Hacker News Reader

Comment by aforwardslash | Hacker News Reader

I'd argue that PoS/PoW have nothing to do with the blockchain concept, and instead are a way of building virtual currencies on top of it. The easiest way (for me) of thinking of a blockchain is a distributed merkle tree (or a patricia trie, if you're into ethereum or whatever), with a consensus mechanism. That's it. What you put there is your own responsability - so if you decide to create a virtual currency and either reward work or reward possession, its up to you - but that's on top of the "blockchain" concept. And I've actually came across some pretty valid use cases - one good example is that blockchain could be used to enhance resiliency on systems like certificate transparency - that are already a centralized, versioned sparse merkle tree.

View on HN

Totally disagree. The thing that was really novel about blockchain was its distributed consensus mechanism, even in the face of bad actors. Without that, just having, say, a cryptographically signed list of transactions really isn't new.

The novel thing was a consensus mechanism that is itself permissionless in the sense that anyone can contribute to it. There is no procedure for becoming whitelisted to mine. You just mine.

If there were such a permission scheme, such as in Facebook's now likely defunct cryptocurrency idea, you would not need proof of work. You'd just have a list of authorized signers of new blocks and if someone does something bad you kick them off the list. This is basically how the web PKI system works with browser certificate whitelists.

Lots of cryptocurrency critics over-argue their case by denying that this was a novel innovation, but yes it was novel. It has big downsides but it works. Personally my big issue with crypto other than the energy waste is that the cryptocurrency ecosystem itself is toxic and full of scams and other trash.

"You'd just have a list of authorized signers of new blocks and if someone does something bad you kick them off the list."

Are you saying that the incentives provided for successfully mining a block in crypto's isn't necessary to their security? I can't think of a single expert that agrees, or an example in the space of that working. Block reward is typically thought of as a tradeoff between security and inflation.

Yeah.

A distributed consensus mechanism really isn't new, either. In fact, some mechanisms are as old as merkle trees. And even within the realm of consensus mechanisms, using BFT (https://en.wikipedia.org/wiki/Byzantine_fault) to achieve consensus is basically an 80's technology with a new spin on it. BFT isn't magical and brings its own set of problems - the most common is the relative simplicity of "majority" attacks.

A lot of the research on Distributed Ledger Technologies (blockchain is an example of a data structure that is a DLT) is in playing with game theory, voting schemes and incentives.

It's a really interesting space to research for economists or political scientists. It's also a lever to turn government, social, financial and other systems into software.

All that to say, yes, PoW and PoS are not necessarily tied to blockchains or DLTs.

You can use PoW as a sort of captcha, for example. If it costs a little bit of CPU to submit a request to a website it's not a big deal for a human. However, if you're running a bot-net suddenly it starts to hurt!

botnet operators rarely pay for the electricity or hardware of their botnet. why would CPU usage hurt them ?

Depending on what the Botnet is trying to achieve, CPU can be “expensive” in a time sense. If you are doing a credential stuffing attack with something like a COMB dump, then every cycle counts because you have a ton of creds to check.

A lot of them are using budget VPS services, or even Amazon/Google/Microsoft cloud services.

True, it won't solve the problem of malware botnets, but PoW or PoBandwidth schemes can make it too expensive to run botnets on cloud services so the only place left to go is illegal botnets.

Isn’t this just bcrypt?

Almost. PoW allows the entity validating the work to determine just how much work it thinks is appropriate at the moment of authorization, whereas the number of hashing rounds for bcrypt is determined at the time the secret is stored.

So if I lock up a secret with bcrypt today and require 100k rounds of hashing, it will still require 100k rounds of hashing to unlock for any brute force attempt of the ciphertext for the rest of time. A server (or blockchain network) could evaluate a PoW submission 50 years from now and determine that difficulty should be raised to 100T rounds instead of 100k rounds.

> You can use PoW as a sort of captcha, for example. If it costs a little bit of CPU to submit a request to a website it's not a big deal for a human. However, if you're running a bot-net suddenly it starts to hurt!

That's exactly what HashCash was conceived to do (to fight email spam). It was a neat predecessor to Bitcoin.

PoS/PoW are decentralized concensus protocols; they could have used alternate data structures but practically something like a log-structured database with cryptographic hashed for integrity is the only thing that makes sense to reach consensus on, so cryptographic-tree structures that existed for decades were the logical choice and the concepts got entangled in people's minds. Cryptocurrency doesn't require decentralized consensus (multi-witness would work just about as well). Merkle trees don't need consensus; IPFS works just fine with content-addressed naming.

Blockchains are unique in their design to be self-sustaining decentralized cryptocurrencies; the self-sustaining part is what, in my mind, distinguishes them from the other technologies which rely on external incentives to keep them running.

Is PoS actually decentralized?

It depends on which PoS algorithm you look at, but some of them are generally considered to be as "decentralized" as PoW in terms of the security properties they give to their network under adversarial conditions.

Is PoW actually decentralized? It's more like multi-witness because miners aren't able to see the block they're working on and validate it before submitting back to the pool operator and they can only trust the pool to keep paying out rewards through non-blockchain means.

In ideal worlds as designed for PoW and PoS everyone has roughly an equal stake and equal incentives. Reality differs.

I think there is a common conflation of distributed and decentralized concepts; Many (most?) distributed systems require a consensus mechanism to operate (as they require coordination/state), but many decentralized(most?) systems don't, as each node operates independently and without a transitive state. As such, I have some difficulty in seeing PoS/PoW as "decentralized" (in a decentralized system, multiple different "end states" may coexist without being a problem), but rather more of a reputation model as a way to solve BFT. And BFT consensus mechanisms are - in my opinion - only one of several possible mechanisms, because (again, imo) "blockchain" and "public blockchain" make different assumptions on the underlying implementation.

The consensus mechanism depends on the existence of the currency as a reward. It financially incentivizes mining and choosing of the "correct" fork.

"this specific consensus mechanism". There are plenty of them, that don't require monopoly money.

PoW/PoS are just consensus mechanisms. Otherwise I agree that a blockchain is a Merkle tree with a (distributed) consensus mechanism. I'd tend to say that only distributed consensus mechanisms count as blockchains (and thus "permissioned" blockchains aren't true blockchains), but that's nitpicking.

A basic question I have is: what stops someone from re-writing history? With Bitcoin, it's the fact that it would be computationally hard. By the time you were able to "mine" more, the network has moved on. If we're talking about a private blockchain (like I see some companies pushing), what's stopping a malicious actor from simply recomputing on new data?

IBM talks about private blockchains, but it kinda seems like without spending a lot of computing power to make it infeasible to re-compute, you're relying on trusting that people won't. If I control all the machines in the private blockchain and the cost to compute the hashes is low, presumably I could replace nodes that I want to remove or alter and then just recompute.

For example, let's say that I'm trying to create an immutable ledger of emails sent for compliance purposes and I have 10,000 emails. Later, I want to hide 5 emails sent and alter 5 email's contents. I replace the remove/alter the nodes and then just re-compute the chain storing the new version of the chain on the servers I control. I then tell regulators that those emails never existed and we have the immutable block chain to prove it!

Even if a blockchain is public, what is protecting against someone disputing history? If it were computationally cheap, I could come up with a longer chain than the current Bitcoin chain with certain transactions altered to favor me. The reason I can't do that is because the proof-of-work makes it infeasible for me to re-compute a block before the network has already put a new block on the chain (presuming I control less than 51% of the compute power in the network). To recompute 10 blocks, I'd need to control an overwhelming majority of the compute power in the network.

But if the data is important and we're not giving people a big reward for mining the chain, it seems like it would be vulnerable to alteration. Sure, you could say "I saw the chain with 1,000 links and then I saw the chain with 1,100 links, but this new chain disagrees with some of the original 1,000 I had. Someone is doing something sketchy, but is it the first person trying to rewrite history and not being able to mine fast enough or is the second person rewriting history and has enough compute power to become the longest chain?" Again, with Bitcoin this isn't such a problem because PoW makes it infeasible to re-compute that stuff. However, if we're talking about a blockchain that isn't secured by PoW (or PoS), what is preventing that?

If I want to create a public ledger of all my shipments with blockchain and there are 10-50 shipments per day at various unknown times, it seems insecure. I put up a piece of shipment data to be added, and then it gets added to the chain. That's the good scenario. Now I want to rewrite history so I go back and re-compute some stuff and then for the next 20 shipments I privately compute the signatures before broadcasting to the network so that I'll have a chain 20 longer than everyone else which will replace the old chain (and with it the data I'm looking to alter).

I could just be missing a piece that makes this comment void, but it's always seemed like Bitcoin's resistance to history-rewriting is the PoW. Without that, I could simply re-compute the history, create a longer chain, and replace the current chain. Yes?

I think you may be missing the proposed use case of private blockchains. They aren't for storing data within a single organization on trusted computers controlled by that organization. They are for sharing data across a limited consortium of organizations who don't trust the integrity of each others' computing platforms.

Also, it's not clear to me whether you're asking about PoS security too, but if so, the reason one person can't go back and rewrite a bunch of it is because the right to create blocks is spread out randomly across the entire active set of stakers. So if an attacker only controls a small fraction of the total stake, then their signature would only be valid in building a small fraction of the chain they would need.

In your scenario, the other members in the chain would trivially detect your modification because they have a record of what you said before. Then they decide to kick you out of the club and/or sue you for breach on contract.

> And I've actually came across some pretty valid use cases

Pics or didn't happen.

This is not a constructive response (in terms of HN commenting guidelines), especially given that the parent provided concrete examples. If you disagree with their examples, or if you think that the cons outweigh the pros, a better approach next time would be stating that and making your counterargument.

The effort we invest to encourage thoughtful dialogue is paid back in the long run by keeping the quality of discourse high(er) and making less work for dang.

Please don't post unsubstantive comments or flamebait. Perhaps you don't owe the blockchain better, but you owe this community better if you're participating in it.

https://news.ycombinator.com/newsguidelines.html

Reply on news.ycombinator.com