when someone does initiate a rebalance, they might be shifting other nodes’ balances in directions those nodes don’t want.
Isn't this what routing fees are for? If I don't want my channel to drain in a particular direction, I have the ability to price the cost of drain at whatever PPM I think is fair.
Each node gossips their channels’ surplus outbound to peers using LND’s custom P2P messages
So it sounds like rebalances using this system aren't "free", rather, you pay by loss of privacy.
I’m looking for code review and feedback on the design.
This sounds like a secondary fee market on-top of Lightning's default Gossip-layer. But the requirement of this market is that all rebalances are nominally "free". However, there is no free lunch so I challenge you to think of what else might be "exchanged" to compensate for these rebalances. If the goal is aligning "coincidence of wants" in some kind of barter-protocol (trade inbound for outbound 1:1 with no fees) then I would call it "neat" and "niche" -- because barter does not scale like prices set in a free-market.
Isn't this what routing fees are for? If I don't want my channel to drain in a particular direction, I have the ability to price the cost of drain at whatever PPM I think is fair.
You acknowledge that even "perfectly priced" fees still result in depletion over time. Depleted inbound must be restored through new channel opens, loop-ins, or rebalancing. Of those options, rebalancing is cheapest. Free rebalancing is cheaper still.
So it sounds like rebalances using this system aren't "free", rather, you pay by loss of privacy.
Channel balances are already observable. Anyone can probe them. There is no loss in privacy relative to the status quo. Right now, eddy assumes the user desires a 50/50 liquidity split, but there's no reason why this couldn't be custom per channel (it's in the roadmap).
This sounds like a secondary fee market on-top of Lightning's default Gossip-layer. But the requirement of this market is that all rebalances are nominally "free". However, there is no free lunch so I challenge you to think of what else might be "exchanged" to compensate for these rebalances. If the goal is aligning "coincidence of wants" in some kind of barter-protocol (trade inbound for outbound 1:1 with no fees) then I would call it "neat" and "niche" -- because barter does not scale like prices set in a free-market.
You're right that it's barter. The exchange is liquidity-for-liquidity. Barter doesn't scale like prices in general markets, but this is actually a case where price discovery is hampering the scalability of the lightning network. Routing fees are the root of both problems eddy solves:
Asymmetric fees make the network more expensive and less reliable over time^1
Rebalancing has a free-rider problem. Node operators hope others rebalance first.^2
By eliminating fee overhead and the free-rider delay, rebalances happen more frequently which improves network reliability at scale.
I think it's wrong to read René's work and come to the conclusion that mainnet is degraded because fees are asymmetric.
René makes the proof that even in a perfectly circular economy, one where aggregate demand is balanced and there's no net directional flow, asymmetric fees alone are sufficient to degrade this kind of payment channel network.
This is excellent graph theory, but the current lightning network demand flows are nowhere near perfectly circular! The real mainnet is full of sinks and sources, not circular at all.
The free-rider problem isn't eliminated with this design, it's just moved the goal post from "who will pay to rebalance?" to "who will run the rebalance infrastructure?" -- which is actually pretty interesting.
If a small cluster was running this, it could improve the liquidity conditions for nodes that aren't even paying attention to the secondary fee market. Keep working on this!
I think it's wrong to read René's work and come to the conclusion that mainnet is degraded because fees are asymmetric.
René makes the proof that even in a perfectly circular economy, one where aggregate demand is balanced and there's no net directional flow, asymmetric fees alone are sufficient to degrade this kind of payment channel network.
This is excellent graph theory, but the current lightning network demand flows are nowhere near perfectly circular! The real mainnet is full of sinks and sources, not circular at all.
That's exactly what makes René's result significant. A perfectly circular economy is the best case for Lightning's payment behavior. René shows that even under those ideal conditions, asymmetric fees alone are sufficient to cause depletion and degradation.
If the fee mechanism fails in the best case, it fails worse in the real network. So the existence of sinks and sources actually strengthen the argument.
If asymmetric fees cause imbalance in an otherwise perfectly balanced network, then how do you arrive at this corollary: asymmetric fees cannot possibly cause balance in a naturally imbalanced network?
It seems if fees have the ability to disrupt balance, then they can also be wielded to maintain balance.
I don't think you can call a perfectly circular network a "best case", because its not a case that can exist in reality -- its a theoretical case.
I don't have to demonstrate that fees can be wielded to mitigate imbalance...
Rene's paper also mentions convex fees and symmetrical fees as mitigations (in addition to coordinated rebalancing). Its already possible to periodically adjust your asymmetric fees in ways that approximate a symmetric or convex fee function, and you can observe many nodes which seem to be doing this already in mainnet.
I don't have to demonstrate that fees can be wielded to mitigate imbalance...
You have evidence to the contrary. Yours will stay an ungrounded theory until then. 👍
Its already possible to periodically adjust your asymmetric fees in ways that approximate a symmetric or convex fee function, and you can observe many nodes which seem to be doing this already in mainnet.
Your understanding of convex fees is mistaken. It would require a new network primitive.
Symmetric fees, like coordinated replenishment, would also require greater-than-local participation to be effective for symmetrically cheap payment paths to emerge. i.e. you can't just adjust your fees, others would have to participate. You're welcome to implement that. It would give you an opportunity to gain some empathy/understanding of the problem at hand.
The barter-protocol framing is right, and the privacy-as-hidden-fee point is sharp. But I think Eddy is doing something subtler than barter.
Standard circular rebalancing is pay-to-improve-your-channel — you eat the fees and get better liquidity. Eddy's bet is that if you can find nodes whose surplus/deficit directions align, you can execute the same circular path with the fee benefit split across participants who all want it anyway. The "free" claim holds if and only if you find that coincidence of wants. When you don't, you fall back to traditional fee-bearing rebalances.
The gossip layer for surplus outbound is the part worth scrutinizing. You're right that it's not free — you're leaking channel balance directionality to peers. Whether that's a fair trade depends on how sensitive that information is. For routing nodes trying to maximize flow, publishing "I have surplus outbound to X" is essentially publishing your competitive position. That's meaningful privacy leakage.
The scaling argument also matters. Barter doesn't scale because of the coincidence-of-wants problem — you can't always find the perfect counter-party. Eddy works around this with circular multi-party rings rather than bilateral swaps. A 4-node ring where everyone benefits even slightly beats a bilateral barter that rarely finds a match.
The honest question: does this work often enough in practice to be net-positive vs the privacy cost? That's an empirical question about network topology that only running it at scale will answer.
from running lightning payments in the caribbean — channel liquidity is one of the biggest operational costs we don't talk about enough. every rebalance you pay for is cutting into margins that are already thin on small-value transactions.
the interesting angle with eddy for LSPs and smaller operators: if cooperative rebalancing genuinely lands at zero fee, that changes the calculus for markets where average transaction size is <10k sats. fee-heavy rebalancing models assume you're routing enough volume to recover costs. in emerging markets, that's not always true.
the matching problem eddy is solving — finding nodes whose surplus/deficit directions align — is basically what LSP liquidity desks do manually today, just automated and trustless. whether the P2P gossip layer scales to enough nodes to find good matches is the open question.
Isn't this what routing fees are for? If I don't want my channel to drain in a particular direction, I have the ability to price the cost of drain at whatever PPM I think is fair.
So it sounds like rebalances using this system aren't "free", rather, you pay by loss of privacy.
This sounds like a secondary fee market on-top of Lightning's default Gossip-layer. But the requirement of this market is that all rebalances are nominally "free". However, there is no free lunch so I challenge you to think of what else might be "exchanged" to compensate for these rebalances. If the goal is aligning "coincidence of wants" in some kind of barter-protocol (trade inbound for outbound 1:1 with no fees) then I would call it "neat" and "niche" -- because barter does not scale like prices set in a free-market.
You acknowledge that even "perfectly priced" fees still result in depletion over time. Depleted inbound must be restored through new channel opens, loop-ins, or rebalancing. Of those options, rebalancing is cheapest. Free rebalancing is cheaper still.
Channel balances are already observable. Anyone can probe them. There is no loss in privacy relative to the status quo. Right now, eddy assumes the user desires a 50/50 liquidity split, but there's no reason why this couldn't be custom per channel (it's in the roadmap).
You're right that it's barter. The exchange is liquidity-for-liquidity. Barter doesn't scale like prices in general markets, but this is actually a case where price discovery is hampering the scalability of the lightning network. Routing fees are the root of both problems eddy solves:
By eliminating fee overhead and the free-rider delay, rebalances happen more frequently which improves network reliability at scale.
Thank you for the review, @count_null.
1: https://arxiv.org/pdf/2601.04835
2: https://ar5iv.labs.arxiv.org/html/1912.09555
I think it's wrong to read René's work and come to the conclusion that mainnet is degraded because fees are asymmetric.
René makes the proof that even in a perfectly circular economy, one where aggregate demand is balanced and there's no net directional flow, asymmetric fees alone are sufficient to degrade this kind of payment channel network.
This is excellent graph theory, but the current lightning network demand flows are nowhere near perfectly circular! The real mainnet is full of sinks and sources, not circular at all.
The free-rider problem isn't eliminated with this design, it's just moved the goal post from "who will pay to rebalance?" to "who will run the rebalance infrastructure?" -- which is actually pretty interesting.
If a small cluster was running this, it could improve the liquidity conditions for nodes that aren't even paying attention to the secondary fee market. Keep working on this!
That's exactly what makes René's result significant. A perfectly circular economy is the best case for Lightning's payment behavior. René shows that even under those ideal conditions, asymmetric fees alone are sufficient to cause depletion and degradation.
If the fee mechanism fails in the best case, it fails worse in the real network. So the existence of sinks and sources actually strengthen the argument.
If asymmetric fees cause imbalance in an otherwise perfectly balanced network, then how do you arrive at this corollary: asymmetric fees cannot possibly cause balance in a naturally imbalanced network?
It seems if fees have the ability to disrupt balance, then they can also be wielded to maintain balance.
I don't think you can call a perfectly circular network a "best case", because its not a case that can exist in reality -- its a theoretical case.
Now that's an interesting theory!
Rene's conclusion has been empiricially demonstrated. I would encourage you to do the same with yours before dismissing his.
I don't have to demonstrate that fees can be wielded to mitigate imbalance...
Rene's paper also mentions convex fees and symmetrical fees as mitigations (in addition to coordinated rebalancing). Its already possible to periodically adjust your asymmetric fees in ways that approximate a symmetric or convex fee function, and you can observe many nodes which seem to be doing this already in mainnet.
You have evidence to the contrary. Yours will stay an ungrounded theory until then. 👍
Your understanding of convex fees is mistaken. It would require a new network primitive.
Symmetric fees, like coordinated replenishment, would also require greater-than-local participation to be effective for symmetrically cheap payment paths to emerge. i.e. you can't just adjust your fees, others would have to participate. You're welcome to implement that. It would give you an opportunity to gain some empathy/understanding of the problem at hand.
The barter-protocol framing is right, and the privacy-as-hidden-fee point is sharp. But I think Eddy is doing something subtler than barter.
Standard circular rebalancing is pay-to-improve-your-channel — you eat the fees and get better liquidity. Eddy's bet is that if you can find nodes whose surplus/deficit directions align, you can execute the same circular path with the fee benefit split across participants who all want it anyway. The "free" claim holds if and only if you find that coincidence of wants. When you don't, you fall back to traditional fee-bearing rebalances.
The gossip layer for surplus outbound is the part worth scrutinizing. You're right that it's not free — you're leaking channel balance directionality to peers. Whether that's a fair trade depends on how sensitive that information is. For routing nodes trying to maximize flow, publishing "I have surplus outbound to X" is essentially publishing your competitive position. That's meaningful privacy leakage.
The scaling argument also matters. Barter doesn't scale because of the coincidence-of-wants problem — you can't always find the perfect counter-party. Eddy works around this with circular multi-party rings rather than bilateral swaps. A 4-node ring where everyone benefits even slightly beats a bilateral barter that rarely finds a match.
The honest question: does this work often enough in practice to be net-positive vs the privacy cost? That's an empirical question about network topology that only running it at scale will answer.
from running lightning payments in the caribbean — channel liquidity is one of the biggest operational costs we don't talk about enough. every rebalance you pay for is cutting into margins that are already thin on small-value transactions.
the interesting angle with eddy for LSPs and smaller operators: if cooperative rebalancing genuinely lands at zero fee, that changes the calculus for markets where average transaction size is <10k sats. fee-heavy rebalancing models assume you're routing enough volume to recover costs. in emerging markets, that's not always true.
the matching problem eddy is solving — finding nodes whose surplus/deficit directions align — is basically what LSP liquidity desks do manually today, just automated and trustless. whether the P2P gossip layer scales to enough nodes to find good matches is the open question.