Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The Limits to Blockchain Scalability

vitalik.ca

Post reply on HN
Print view View original

421–430 of 465 posts

Re: The Limits to Blockchain Scalability

#421

Post by dmitriid »

crazypython wrote: post #408
dmitriid wrote: post #358

Earlier quoted context omitted.

> Yes, the blockchain data structure ensures that if you verify the older transactions, the newer transactions on top of it are verified. That is not enough. Blockchain is proposed for various things like, for example, land registries. They have to be kept around indefinitely long. In many countries financial institutions are required by law to keep financial transactions around for 4 years. Ans so on. But yeah, sure…

> Blockchain is proposed for various things like, for example, land registries. They have to be kept around indefinitely long. In many countries financial institutions are required by law to keep financial transactions around for 4 years. Ans so on. If you have a copy of the latest block and the older verified block headers, you can trustlessly verify older blocks that people give you. > But yeah, sure, go ahead and…

> Archival nodes– nodes that store full history– still exist and are still needed.

Still needed as in "must be there for the chain to function"? Is it possible to have a situation where every node on the chain has truncated its history?

Note also that currently bitcoin needs 300 GB to store data for a fraction of a fraction of transactions needed for the world to run [1]. And crypto enthusiasts people want to move everything onto a blockchain.

[1]

- Total number of transactions is ~640 million: https://www.blockchain.com/charts/n-transactions-total

- Visa alone processes 4.8 billion transactions per quarter: https://s1.q4cdn.com/050606653/files/doc_financials/2020/q3/... which amounts to ~52 million transaction per day

- To handle Visa traffic alone, for that same quarter, Bitcoin blockchain would grow by 9 terabytes. 36 terabytes a year.

And that's for Visa volumes alone. Even though Visa is the biggest one, there's the long tail as well.

"the total number of general purpose credit card transactions in the U.S. was 40.9 billion in 2018" https://www.creditcards.com/credit-card-news/market-share-st...

So, how long before even the archive nodes start truncating data?

Re: The Limits to Blockchain Scalability

#422

Post by thecrash »

dannyw wrote: post #348
seibelj wrote: post #31

Earlier quoted context omitted.

It is always preferred to require less resources. There must be a healthy balance. If Ethereum accomplishes what a lot of us hope, allowing home-hobbyists to run nodes and at least break-even in staking or profit a small amount will be crucial. Relying on goodwill from a small amount of technically-adept-yet-centralized operators is how we get to the Tor situation where everyone knows the government runs most of the…

Genuine question: can crypto economics provide incentives for a more decentralised Tor? I'm talking about being paid crypto for hosting a tor node, and charged crypto for using a tor node. Using privacy technologies like ring signatures of course.

It's an interesting idea, and being worked on by Nym https://nymtech.net/

Re: The Limits to Blockchain Scalability

#423

Post by jude- »

DennisP wrote: post #412
jude- wrote: post #294

Earlier quoted context omitted.

Why is the chain's liveness and resiliency to server failures dependent on where people happen to be keeping their validators? This isn't a problem in Ethereum 1.0.

I mean, if all the miners were on one datacenter and that datacenter went down, so would Ethereum 1.0. That's the same sort of failure you're worrying about for 2.0.

A blockchain's resiliency in part depends on how many full replicas are available -- if at least one replica exists somewhere, the chain can be resuscitated even if all block producers die.

In Ethereum 2.0, nodes don't maintain a full copy of the chainstate -- not even block producers do. Therefore, Ethereum 2.0's resiliency depends on the continued availability of N > 1 partial copies of the chainstate, such that the full chainstate can be reassembled from them. This is strictly less resilient.

Re: The Limits to Blockchain Scalability

#424

Post by pfisch »

legulere wrote: post #113
pfisch wrote: post #102

Earlier quoted context omitted.

Polymarket is a real use case.

Why do you need a blockchain for that? Gambling and Insurance existed both long before blockchains.

Because of the blackmarket nature of the betting, which is illegal in many countries.

Re: The Limits to Blockchain Scalability

#425

Post by wespiser_2018 »

billytetrud wrote: post #359
wespiser_2018 wrote: post #315

Earlier quoted context omitted.

It's so incredibly redundant I think people lack a mental model for just how many messages need to be sent for every node participant to ensure all others are reliable (byzantine fault tolerance). The number of messages is is about n^3, so that's like asking 5 people to go to lunch with 125 emails. https://scholar.harvard.edu/files/mickens/files/thesaddestmo...

Lol, nice link. Why do you think number of messages is n^3? Each message only needs to be sent once to each network participant, that's n messages. Additional messages will be needed to tell their connections which messages they've received, but this can be a single metadata message talking about many primary messages. So if you send 1000 messages through a network of 3000 people, that's not 3000^3*1000 messages, it'…

Because that's how many messages are required to solve for consensus given byzantine failures, at least with relatively simple algorithms like pratical byzantine fault tolerance (p-BFT). The exact bound is O(m*N^2) for pBFT, where m is the number of rounds, at up to 1/3 of N. Blockchains use a different consensus mechanism, but the consensus mechanism is still incredible inefficient compared to something like 2PC which drives Paxos, and can make decisions in O(N) messages like you said. http://www.cs.albany.edu/~maniatty/teaching/os/bft/lectnotes...

Re: The Limits to Blockchain Scalability

#426

Post by BigBubbleButt »

puranjay wrote: post #401
BigBubbleButt wrote: post #257

Earlier quoted context omitted.

> Unless you count making money as not enough of a use case. So then you're agreeing that the only real use case for crypto is speculation, and not something normal people should care about. > By that standard, majority of Wall Street has no use case either, since much of the money is made entirely on speculation. I think most educated people would agree that Wall St is a drain on society the same as targeted ads. Th…

Right now, if I want to invest in US stocks, I have to go through a handful of local brokers. They want a minimum of 10k inr in the trading account. I have to clear kyc and get my account approved. Once approved, they can shut down the account anytime for any reason. And all of these are recent developments - until a couple of years ago, it wasn’t even possible to invest in US stocks. Alternatively, I can buy tokeniz…

I know plenty of people in India who trade US stocks and they don't have to put up that much in assets. You've made so many other factually incorrect statements I don't know if you're lying or ignorant but I know for a fact you're wrong.

Re: The Limits to Blockchain Scalability

#427

Post by dmitriid »

legionlegion wrote: post #420
dmitriid wrote: post #407

Earlier quoted context omitted.

> Blockchain is what makes it freely accessible to anyone from anywhere without requiring a sign up Except... You need to signup because there's literally no way to obtain anything on blockchain if you're not a miner. Well, you could probably sell something for bitcoin, but then oops, you need to sign up somewhere to convert it into actual real money. > I can’t sign up for a US trading account without a certain amoun…

> Well, you could probably sell something for bitcoin, but then oops, you need to sign up somewhere to convert it into actual real money. Not quite, you can cash out BTC without going through any KYC. You can use non-kyc exchanges like hodlhodl, ATMs which just give you cash, or in person p2p trading. > Until you pay for something and that something never arrives. Then you would be very glad to have a central authori…

> ou can cash out BTC without going through any KYC. You can use non-kyc exchanges like hodlhodl, ATMs which just give you cash, or in person p2p trading.

So, all options significantly more involved or less accessible to people than regular cash.

> Lol, you clearly do not live in a country with double digit inflation, government mandated bank limits, oppression

I have. Moldova after the fall of the Soviet Union. Then in Turkey which historically had ~100% inflation for most of its history and then a crash in early 2000s.

I lived through it all.

"You need to set up a wallet for a digital token, somehow buy it for real money, then somehow convert it into real currency" is not an option for the absolute vast majority of people even in the countries with high levels of inflation, bank limits etc.

If anything, cryptocurrencies are a very, very first-world thing.

And, of course, there's also this, much more eloquently put: https://news.ycombinator.com/item?id=27259374

> Cryptocurrency is literally saving lives in some parts of the world

For a very small number of people.

> And you clearly need to workout that empathy muscle

I prefer reality to empathy when it comes to cryptocurrency.

Re: The Limits to Blockchain Scalability

#428

Post by antender »

noobly wrote: post #302
noxer wrote: post #85

This has some really questionable assumptions. Like the part about permanence. "An important property of a blockchain that users really value is permanence. A digital asset stored on a server will stop existing in 10 years when the company goes bankrupt or loses interest in maintaining that ecosystem. An NFT on Ethereum, on the other hand, is forever." This is wrong 2 times. First, there is no general requirement of…

>I dont care how my coins was moved 10 years ago. The only storage that is relevant for me, now, for value transfer is the current undisputed state I don't have anything to add, but this has always struck me as inelegant, especially for Ethereum, which is more of a dApp platform. Can someone point me towards something that can run dApps and maintain current state with distributed consensus _without_ a blockchain (or…

You mean something like Hyperdrive (formerly Dat) + Beaker browser?

Re: The Limits to Blockchain Scalability

#429

Post by mxd3 »

meowkit wrote: post #323
grubles wrote: post #298

Earlier quoted context omitted.

>Vitalik is consistently one of the most interesting people to follow Really? He has consistently argued for on-chain scaling and for people to not validate the blockchain state much like Elon Musk, with his seemingly 101-level understanding of blockchains. This post is a strange 180 from Vitalik's usual "do the opposite of Bitcoin because that is good marketing". Now Vitalik is aware of the importance of running a f…

Provide sources please. This post is full of odd conclusions and non-sequitors.

Vitalik was a big blocks supporter for a long time. I saw this post as a 180 as well.

"Now I personally can see that it’s not axiomatically true that doing nothing is safest, especially in the context of a changing environment (for example I continue to believe that Bitcoin’s failure to raise its blocksize by a significant amount in 2016–17 was a travesty and a great violation of many people’s expectations of the protocol, and one that led to more total losses due to excess txfees than the amount lost in the MtGox hack), but this is the argument that you need to be arguing against."[1]

[1] 12/26/2018 https://medium.com/@VitalikButerin/he-imagines-a-world-in-wh...

Re: The Limits to Blockchain Scalability

#430

Post by billytetrud »

natch wrote: post #317

I feel like the idea that "a million transactions per second ought to be enough for anybody" (not a quote, just a paraphrase echoing Gates for those who missed it) could come back as another storied mistaken estimate at some point. Just taking the human population as a starting point, if we posit that in the future, non-human entities will also be doing transactions, and there may be orders of magnitude more of those…

Ambition isn't magic. The point of the post is that there are physical limits of how much you can scale up without compromising the integrity of the system.
Post reply on HN