Blockchain seems like a brilliant solution when multiple anonymous untrustworthy parties need to get together and agree.
But to the point of this article, most situations aren't like that. You have identities, contracts, and trust at least as much as can be enforced with the legal system.
I have also seen a lot of over-application of blockchain to "standard" banking problems. I guess it's real value will depend on whether or not it's successful in "decentralizing all of the things". If that doesn't happen, then it may just be a very cumbersome way of doing something we already knew how to do.
Firstly Proof of Work is incredibly energy inefficient, to extent that electricity consumed to append to public blockchains is comparable to electricity used by nation states.
Secondly POW by its inherent design recentralises trust. When mining evolves to deploying on ASIC and deploying it where electricity is cheap you are recentralising trust to a tiny few who are willing and able to fulfil the ever higher and steadily increasing bar that POW, by design, demands. The trust model at this point has evolved from technical difficulty of fabricating & maintaining long duration forked chains to one where key block creators are too financially invested in success of the chain overall to allow fraud to happen. At this point, when your trust model is essentially too much to loose to cheat, there are scant distinguishing features between a blockchain ledger and a traditional centralised clearing house.
To steal a line from a good book I read recently about blockchains: "The good bits of blockchain are not original, and the original bits of blockchain turn out to be much good."
Blockchains and POW are mostly orthogonal concerns. They've become associated because one implementation chose to use it, and is very well known. Also because POW is an effective way to grow network-effects by distributing profits to those that participate in 'mining'.
My definition of "blockchain" is a distributed ledger in a entirely trustless p2p network. To which PoW is the most widely deployed mechanism in use to secure against the double spend problem. Pow being a mechanism that remains un-superseded inspite of significant efforts to do so. All this making PoW hardly orthogonal.
Of course this definition of a "blockchain" can be loosened. One of two tweaks and suddenly we need to label prosaic everyday items such as a git repos as "blockchains".
At which point it becomes difficult to even have conversation about "blockchains" since our industry has yet to bring to bear a sufficient amount of focus to even define the term with any sort of useful precision.
While your criticisms are valid, you're criticizing just one implementation of "a blockchain", namely chains secured by CPU-bound PoW. There's plenty of other implementations with different trade offs. You could design your PoW algorithm to be more ASIC-resistent, eg https://crypto.stackexchange.com/questions/29890/memory-hard.... You could switch to proof of stake. You could use whatever algorithm XRP is using.
Also, by saying "blockchain is a terrible solution", I assume you mean "there are other better solutions [to the problem of creating a decentralized ledger, or whatever]". What are those better solutions?
"blockchain is a terrible solution" doesn't imply "there exist better solutions". It can mean that it's terrible in absolute terms, not relative -- i.e., that the value of being able to create a decentralized ledger is much less than the cost of blockchain(s).
I agree that bitcoin's use of sha256 nonce seeking is a very easy target for critique. My broader assertion though is that proof of work, by design, will always centralise into a few, one way or another. Other theoretical schemes may be more resistant to centralisation though my instinct is that unless you find some sort of PoW mechanism that strongly anchors to some universal and impossible to simultae physical constraint of the universe we inhabit, POW mechanisms of any economic value will all race towards centralisation shrinking number of signing peers.
Proof of stake at this time is purely theoretical. It lacks strong proof that it solves the problem it sets out to solve. I personally doubt it is a viable solution and my belief it is a dead end avenue of research. Notable there are no large scale implementations yet, even eth rolled back the clock on POS.
Never heard of XRP before. Very briefly scanning, looks like a distributed consensus type protocol, which isn't anything new doesn't attempt to solve the core concern about decentralised trust. Realise am being a bit flippant here, I need to study more which I will, but this is my initial impression.
>> I assume you mean "there are other better solutions [to the problem of creating a decentralized ledger, or whatever]". What are those better solutions?
No I am not saying this at all. Why is it incumbent on me to solve this problem merely because I advance a critique of the current solution? Does my failure to bring an alternative to the table make my critique invalid? I don't think so. I have been puzzling over this problem space and exploring some ideas for alternatives. The problem is hard and honestly I am probably not smart enough to solve it. But the problems are interesting so I'll keep turning it over.
In terms of my thinking right now, I am not even confident that problems solved blockchains are even problems worth solving, that blockchains provide us, as a society, anything of value at all. Too much is given up: notably for example dispute resolution. Look at all the long parade of thefts and errors that have happened on Eth, the legacy to date of the idea of "the code is law" is problematic at best. To date I can only think of one real use case that has made me think blockchain/bitcoin is actually even useful : when US DoJ leaning on Visa/Mastercard to cutoff funding to Wikileaks, bitcoin provided good stopgap technical solution to this grotesque legal excess.
> Never heard of XRP before. Very briefly scanning, looks like a distributed consensus type protocol, which isn't anything new doesn't attempt to solve the core concern about decentralised trust.
To save you some googling, it _does_ try to solve decentralized trust. It goes for a more traditional scheme, a variant of Byzantine agreement.
However, there was a dispute wherein David Mazières (of Kademlia fame) claims that Ripple's design has a flaw, that he fixed in SCP (Stellar Consensus Protocol), by designing FBA (Federated Byzantine Agreement) as described in his paper: https://www.stellar.org/papers/stellar-consensus-protocol.pd... (which features a comparative description with Ripple and Bitcoin).
That's why all the "enterprise blockchain" stuff doesn't use POW. (These are technically "ledgers" not "blockchains" but people don't understand the difference anyway.)
But then you have nothing new. Hash chains are at least as old as a 1981 Lamport paper, and merkle trees (the data structure inside a blockchain block) were patented in 1979.
I'm not talking about hash chains. I'm talking about systems like Corda, Quorum, or Hyperledger that use BFT to provide mostly-trustless consensus without POW.
> Firstly Proof of Work is incredibly energy inefficient, to extent that electricity consumed to append to public blockchains is comparable to electricity used by nation states.
Proof-of-work is designed to be inefficient -- to consume exactly as much power as the bitcoins produced from it are worth. The more efficient PoW is, the less energy it requires to rewrite history, which is exactly what we're trying to avoid.
All of them acknowledge difficulty in computation. It's like trying to solve the drake equation. Unless you know all the terms in the equation accurately the number at the end isn't terribly useful.
Many different computations seem to be in same order of magnitude.
The result of high energy consumption is unsurprising and intuitive, given the design principal of PoW. It's all about who is prepared to do the most work, up until economic limits set by current BTC rate.
Let's do an upper bound, assuming miners are making enough to pay for their electricity (at an extremely low price), but not to pay any of their other costs (cooling, hardware, server operation, profit, etc):
> Blockchain seems like a brilliant solution when multiple anonymous untrustworthy parties need to get together and agree.
No. What you described is the problem of Byzantine generals, for which
efficient solutions are known since '80s. Think again what blockchain is,
because it's not agreement protocol.
Well, Lamport didn't know that the generals are only allowed not to vote when
he was writing the original paper. I read the original publication. Have you?
Given the amount of people that suggest blockchains in a trusted environment it seems perfectly reasonable to reiterate that you don't need a blockchain to have distributed consensus.
I was hoping for something different given the title though.
Except blockchains are incredibly complicated, and as the hacks have shown, programming on the blockchain is a great way to create a bug bounty that also shuts down your company
Definitely, blockchains aren't secure in the way that many Ethereum apps seem to think they are.
But for certain types of lower stakes applications (say a log running on a mesh network of sensors) blockchains can provide a simple way to do something that could otherwise be fairly complicated. Certain blockchain implementations (I like MultiChain) are really easy to use.
There is no such thing as trust anymore at our level of information and 'evil' density. Yes 'evil density' meaning the amount of unemployed outsiders working from a different mission statement than their target. Trust is a relic of small communities and shared goals. As we break down barriers, go global, increase information density - we tempt fate by assuming all new members will function honestly. Or that established members will remain honest
The dynamics of trust in large communities is indeed interesting. I found an interactive explanation about how fragile it can be. http://ncase.me/trust/
Trusting the lawmakers, and elected / appointed agents in general, to do the right thing is another interesting problem. It does not seem completely solved yet, too.
"Law and order" might arguably about trusting such agents; but rule of law is more the opposite. It's about putting those agents in a situation where they tend to enforce rules that they cannot easily manipulate.
You are right that is a hard thing to do, and no society has ever done it perfectly. But western democracies do it surprisingly well, and do so using a playbook of principles that has been understood and written about since the 18th century.
I think the inclusion of blockchain is misleading. I had to wrap my brains around the straw-man first to understand what the article is trying to show-case. It should simply say FaunaDB supports distributed transaction.
Separately, distributed transaction is not the primary motivation for using blockchain. For ledger/financial use cases its the verification and auditing confidence it provides.
it's kind of annoying how people abuse the term "blockchain" nowadays whenever there's something remotely close to distributed systems.
"Distributed Ledger without the Blockchain" is just a distributed database, which has been possible forever. Check out couchdb/pouchdb, etc.
The innovation of Bitcoin's blockchain was that it introduced distributed ledger that didn't require centralized trust. If you take that way, you're just a database no matter how you try to spin it.
This is like saying "We've built orange juice without the orange!" Yeah, you've built bottled water.
Yes. The point of the post is that with global ACID, distributed ledgers among trusted parties are just regular applications. You can’t do this with eventually consistent databases. What I’m describing is a genuine solution our customers are choosing FaunaDB for. For their distributed ledger requirements, a blockchain is overkill.
A distributed database with global consistency, which is a new category. FaunaDB (and similar databases) allows a distributed ledger among trusted parties to be “just a database application.”
How does FaunaDB guarantee that all partitions of a transaction commit or abort? The Calvin paper just says "see here for our distributed pessimistic locks that make this not a problem" but if FaunaDB uses oplocks then that doesn't apply.
In FaunaDB all transaction resolution is deterministic. Once global log order is determined, partitions will wait for the transaction inputs at the correct version and will each independently come up with an identical result.
> Each database has a single partitioned RAFT-replicated log that is used to derive a total order of all transaction effects, and that handles all cross-region network communication.
But if each partition knows about all the reads in a transaction and verifies all the timestamps, that sounds workable. Not that I'm in any way qualified to pass judgment on something like this, but I feel like I have a little bit of understanding. Overall it sounds really cool and the idea of agreeing on inputs instead of outputs is tremendously clever.
Admittedly, the specific mechanics of how the log interacts with data partitions was glossed over a bit in that post.
However, to answer your question, the input log is partitioned: Transactions for each epoch are striped across the log partitions. A data partition will need to read a batch from each log partition in order to be able to process the transactions for any given epoch.
I think this is because people mean different things when they talk about "blockchains".
it is not interesting if we just see it as some glorified chunked linked list using cryptography.
However it IS a novel new piece of technology if we don't define it as that but instead see it for what it actually is. Blockchain is interesting because it uses that chunked linked list data structure to create distributed consensus. So it's unfair to just isolate the data structure on its own and say it's not interesting or novel.
It's only because it used that sort of data structure and applied it to consensus mechanism that it became interesting, and that's what blockchain is. Which is why I think this headline is ridiculous.
> it is not interesting if we just see it as some glorified chunked linked list using cryptography.
See I think the problem here is that people don't realize how many things they use are already blockchains and have been blockchains for years. Git is a blockchain, darcs was a blockchain.
Hell, lots of Wiki implementations are actually secretly blockchains.
What was novel about Satoshi's vision was POW races over the commit rights with enough smarts around how folks would accept blocks. But even then that's a game theoretical way to build the consensus.
> What was novel about Satoshi's vision was POW races over the commit rights with enough smarts around how folks would accept blocks. But even then that's a game theoretical way to build the consensus.
Like i said, people think of different things when they say blockchain so I don't blame you or anyone. I'm just pointing out the phenomenon. Even the most passionate cryptocurrency evangelists seem to have different ideas on what they think "blockchain" is when you watch their speeches.
As you pointed out, the PoW approach to commit rights to a canonical ledger that's distributed is what's novel about this. Some people think this "canonical ledger" concept is not part of what blockchain is, some people do. No one is right or wrong because it's just semantics.
When you compare Git to blockchain you are comparing just the data structure aspect, and I'm trying to point out that some people define blockchain as the distributed ledger network itself, which is where the confusion comes from.
> When you compare Git to blockchain you are comparing just the data structure aspect
Not really, though. While the git client chooses not to verify it continuously, github does. This is a very powerful feature of git, the history of code is difficult to alter once held (especially if you use modern variants of git that use a secure hash).
In a very real sense, git is a blockchain and it's strictly incorrect to suggest that you need an unproven POW game theoretic system designed by an anonymous dude with a fake japanese name.
> No one is right or wrong because it's just semantics.
Why is it that people cheerfully offer this when called out on blockchains but think I'm being improper for calling Etherium an obvious ponzi scheme designed to enable a bunch of quasi-legal and obviously scammy ICOs and then not have to issue refunds even as the ICO folks get fined and outlawed?
> some people define blockchain as the distributed ledger network itself
Some people are just using words in a very wrong way. This is technology, not postmodernist social theory. Even if it was postmodernist, the people using these words in this way are just wrong and not speaking to the academic or even industry consensus. "Blockchain" is "web 2.0". It actually meant something specific for a short period and then marketers got ahold of it and it was basically a non-word.
Even as you pointed out, "git" is a decentralized version tracking system, and "github" is a centralized website that hosts it. In this case it's pretty clear where git starts and ends.
however bitcoin blockchain is a data structure AND a network. And the innovative part is that it manages to stay decentralized (unlike github) while keeping a single canonical ledger. This has nothing to do with postmodernist societal theory, it's a fact.
Also I'm not sure if you actually took a look at the code and ran a node and hacked around with a bitcoin wallet, but from my experience there's a lot more to Bitcoin blockchain than your simplified version of what it is.
> In a very real sense, git is a blockchain
Let me bite and go into technical details and argue with you. Even if we're talking on a purely data structure level, git is NOT a "block" chain. There's a clear reason why blockchain had to use the "block" structure to group transactions together instead of chaining them individually.
Thinking git is a blockchain is as ridiculous as saying anything that uses this type of linked list structure is a blockchain.
Finally, let's not bring in these ludicrous ICO stuff into this argument. I am as much of a skeptic as you about these scams, but just because there are many scams out there, doesn't mean that the original technology is not a work of genius.
If you really think it is not, then I recommend you actually reading the white paper, running a node, and looking into the code. I bet you will realize what you've been skeptical about is nothing more than what the ignorant media have been hyping up, but not the actual technology.
p.s.
Note that i'm the GP of this entire thread which you commented to, so I am as cynical as--probably more than yourself--about people using the term for marketing purposes.
> however bitcoin blockchain is a data structure AND a network. And the innovative part is that it manages to stay decentralized (unlike github) while keeping a single canonical ledger. This has nothing to do with postmodernist societal theory, it's a fact.
But it's not true. Neither the blockchain nor the POW hashes assert consensus. What asserts consensus is technically fear of losing money and the desire to make more money. PoW blockchains only work so long as a majority of computing power in the system is not hostile.
Let me reiterate: THE BLOCKCHAIN DOES NOT ASSERT DISTRIBUTED CONSENSUS AND THE POW HASH JUST DEFINES WHO CAN "PLAY" THE MINER GAME AS A SMALLER SUBSET OF INVESTED PEOPLE.
That is fact. This is why distributed system engineers are very eye-rolley about the ide seea that the blockchain has a Byzantine problem solution. It really doesn't, it just bribes them. As long as those bribes work for a majority of the generals, the system moves one step forward.
This is fundamentally different from other algorithms in the space, and does fundamentally different things.
> git is NOT a "block" chain. There's a clear reason why blockchain had to use the "block" structure to group transactions together instead of chaining them individually.
Squashed commits exist, dude. And quite often Bitcoin's chain has short lived forks. Heck, it's even got a long-lived fork.
The folks with merge rights are your miners in this system. You could also make an analgoue to POS chain systems where the stake is not a majority share of currency but some sort of community mandate or grant.
What's funny to me: cryptocurrency is moving away from Bitcoin's model and towards the POS world rapidly. What's more, people are redefining "stake" in exotic and exciting ways besides Victorian-era money worship models.
> Thinking git is a blockchain is as ridiculous as saying anything that uses this type of linked list structure is a blockchain.
A blockchain is just a Merkle Tree. The "genius" you're so lovingly referring to with Bitcoin is the game theoretic part of Bitcoin. The technical parts are ancient.
It's worth noting that there are lots of attacks that would count as Byzantine faults that neither POW nor POS shared consensus models can handle.
> I recommend you actually reading the white paper, running a node, and looking into the code. I bet you will realize what you've been skeptical about is nothing more than what the ignorant media have been hyping up, but not the actual technology.
I've written them. They're the kind of thing I write over a few glasses of whiskey when the family is away. It was boring. They're not hard unless you're writing them in some crazy makework language like C or Go.
The most exciting parts of cryptocurrency are how we start redefining shared value in a way that offers some of the adaptive benefits of capitalist markets while still offering societally confirmed limits.
The thing that really strikes me about your (and so many other people's) misunderstanding of cryptocurrency is that you have conflated the notion of machine consensus with human consensus. At the speed of the current Bitcoin network, machine consensus is an old and solved problem. A boring problem. What the Bitcoin paper is about is a system that builds human consensus. That's a trickier problem.
> But it's not true. Neither the blockchain nor the POW hashes assert consensus. What asserts consensus is technically fear of losing money and the desire to make more money. PoW blockchains only work so long as a majority of computing power in the system is not hostile.
What exactly is 'not true'? Go back and re-read the sentence you said 'that's not true' to. i was talking about the distinction between a data structure and a network. if you think THAT's not true, then explain why blockchain is not a data structure and a network simultaneously, instead of your irrelevant rant about how Bitcoin is not perfect, because I never said bitcoin was perfect either.
> Let me reiterate: THE BLOCKCHAIN DOES NOT ASSERT DISTRIBUTED CONSENSUS AND THE POW HASH JUST DEFINES WHO CAN "PLAY" THE MINER GAME AS A SMALLER SUBSET OF INVESTED PEOPLE.
You can reiterate as much as you want, and no one is stopping you from believing in your definition of blockchain, but remember that you're just rambling on and on about a topic I never was discussing about.
> Squashed commits exist, dude
Do you really think just because you CAN achieve something in a piece of technology, that makes it same as another piece of technology that is specifically designed to work in that manner? By that standard, iPad, iPhone, laptop, smart watch, they're all same thing. They're made up of Cpu and memory and fundamentally same tech if you open them up.
> A blockchain is just a Merkle Tree. The "genius" you're so lovingly referring to with Bitcoin is the game theoretic part of Bitcoin. The technical parts are ancient.
The game-theoretic part is what people are most excited about, and the only proven way so far to make this type of model work is through this specific data structure AND a network called blockchain. PoS, PoW, doesn't matter, the only proven network is Bitcoin so far. There are other newer approaches like 'tangle' from iota, but you'll probably say that's just nothing more than a graph database haha.
To you blockchain may just be that data structure but to me and many other people, it is also the entire network. my only argument was that people can't agree because people have different views on the terminology.
if you want to talk about whether bitcoin's model scales or not, go to twitter, take sides, and find your friends and fight there, there are plenty of trolls there who will be happy to play with you.
> I've written them. They're the kind of thing I write over a few glasses of whiskey when the family is away. It was boring
no you haven't. just listen to yourself, you sound like a guy just out of coding bootcamp who says 'i can easily build a facebook over a weekend'. Anyone can theorize about a concept and anyone can even write a 'facebook' over a weekend. But that's a child's play, what matters is whether you actually create something in the real world that works on a scale. if you just read just another medium blog post about 'how i built a blockchain in 200 lines of code' and did the same thing and think "there! i've built a blockchain, that was nothing!", then i really can't help you there buddy.
> The thing that really strikes me about your (and so many other people's) misunderstanding of cryptocurrency is that you have conflated the notion of machine consensus with human consensus
All i talked about in this thread is that blockchain is a network AND a data structure. Some people like you think it's just the pure data structure part, some people think it's both. no one is right or wrong about that because there's no fixed definition that says which is which, unlike the internet where they have specific terms like 'tcp', 'ip', 'the internet', etc. You can discuss whether that argument is correct or not, but that's not what you've been doing. you're just ranting about Bitcoin itself and i don't really care about that.
Man this must be what talking to a wall must feel like.
Yes there is, Bitcoin is a p2p network that maintains the blockchain. without the network you don't have bitcoin.
The only difference is you think the blockchain is separate from the network, whereas my point was there are people who also see them as inseparable. my point was there are two different ways of seeing things, including your view. I didn't even say you were wrong, and i didn't even say the other pov was wright. i was just providing another perspective.
Suit yourself, but it's impossible to have civil conversation when all you get in response is condescension.
>> Blockchain does not build distributed consensus, it's a totally different thing: document timestamping service.
> i think you missed the part where i said "I think this is because people mean different things when they talk about "blockchains"."
Oh? So you say that they don't mean a cryptographically linked list of
blocks, like Wikipedia (https://en.wikipedia.org/wiki/Blockchain) describes?
Because that's exactly how timestamping works: unmodifiable list of documents,
from which you can tell which document was issued earlier. Then what else do
they mean?
Most people in IT already know what distributed databases are. They don't want a distributed database. They want blockchain and nothing else. The only reason they want it is because it's something they don't understand, and they naively hope that it will help them somehow.
But they already understand what a distributed db is and know they don't need it.
Which means people will not suddenly start using FaunaDB just because they try to spin it as an alternative to blockchain. people may not yet have full understanding of the implications of the blockchain but they're not idiots.
Actually, you don't have to trust all of them. Some miners can go rogue and it won't matter; everything will still work. The software of blockchains is almost always open source, so you don't have to blindly trust it either; reading the source code and checking it for problems it is always an option.
Consider enterprise B2B transactions - There will always be an old school legal contract in place and hence trust in the existing system. The overhead of the blockchain can be avoided in such cases.
But to the point of this article, most situations aren't like that. You have identities, contracts, and trust at least as much as can be enforced with the legal system.
I have also seen a lot of over-application of blockchain to "standard" banking problems. I guess it's real value will depend on whether or not it's successful in "decentralizing all of the things". If that doesn't happen, then it may just be a very cumbersome way of doing something we already knew how to do.