulatory arbitrage. Normally, companies get listed under strong supervision. Yet today anyone may issue coins and reshuffle. Regulation is necessitated; the usage of ICO funding has to be transparent on blockchains.
Now we will talk about applications. If we have an algorithm with low energy consumption, high transaction speed, and large scale, where can we apply blockchain technology?
Super node is a genius marketing, but it can’t last long. Compared with centralization, blockchain technology has its innate weakness, so it should be applied only in certain scenarios with proper regulation.
Blockchain can be applied where centralized organizations are incapable or unwilling to do the job, where cost or moral risk is high and information asymmetry is huge. For example, in some South African countries, like Venezuela and Zimbabwe, the government is corrupted and in lack of self-discipline, so its policy enforcement is even less dependable than blockchains. Blockchains can also be applied in payments there. Assuming that technical problems are solved, we may use blockchain technology in small loans and international loans.
In each of its application, blockchain should be compared with centralized organizations to see whether its reliability and justness can outrun its low efficiency and high energy consumption. In addition, public chains need to cooperate well with application chains, which takes delicate bottom design to avoid traffic jam. The design of public chains has to be altruistic — it has to focus on the community and allow the community or the public chains to be independent of the designer. That’s the reason why BAT can never make great public chains — they can never be altruistic.
Let me talk about Usechain next please. Our aim is to embrace regulation and run the system legally. So, we want KYC/AML and decentralized applications with high speed, low cost, large scale, and less dependence on computing power. Our mirror identity mapping separates identity verification and identity on chain via proof of zero-knowledge. Then, each individual has and only has one main account and limitlessly many subsidiary accounts. The government can audit the accounts to prevent tax evasion, money laundering, drug trafficking, etc. While the accounts remain anonymous to the public, in the case of losing your private key, the several groups of people responsible for identity verification and identity logging in may help restore your account based on a voting by the community.
With a one-to-one mapping, we will introduce the concept of RPOW. RPOW lets everyone generate a pseudo-random number and a hash afterward based on a distinct verifiable random function (VRF) and the public address of his main account. Then we compare the hash with a hash generated by a public nonce and examine the last 32 digits (in binary). If they are the same, the mining probability is maximized; if the last 31 digits are the same, the mining probability is lowered by 16 times, and so on. In this way, the max is 16 to the 32 power times larger than the min, so the miners with low probability will give up. As a result, only 10 to 20 nodes are mining at each time, the energy consumption goes low, and the speed goes up. Compared with EOS, we have 10 to 20 each time, while DPOS has only 1. Though DPOS is faster, our 10 to 20 are random, or say they are re-selected by God each time.
Therefore, we can set up an equal, decentralized system.
We have solved the impossible trinity by identity mapping and randomization. In addition, we may do slicing with identity mapping and isolate subsidiary chains from the main chain in a tree structure, thereby bringing blockchain to a larger scale. We also improve smart contracts by preventing bugs with formalized proofs. Nodes are allowed to enhance the speed by doing parallel verification to the smart contracts. In a word, we surrender our anonymity to the government and make everything legal.
We give the financial application of blockchain what it wants — high speed, low consumption, large scale — with Mirror Identity Protocol and Randomized Proof of Work. We add Tokenized Proof of Stake in the subsidiary chains. Our goal is to build a blockchain for practical uses instead of public stunt. Blockchain is for the community and for everyone, so we shall set up and remain our credits. Every public chain belongs to the community, so Usechain belongs to the community as well.
Host:
Jun Du, what do you think about the exchange that can “randomize the centralized pairing and decentralized settlement, enhance transaction speed and user experience” mentioned by Professor Cao? Is it hard to realize today?
Jun Du:
It seems pretty hard for today’s technology, and it may take some time.
Huining Cao:
Today’s technology cannot support a decentralized exchange with great user experience. Thus, breakthroughs in two ways are required: First, the decentralized exchange must have a breakthrough in speed, consumption, and scale. Second, it must still take centralized pairing, but the settlement can be decentralized. Centralized pairing can be split and dealt with by several groups.
One more thing I forgot to mention previously. If the mining frequency of an account is x times higher than average, then its mining probability for next time will be lowered by 2^(2^x-1) times. When x=5, it is 2³¹ which is approximately over 2 billion. So the mining machine has a quite small influence.
Reach us at:
Website: http://www.usechain.net/
Weibo: https://weibo.com/usechain
Twitter: https://twitter.com/usechain
Facebook: https://www.facebook.com/UsechainFoundation/
Telegram: https://t.me/usechaingroup
Medium: https://medium.com/usechain/
Bitcointalk: https://bitcointalk.org/index.php?topic=3432061.0
Reddit: https://www.reddit.com/user/Usechain
