4
$\begingroup$

I've read all other related answers and googled this extensively. My intention and background in asking this question is that I am delving into cryptocurrency and blockchain technology in general. There is considerable misinformation out there regarding Byzantine Fault Tolerance (BFT) and it's relationship to cryptocurrency, that much I can say for certain even with my limited understanding. I will share my own understanding as follows

  1. BFT is not comparable to PoW or PoS consensus algorithms as it is merely a quality that a consensus algorithm can have, not a consensus algorithm in and of itself
  2. Practical BFT (pBFT) systems could be seen as antithetical to the stated goals of Satoshi Nakamoto in that they are centralized and must be centralized

Again, I've read nearly all the information that I'd be capable of following based on my background and have only a cursory understanding of how it is that pBFT functions, about how it is more final than PoW as a consensus algorithm, as PoW is probabilistic.

So, with the above in mind, can you please elaborate on pBFT for delegated BFT (dBFT) in relation to PoW and PoS algorithms?

Thanks to anyone willing to attempt to elucidate what I find to be quite esoteric

Improve this question
$\endgroup$
11
  • 1
    $\begingroup$ Welcome to Cryptography.se. This is clearly off-topic for our site: Do we want bitcoin/cryptocurrency related questions? A cryptocurrency tag? $\endgroup$
    – kelalaka
    Commented Apr 17, 2021 at 17:43
  • 1
    $\begingroup$ @kelalaka actually, the consensus algorithm (which is what is being asked about here I think) of a cryptocurrency might be a cryptographic component, so I wouldn't rule it out a-priori. $\endgroup$
    – SEJPM
    Commented Apr 17, 2021 at 17:48
  • $\begingroup$ it's certainly not off topic, in any case, in the linked exchange it's mentioned that a given individual is free to answer topics pertaining to cryptocurrency if they so desire. BFT stands for byzantine fault tolerant, which describes a systems tolerance to the byzantine generals problem, which finds it's present usage mostly in tandem with cryptocurrency, but BFT is definitely not a topic exclusively related to cryptocurrency. That being said, being asked what BFT stands for shows that probably it's better to take this to a different forum. $\endgroup$
    – josh
    Commented Apr 17, 2021 at 18:58
  • $\begingroup$ Further, clearly implies that cryptocurrency related questions are explicitly disallowed, in the form of some type of mandatory notice, header, etc. Nowhere is it clearly stated. Your usage of the word clearly is superfluous and muddles what would have been an otherwise helpful comment. I read your profile in which you requested that people correct your language. Affirming subjective opinion as objective fact is rarely a good thing. Thanks anyways. $\endgroup$
    – josh
    Commented Apr 17, 2021 at 19:07
  • 3
    $\begingroup$ I’m voting to close this question because this about distributed network's consensus. $\endgroup$
    – kelalaka
    Commented Apr 18, 2021 at 10:40

1 Answer 1

Reset to default
1
$\begingroup$

I'd sum it up as follows. There are two classes of byzantine consensus algorithms (also known as BFT state machine replication, or BFT consensus):

  • Permissioned consensus. Here, the set of $n$ players is fixed and known to everyone ahead of time. The assumption is that at most $1/3$ (or $1/2$ depending on network assumptions) of the players are malicious. (These have been studied since the late 1970s. PBFT is a specific permissioned consensus protocol published in 1999.)
  • Permissionless consensus. Here, the set of players is not known ahead of time, and indeed participants can join and leave the network "at will" over time. The main innovation of Bitcoin/Nakamoto consensus (2008) is to enable some sort of "permissionless consensus".

Permissioned consensus can be used as a building block to build permissionless proof-of-stake consensus. In proof-of-stake, each coin gets one vote in the underlying consensus protocol. Thus, we can run permissioned consensus using each coin as a player. It is permissionless (in some sense) because we don't need to know who owns the actual coin, and because coins can change hands.

So protocols like Ethereum and Algorand essentially run permissioned protocols using the set of current coins as a set of virtual players. Algorand in particular runs a protocol very similar to PBFT. (Their sortitioning technique applies to permissioned consensus as well.)

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.