Shower thought for Lightning Network Observers, will Lightning fees be cheaper and cheaper as the network grows, and its participants increase? Or will it be higher and higher?
I believe it’s another important question before we assume anything about the Lightning Network.
I have given this some thought.
Things depend on what we call a participant and how the distribution of them looks. Here are the sorts of participants as I see them:
1. Bank nodes
These may actually end up BEING banks, but we will also see Lightning Service Providers in this role. Examples are the Wallet of Satoshi node. Or the IBEX node which I believe is processing the Chivo payments. These nodes are big, and deeply connected. They are the scary "banks" that "Blockstream has replaced the miners with". And they are correctly identified as hubs. These are the ones the big blockers will point to when they are arguing that lightning centralizes control of Bitcoin. In the future I believe actual BANKS will also have nodes like this. Western Union, and Wells fargo better be spinning up BP/LNP nodes RIGHT NOW if they want to remain relevant. Rarely would an individual run a node like this, but Alex Bosworth's node probably classifies. I personally do not see them as evil, or negative unless they take over the entire network, which I think is unlikely. These nodes can see large traffic in both directions.
2. Merchant nodes
These would be big "takers" of liquidity like Bitrefil and other places bitcoin is SPENT. Eventually, companies like Starbucks might end up spinning up their own nodes once they realize this is a better return for them than having a service provider handle it. Hard to say... some businesses will likely run their own, but for a while I imagine many businesses that TAKE bitcoin will use the services of an LSP. These nodes will also likely be BIG and well connected HUBS. Eventually I envision large retailers like grocery stores, and big box places running their own nodes just because of the amount of money moved around as well as thin margins. They also have an incentive to be able to capture the data of their customers directly. "2%off if you are connected to our node!" It could be advantageous for merchants to have a direct connection to customer nodes. These nodes will see more incoming than outgoing traffic. I can even see big merchants becoming LSPs. Your grocery store also becomes your one of your banks so to speak.
3. Non custodial LSP nodes
These are unique. Wallet makers like Breez, Phoenix and Muun makes apps that allow users to run Neutrino nodes on clients like phones. These wallets currently handle all lightning stuff in the background. They open a channel (or more?) on behalf of the customer. Presumably they are using their own nodes which are sort of a subset of the #1 nodes up there. Again, these are well connected HUBS. But the users are not really nodes but dead ends on the network. The business model for these LSPs will be to charge fees on the first hop.
4. Hobbiest/Pro routing nodes
This is what I run. This is a generally smaller node with strategic targeted connections to the network. They can be run for several reasons. A computer hobbyist with a rPi will run them. Self-sovereign bitcoinners that want to "be their own bank" and preserve their privacy might run them. Idealists who want to keep bitcoin as decentralized as possible might run them. I think they are quite important. These are the nodes that keep the network from ONLY being run on commercial grade nodes and captured. We might be motivated by making a little profit for routing, or like me motivated to provide low/zero cost connectivity between 1,2 and 3 as well as the ability for user to user connections. I personally connect to several of the above nodes as well as rings of peers to keep the network as distributed as possible and drive fees down. It is hard to imagine that this will ever be more than a minority percentage of the network, in connectivity, and liquidity. But i do think enough of us will have a palpable effect on the overall network. We are the X degrees of separation that will route around the big powerful hubs, thereby forcing them to be better network citizens.
5. Dead end users
This could be the majority of users. These are users who use a custodial wallet (arguably not really lightning users at all, but customers of lightning businesses), or a non-custodial wallet that does not route at all. So these are the dead end spokes on the hubs in 1, and 3 (and maybe 4?). Since they do not route payments they do not really have an effect on the network other than providing velocity and liquidity for the routing nodes to handle. These "nodes" will handle way more outbound transactions, but will also process "Venmo" type user/user payments as well as refunds.
To your point, I think #4 is the lynchpin for how fees end up working. If there are enough of us providing low fee connectivity I think the cost for payments could be kept very low or even free for the archetypal "coffee transaction". Small transactions from customers to businesses, or from user to user could find lightning pathways that cost very little and do not even have to touch some of the hubs. I would guess we see the hubs being involved in transactions that are big enough that it becomes hard to find pathways between the idealists. But my node has enough inbound and outbound liquidity that I could route payments approaching $10kUSD at today's prices.
Anyway.. it is VERY interesting in my opinion... and we are still in the infant stages of it. It will be fun to watch!
But do you believe node operators would run their nodes altruistically by taking fee rates lower and lower, or do you believe they will eventually look for incentives? Because opening/funding channels require capital, which is limited, resources to maintain hardware costs which is also limited, and technical knowledge/labor maintenance.
