Topic: Could the lightning network solve the block size problem? - page 2. (Read 3870 times)

As far as I'm aware, OP_CHECKLOCKTIMEVERIFY is used to ensure that only the most recent tx is valid.

In case of traditional approach, it's not trustless. Imagine that Alice on BitPay transacts 100 BTC to Bob on Coinbase. BitPay deducts Alice's balance by 100 BTC and deducts 100 BTC from Coinbase's account with it. Now Coinbase credits Bob by 100 BTC. Effectively, it means that Coinbase is being owed 100 BTC by BitPay, which is supposed to clear at the end of the day. But if BitPay runs away with money, Coinbase is short 100 BTC!

LN are no more of a learning curve than what Bitcoin was 4 years ago.

In the case of a dispute, then the most resent tx is sent (with a big fee) to the Bitcoin network. LN work by constantly trading would-be-valid Bitcoin transactions, these transactions never need to go on the Bitcoin network unless one of the parties doesn't respond.

In the case that one of the parties doesn't respond, then after a timeout, the latest Bitcoin tx is released on the network, spending both the anchors, and thus closing the channel. -  Everything was agreed upon up-to that point, so very little to Bitcoin was lost (maybe a few cents in fees).

Think about it: it works in the opposite way to how things work atm:

1. You go to the cafe, and buy a coffee with Bitcoin.
2. The cafe gets a Bitcoin TX, and has to wait for a block for it to be confimed.
3. Everything is settled after the tx is in the block chain. (about 10 min or more).

(with lightning):
1. You go to the cafe, and buy a coffee with LN.
2. You update the anchor TX to the cafe, and send the updated TX to the cafe. (instant).
3. No need for settlement since you come in every day.

- See in the general case, nothing ever touches the blockchain. - there is no 'confirmation times'

- Only in the case where there is disagreement dose the 'confirmation time' matter - only to resolve things back to the 'last known good state'.

- It isn't possible to double-spend a LN transaction, since the lock time is longer than the time for the Cafe to release the last know tx.

- if the Cafe goes out of business, then you can get your Bitcoin back, maybe after 6 weeks.  But you only put 20$ at a time, so this doesn't bother you so much.

I still don't really understand how a channel works under dispute, e.g. both party claim different truth. I think that is not possible without a judge-like third party involved

And the traditional approach is also trustless: Bitpay have a 1000 btc account at Coinbase, Coinbase have a 1000 btc account at Bitpay, if Bitpay run away, his account at Coinbase will belong to Coinbase, vice versa

Fair points, I like it. To be honest, I doubt Bitcoin can bear $20 fees even with LN, as LN still depends on the blockchain. I'm also not sure why $20? E.g. if we remove subsidy now, it will require only ( ) $2.5 fee per tx in order to keep the dollar reward per block steady. Of course, assuming other things being equal

IMO the most realistic scenario is that we raise the limit in 2016, and it buys is time to explore all options (be it LN or some kind of fliexible blocksize limit et cetera).

Because LN provides the way of doing so trustlessly -- one cannot steal funds, and in case of uncooperative behavior the channel is resolved on the blockchain.

I'd say both

The lightning network is a great invention.  It supports low-overhead nano-payments with near instant confirmation times that are fully backed by bitcoins.  It has also a few disadvantages:

• If your hub is down, you can't send or receive payments.  You either need to wait until your hub is up again or receive the bitcoins by the pre-signed transactions, which takes several days.
• Hubs may censor your transactions.  Sure, you can always choose another hub or use bitcoin directly.
• You need to establish a payment channel on the blockchain beforehand.  This is a locked coin on the blockchain that contains enough funds to cover your funds that you have at the hub.  The hub must fund it with his own coins to guarantee for the maximum balance that you may receive.  So it is quite capital intensive to run a big hub.  Not only does the hub not get the funds that the customer put there (which is a good thing) but the hub also has to cover for the maximum that its customers may put in their accounts without requiring a new on-chain transaction.
• The security of the lightning network relies on the fact that nobody can spam the blockchain to prevent the confirmation of a fixed fee transaction for a few days.  Do you think this is possible with the current block size?
• If you don't constantly watch the blockchain, your hub may cheat you and steal your coins after a few days.
• There is no implementation, yet.

We need a larger block size anyway.  The lightning network may help to reduce the current growth, but the size of the blocks have to grow.  Also we should at least kick the can to ensure that the current growth can continue short term until we know the lightning network works and is accepted by the community.   And if we keep the block size small to get fees of 20 $ per on-chain transaction (like some small block proponents want), we would basically prevent that people can avoid LN because of censoring or other reasons.

You got it in 1 (well, two really).


BIP101 is the most crude way of "solving" the issue, other than just deleting that one line of code completely, of course.



Sadly, you will also "solve" the decentralisation "problem" at the same time. Please spare us these armchair software engineering decisions, you are either incapable of understanding the issue, or willfully ignorant of the consequences.

The (technical) debate has long since moved on; no one except a bunch of shills and ignorants are trying to offer Hearn and Andresen either as the new team or their hijack-as-a-technical-proposal as the new software.


Give it up, the Bitcoin userbase doesn't want to be constantly harassed by mindless acolytes and sociopathic pushers. XT and BIP101 threads are nearly gone, mostly people are having more productive discussions about totally different topics, scaling or not.

And if you've got a better idea than the lightning network to scale transaction rates, let's hear it (no, not BIP101....)

You were talking about a systematic failure, which this could not cause. Unless everyone on the planet was transacting through Bitpay, this can not cause big problems. Transactions can't be "stopped". Transactions on LN are near instant and get confirmed in a similar matter. If there were XXX numbers of people transacting through Bitpay, and Bitpay gets shut down or something, then they would just need to find an alternative over which they will transact (unless they decided to make direct payment channels to each other).

This is not correct. Even if they do increase it, it will not be via changing a single line of code. Besides, the amount of TPS that LN provides is not even measurable to what a increase in block size would give us.

Note: the first case has 250 bytes per transaction, and the second 500.

In a way it works the same like what the exchangers are doing daily. Deposting bitcoin into an account created and given by the exchanger, you deposit the amount and whatever happens next took place within their own accounting ledger.

FRB is a fact, it happens on every centralized financial institution, simply because most of the funds are not moving: Suppose that a web wallet provider have 1000 customer, and each customer constantly maintain 1 bitcoin buffer in his web wallet, the total amount of coins in web wallet provider would be 1000 bitcoins. Those coins never get less as long as the customer base does not change, then it gives web wallet provider an incentive to lend out those coins to earn some interest

Of course bitcoin lending are extremely risky, so today those institutions just throw those coins into a cold storage. But when the market matures, they can essentially lend out most of those coins without causing too much trouble

Some of the exchanges like btcchina provide others transparent audit of its customer fund database and its cold storage, to prove that they do not do FRB, but the liquidity condition is usually a guarded secret for most of the institutions. And from risk point of view, a 50% reserve ratio can be regarded as extremely safe

I also have this question

Suppose that there are 2000 transactions happened in a payment channel during a day, how could these transactions hit the payment channel and be cleared every 24 hours before the final settlement are written into blockchain?

In legacy financial system, Bitpay and Coinbase would need to keep and update a common list of all their customers' address, and credit/debit each other when a payment between their customer happens. At the end of day, all the to/from payment cancel each other and the net result is written into blockchain

It seems this list of all the customers would need to be maintained outside of blockchain, and the blockchain is just used as a final settlement mechanism

Then the question arises: If institutions still need to use the traditional way of clearing, then why do they use LN at all? They can just open an account at each other like how it is done between banks today

If they are in fact using Fractional reserve then perhaps you do have a point.

Fractional reserve banking is the equivalent of not being solvent.

SO funny how banks are legally able to operate while being insolvent.

How does a "weak" link on the chain have anything to do with systemic failure of Lehman brothers?

The fiat system is already fundamentally flawed given there is more digital fiat than actual paper redeemable fiat you can touch. that has nothing to do with a "weak" link of a network of banks (or transfer channels between banks).

HMM

1. Increase the blocksize by a "decent" amount to accommodate transaction volume by changing ONE LINE OF CODE.

OR

2. Create "the lightning network" to accommodate transaction volume by creating an entirely new system to piggy back on top of bitcoin by adding THOUSANDS OF LINES OF CODE.




Answer = 1  

The risk comes from the centralized organization, not LN in particular. If each LN payment channel is totally independent from others, then the fail of one channel would only affect two institutions. However if many channels are chained together and many channel carries third party payments, then if one of the main channel failed (for example hacking of one of the private key in Bitpay), then all the transactions are stopped due to the chain is broken

Conducting off the blockchain transaction will sure put more pressure on centralization But I'm just wondering if that will involve another ledger to be created since i wonder how are they going to record and capture all these transactions before broadcast it to the blockchain.

Could the lightning network solve the block size problem?

I say it can't by definition. The lightning network is not bitcoin. It is an altcoin. You can't fix bitcoin by using an altcoin. That is no fix, that is simply only not using bitcoin.

And it doesn't matter that the transactions land in the bitcoin network somehow at the end. You did not fix bitcoin by using LN.

I don't like all these claims about bitcoin not being able to fulfill it's role as a currency so that we need crutches to let bitcoin still work somehow. No, bitcoin itself has to be made capable of dealing with this itself.

Just to demonstrate my laziness when it comes to looking at this system, I'm not sure how true the above is. I kept listening to the Lightning presentation in parts while distracted, and I've wondered ever since I heard about it years ago "where's the side chain part?".

I think the answer is that there are no side chains; lightning is a way of aggregating transactions on the main chain. They can take a while to actually hit the main chain (while the channels resolve, which are basically a bunch of chained multi-sig addresses).

So it's on the main chain, but somewhat indirect. And it should enable a load of stuff straight on the main chain too (recurring/micro payments is a significant one for bitcoin, hopefully the protocol will be sufficiently economic for any merchant to open their own)

At some point I'll actually look at lightning properly, just don't ever seem to make time for it

I do like the lighting network, but it should not be considered as an alternative to increasing the block size from where it is now. Since we should bring the benifits of using the Bitcoin blockchain directly to as many people as possible, as far as the technology allows at least.

Sounds like a good solution. There is really no much sense on filling the blockchain with very small transactions.