What is evil

in evil •  4 years ago 

Gè zhǒng gòngshì suànfǎ, jiédiǎn jīběn dūhuì jìnxíng shǔ jù jiǎnyàn gōngnéng, yīn'ér shùjù wèizào zuò'è jīběn bù kěnéng. Nà zhēnzhèng de zuò'è shì shénme? Zhǔdòng zuò'è: Jiédiǎn zhǔdòng shēngchǎn liǎng gè qū kuài, gěi yībùfèn jiédiǎn fāsòng xīnqū kuài A, lìngwài yībùfèn jiédiǎn fāsòng qū kuài A1, zhèyàng xìtǒng jiù chǎnshēng fēn chāle. Bèidòng zuò'è: Kuài chuánbò yīnwèi wǎngluò yán shí huòzhě zhōngduàn, huì dǎozhì yībùfèn jiédiǎn quēshī gāi kuài de xìnxī, jìn'ér huì chǎnshēng fēn chā, ránhòu hòuxù qū kuài shēngchǎn zhě fāxiàn fēn chā hòu, rúguǒ zūnxún zhǐ xuǎnzé yīgè fēnzhī shēngchǎn qū kuài jiù huì ràng fēn chā shōuliǎn (néng shōuliǎn zuìhòu jiù néng dáchéng gòngshì), dànshì rúguǒ jìxù bǎochí bìng jiāzhòng zhège fēn chā, zuìhòu jiù huì dǎozhì zhěnggè liàn biàn chéng yī kē sēnlín, zuìhòu jiédiǎn jiù yǒngyuǎn méi fǎ dáchéng gòngshì, zhè shì bèidòng zuò'è. Bǐrú POS de noting at stake gōngjí, dāng xìtǒng chūxiàn fēn chā shí, shēngchǎn zhě jīyú zìjǐ de lìyì huì zài liǎng gè fēn chā qū kuài shàng gèzì shēngchéng yīgè qū kuài (zhèyàng bùguǎn nǎge fēn chā huòshèng, zìjǐ dōu néng huòdé shōuyì), zhè qíshí jiànjiē zhùzhǎng hé jiāzhòngle fēn chā yīn'ér yào zǔzhǐ zhèxiē zuò'è xíngwéi, yǒusān zhǒng fāngfǎ: 1) Jīzhì bǎozhèng, ràng shēngchǎn zhě méiyǒu shíjiān lái shēngchéng 2 gè qū kuài (POW) 2) lìyì bǎozhèng, shēngchéng liǎng gè qū kuài zuìhòu kěndìng méi yòng (BFT suànfǎ, dàjiā zuìhòu zhǐ huì rèn kè yīgè kuài), shēngchǎn zhě huòqǔ bùliǎo shénme hǎochù, zìrán méi bìyào shēngchǎn liǎng gè qū kuài 3) shìhòu chéngfá, bǐrú DPOS, zhè zhǒng fāngfǎ qíshí méi shà yòng, yīn wéi shìhòu fāxiàn zuò'è jiédiǎn hòu, xìtǒng méi fǎ huísù qū kuài, zhǐ néng chéngfá zuò'è jiédiǎn, dànshì zhège zuò'è sǔnshī shì bù quèdìng de (qiánmiàn tí dàoguò), qǐ bù dào chéngfá zuòyòng, qiě dǐyā wù móshì shì yī zhǒng xiànzhì, bù lìyú shēngtài fāzhǎn
Show more
589 / 5000
Translation results
For various consensus algorithms, nodes basically perform data inspection functions, so data forgery is basically impossible. What is the real evil?

The main action is evil: the node actively produces two blocks, sends a new block A to some nodes, and another part sends block A1, so that the system will fork.

Be evil by action: Block propagation will cause some nodes to miss the block information due to network delay or interruption, which will result in a fork. Then, after the subsequent block producer discovers the fork, if only one branch is followed to produce the block Let the bifurcation converge (convergence can finally reach a consensus), but if you continue to maintain and increase the bifurcation, the entire chain will eventually become a forest, and finally the nodes will never reach a consensus, which is evil. For example, the noting at stake attack of POS, when the system is forked, the producer will generate a block on each of the two forked blocks based on their own interests (so that no matter which of the forks wins, they can get profits), This actually indirectly encourages and aggravates the fork

Therefore, there are three ways to stop these evil behaviors:

1) The mechanism guarantees that the producer does not have time to generate 2 blocks (POW)

2) Benefit guarantee, it is definitely useless to generate two blocks in the end (BFT algorithm, everyone will only recognize one fast), the producer can not get any benefits, naturally there is no need to produce two blocks

3) Punishment after the event, such as DPOS, this method is actually useless, because after the malicious node is discovered, the system cannot backtrack the block and can only punish the malicious node, but the loss of the malicious node is uncertain (mentioned above) , No punishment, and the collateral model is a restriction, not conducive to ecological development
Authors get paid when people like you upvote their post.
If you enjoyed what you read here, create your account today and start earning FREE STEEM!