Lightning Network's effect on miner fees
7 messages, 14 to 15 Oct 2015. Started by s7r.
s7r (2), Paul Sztorc (2), Bryan Bishop, Daniel Newton, Daniel Stadulis
Hello,
I am reading about the Lightning Network and the BIPs which need to be deployed until it can be fully functional. I have to say it's a neat solution to scale and have almost instant transactions in a peer 2 peer, distributed and trustless way. I already knows what the needed BIPs are and what each one does, I am curios about the impact this will have on miner fees.
If transactions happen in a big percent offchain, and they are only broadcasted on the mainchain where funds are moved in or out of the lightning network, this means there will be less transactions on the mainchain -> less fees collected by the miners. What will happen when the block reward will go away? Either the fees for the little amount of onchain transactions will increase to unpractical levels, either the miners will find it not profitable to keep their hardware plugged in to mine, so will leave and the effect will be that the hashing power of the network will decrease. Since the network's hashing power is a security feature (it makes some attacks impossible or insanely expensive) I think it's important to anticipate what will happen in this scenario.
LN transactions are a substitute good for on-chain transactions.
Therefore, demand for on-chain transactions will decrease as a result of LN, meaning that fees will be lower than they would otherwise be.
However, the two are also perfect compliments, as LN transactions cannot take place at all without periodic on-chain transactions.
The demand for *all* Bitcoin transactions (LN and otherwise) is itself a function of innumerable factors, one of which is the question "Which form of money [Bitcoin or not-Bitcoin] do I think my trading partners will be using?". By supporting a higher rate of (higher-quality) Bitcoin transactions, the net result is highly uncertain, but will probably be that LN actually increases trading fees.
On 10/14/2015 6:14 AM, s7r via bitcoin-dev wrote:
Quoted text (22 lines)
> Hello, > > I am reading about the Lightning Network and the BIPs which need to be > deployed until it can be fully functional. I have to say it's a neat > solution to scale and have almost instant transactions in a peer 2 > peer, distributed and trustless way. I already knows what the needed > BIPs are and what each one does, I am curios about the impact this > will have on miner fees. > > If transactions happen in a big percent offchain, and they are only > broadcasted on the mainchain where funds are moved in or out of the > lightning network, this means there will be less transactions on the > mainchain -> less fees collected by the miners. What will happen when > the block reward will go away? Either the fees for the little amount > of onchain transactions will increase to unpractical levels, either > the miners will find it not profitable to keep their hardware plugged > in to mine, so will leave and the effect will be that the hashing > power of the network will decrease. Since the network's hashing power > is a security feature (it makes some attacks impossible or insanely > expensive) I think it's important to anticipate what will happen in > this scenario. > _______________________________________________ > bitcoin-dev mailing list > bitcoin-dev@lists•linuxfoundation•org >
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev
On Wed, Oct 14, 2015 at 10:19 AM, Paul Sztorc via bitcoin-dev <bitcoin-dev@lists•linuxfoundation•org> wrote:
Quoted text (1 line)
> However, the two are also perfect compliments, as LN transactions cannot take place at all without periodic on-chain transactions.
Additionally, lightning network hot wallets are not an ideal place to store large quantities of BTC and users that don't expect to be actively using LN should in general prefer confirmed UTXOs for long-term cold storage. So far the guess that I have seen floating around is that LN usage will at first be restricted to very tiny amounts of BTC in tiny hot wallets, since nobody is particularly interested in running large hot wallets.
- Bryan http://heybryan.org/ 1 512 203 0507
On 10/14/2015 6:19 PM, Paul Sztorc wrote:
Quoted text (16 lines)
> LN transactions are a substitute good for on-chain transactions. > > Therefore, demand for on-chain transactions will decrease as a > result of LN, meaning that fees will be lower than they would > otherwise be. > > However, the two are also perfect compliments, as LN transactions > cannot take place at all without periodic on-chain transactions. > > The demand for *all* Bitcoin transactions (LN and otherwise) is > itself a function of innumerable factors, one of which is the > question "Which form of money [Bitcoin or not-Bitcoin] do I think > my trading partners will be using?". By supporting a higher rate of > (higher-quality) Bitcoin transactions, the net result is highly > uncertain, but will probably be that LN actually increases trading > fees.
Probably yes. But probably no. Having less hashing power is not good, and it's unrelated to scalability and decentralization, it's related to security. Of course we could argue that the hashing power is not super decentralized at this moment but it's unrelated to the topic.
I'd rather have less decentralized big amount of hashing power as opposite to less hashing power.
One theory, very close to yours, is that if Bitcoin transactions demand grows so high that we need the lightning network, there should be plenty of on chain transactions for miners to collect fees from.
I haven't yet seen the incentives of everyone involved in lightning network (payment channel end points, hub operators, miners, etc.) but would it make sense to enforce a % of the fees collected by on payment hubs to be spent as miner fees, regardless if the transactions from that hub go on the main chain or not?
You could make the same argument about changetip, coinbase, bitstamp or any other entity that operates off chain transactions.
1) There is probably no way of blocking them or enforcing fee collection from entities that operate off chain transactions 2) They all have to settle on chain eventually
On Thu, Oct 15, 2015 at 11:37 AM, s7r via bitcoin-dev < bitcoin-dev@lists•linuxfoundation•org> wrote:
Quoted text (54 lines)
> -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA256 > > On 10/14/2015 6:19 PM, Paul Sztorc wrote: > > LN transactions are a substitute good for on-chain transactions. > > > > Therefore, demand for on-chain transactions will decrease as a > > result of LN, meaning that fees will be lower than they would > > otherwise be. > > > > However, the two are also perfect compliments, as LN transactions > > cannot take place at all without periodic on-chain transactions. > > > > The demand for *all* Bitcoin transactions (LN and otherwise) is > > itself a function of innumerable factors, one of which is the > > question "Which form of money [Bitcoin or not-Bitcoin] do I think > > my trading partners will be using?". By supporting a higher rate of > > (higher-quality) Bitcoin transactions, the net result is highly > > uncertain, but will probably be that LN actually increases trading > > fees. > > Probably yes. But probably no. Having less hashing power is not good, > and it's unrelated to scalability and decentralization, it's related > to security. Of course we could argue that the hashing power is not > super decentralized at this moment but it's unrelated to the topic. > > I'd rather have less decentralized big amount of hashing power as > opposite to less hashing power. > > One theory, very close to yours, is that if Bitcoin transactions > demand grows so high that we need the lightning network, there should > be plenty of on chain transactions for miners to collect fees from. > > I haven't yet seen the incentives of everyone involved in lightning > network (payment channel end points, hub operators, miners, etc.) but > would it make sense to enforce a % of the fees collected by on payment > hubs to be spent as miner fees, regardless if the transactions from > that hub go on the main chain or not? > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v2.0.22 (MingW32) > > iQEcBAEBCAAGBQJWHtksAAoJEIN/pSyBJlsR9Y0H+QE/XdW7yauhrNJtp2eIBPg9 > zVUanzR2LT0zAkeF5/Xsx3PFoypALOV7R0YNL29jI3F2XkZA8v24wfNvPi0DETcC > ZOxw4G1erIEjjj51Qz4M7okjQecJxPHOJ+Nz6iNZEDFcZG2b15phCRSQKZwSHP+b > Erw6a4NPs1foieZyk260KSOB8lFs9e8bUJfXd4FfA7l60RA9582K6p05aqVtehFW > ONTe8ULv8F0ba+EzVyTodzzY6ehjD+uc31zL6mDFIbiW+InivFbfi2uDVN1BP/US > m99lLHvDEthnkTokFrbDu81kXdD0lHwIu4O0EMzCnw2E0vWi3sGKd+M0P0sv4WA= > =1qxh > -----END PGP SIGNATURE----- > _______________________________________________ > bitcoin-dev mailing list > bitcoin-dev@lists•linuxfoundation•org > https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev >
On 10/14/2015 6:37 PM, s7r wrote:
Quoted text (22 lines)
> On 10/14/2015 6:19 PM, Paul Sztorc wrote: > > LN transactions are a substitute good for on-chain transactions. > > > Therefore, demand for on-chain transactions will decrease as a > > result of LN, meaning that fees will be lower than they would > > otherwise be. > > > However, the two are also perfect compliments, as LN transactions > > cannot take place at all without periodic on-chain transactions. > > > The demand for *all* Bitcoin transactions (LN and otherwise) is > > itself a function of innumerable factors, one of which is the > > question "Which form of money [Bitcoin or not-Bitcoin] do I think > > my trading partners will be using?". By supporting a higher rate of > > (higher-quality) Bitcoin transactions, the net result is highly > > uncertain, but will probably be that LN actually increases trading > > fees. > > Probably yes. But probably no. Having less hashing power is not good, > and it's unrelated to scalability and decentralization, it's related > to security. Of course we could argue that the hashing power is not > super decentralized at this moment but it's unrelated to the topic.
Who are you talking to? Who said anything about any of this? If you are talking to me, please don't imply that I don't already know these things.
Quoted text (7 lines)
> > I'd rather have less decentralized big amount of hashing power as > opposite to less hashing power. > > One theory, very close to yours, is that if Bitcoin transactions > demand grows so high that we need the lightning network, there should > be plenty of on chain transactions for miners to collect fees from.
For a given fee amount, LN transactions are worse than on-chain transactions. So people would only use LN if they preferred cheaper txns.
Quoted text (6 lines)
> > I haven't yet seen the incentives of everyone involved in lightning > network (payment channel end points, hub operators, miners, etc.) but > would it make sense to enforce a % of the fees collected by on payment > hubs to be spent as miner fees, regardless if the transactions from > that hub go on the main chain or not?
If you want fees to go up, either decrease supply (lower the blocksize limit) or increase demand (a popular Bitcoin). There's no need to do anything roundabout.
Regards, Paul
It makes economic sense to include a transaction on the Lightning Network, iff the the fee to include the transaction on the blockchain is more than than the Time Value of Money of the encumbered funds on the Lightening Nodes amortized across the number of users pushing funds through a LN node.
Large-value transactions are going to hit the blockchain while smaller, predictable, closer-to-median transaction-rates transactions will go into the Lightning Network
Blockchain: Car Purchase, House Purchase, Unexpected Medical Expense Lightning Network: Utility Bill, Groceries, Rent, Mortgage.
Quoted text (4 lines)
> If transactions happen in a big percent offchain, and they are only > broadcasted on the mainchain where funds are moved in or out of the > lightning network, this means there will be less transactions on the > mainchain
This is optimal because the network has minimized the set of costs/externalities to the minimum necessary to conduct a series of transactions
Quoted text (1 line)
> -> less fees collected by the miners.
“It's tough to make predictions, especially about the future.” The effect on fees is going to be hard to predict. 1.) One part depends on user behavior around the dynamics of bid-side demand of fees. I.e. If there is a health ratio of -users who want 1-block- times but are unwilling/unable to bid up the fee of their transactions to push out other 1-block-confirmation transactions (AKA how firm is that fee support presently and under dynamic conditions) to -users who take their transactions off the blockchain to LN
2.) New classes of transactions will be possible that aren't possible today.
3.) What market effect will the financial/technical potential of 'instant' transaction (after a network-joining-intro period) have on Bitcoin's utility/price/adoption?
It would be elucidating if any blockchain data scientists could study the effect of the fee market when high-volume exchanges unexpectedly halted trading.
Quoted text (1 line)
> What will happen when the block reward will go away?
I believe a more specific question to ask is: What will happen when there isn't a convincing economic reason for a large majority of hashing power to be bolstering PoW defense on the main blockchain? Right now we have a pretty good handle on the amount of hashing power that's pointed at extending/defending the 'main' chain but don't have as good intel on how much idle hashing power there is. Idle hashing power becomes more of a threat in market scenarios where chain-extending PoW is scarce (late-game Bitcoin).
My humble prediction is that the necessary number of block confirmation will go up and there were be non negligible mining power idle ready to defend actors' preferred chain. If this is the scenario that plays out, I don't think it'll be very concerning; large-value transaction that will be on the blockchain have more flexible time-settlement tolerance (no one needs their home-buying escrow to settle in <= 1 day) and lower-value transaction that users want/need to be confirmed quickly will be confirmed 'instantly' over Lightning Network or another Bitcoin-anchored protocol.
P.S. I see lots of concern with respect to fee reduction directed at LN while today there are already off-chain databases that remove fee pressure. Like the LN / off-chain databases or not, they will exists.
Daniel Stadulis
On Wed, Oct 14, 2015 at 8:37 AM, Bryan Bishop via bitcoin-dev < bitcoin-dev@lists•linuxfoundation•org> wrote:
Quoted text (21 lines)
> On Wed, Oct 14, 2015 at 10:19 AM, Paul Sztorc via bitcoin-dev > <bitcoin-dev@lists•linuxfoundation•org> wrote: > > However, the two are also perfect compliments, as LN transactions cannot > take place at all without periodic on-chain transactions. > > Additionally, lightning network hot wallets are not an ideal place to > store large quantities of BTC and users that don't expect to be > actively using LN should in general prefer confirmed UTXOs for > long-term cold storage. So far the guess that I have seen floating > around is that LN usage will at first be restricted to very tiny > amounts of BTC in tiny hot wallets, since nobody is particularly > interested in running large hot wallets. > > - Bryan > http://heybryan.org/ > 1 512 203 0507 > _______________________________________________ > bitcoin-dev mailing list > bitcoin-dev@lists•linuxfoundation•org > https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev >