Oddbean new post about login | logout 4 years ago I set up my primary lightning node hosted server. I funded it with several coins, figuring better to have too much liquidity than not enough. I opened dozens of channels under the same assumption. Over the years I've barely touched it other than to update release versions. Today it only has half a dozen channels due to peers disappearing or closing channels, but the value of my liquidity is higher than ever, so I don't see a need to open more channels. Which begs the question: is this centralization? 🧐
It doesn't matter in a free market. The issue with centralization is when others aren't allowed to compete. I don't think there's a magic number of providers that separates good from bad (or 'centralized'). I think we need to reframe the term to reflect the actual problem. You're providing a service that people want. Anyone who can do it better is free to.
The protocol is not centralized, nor the nodes. Loot at LN like the Internet, there are big sites and small sites, and they coexist pacefuly. Anyway, I recently closed my node, similar situation as you, after I discovered it was really not needed anymore and it is not nice to maintain it, specially if you have a lot of channels. Have a look: https://stacker.news/items/486283
Your experience is really matches mine...I've been running a node for a couple of years, testing different implementations, discrete liquidity and few tens of channels. Recentrly switched to automated channel management scripts and it kind of works better, but overally running a public node it's probably not worth the burdain. A private node with private channels is still annoying to manage if compared to all the other mobile solutions out there.
I've seen that Electrum on Desktop with 5 channels, and Zeus on mobile (Electrum also works on mobile) is really all I need for LN. And I use it everyday. Have a look at those solutions. 👍
Why do you see it as centralization?
The natural result of operating in this system with the value regularly going up over time results in disincentive to spread out your liquidity amongst more counterparties.
With lightning you have centralization of payments and custodianship of btc. Whilst the big blockers supported centralization of nodes running/mining the protocol. Small blockers helped nerds run nodes, big blockers helped users hold/spend their own btc
Fair enough, but it's not a matter of "nerds running nodes VS plebs spending bitcoin". Since we need bitcoin to be technically sound, we need infrastructure before everything else. Infrastructure means that Bitcoin needs to keep its properties through time, thus full nodes must enforce protocol rules. If nobody enforces protocol rules, we are left with a new paypal technology, so what's the point?
Listened to an Andreas talk last night where he advocated for many smaller channels over a few big ones. It's only centralization in the sense that you are tied in with a few other nodes, and it's your decision to distribute the channels.
It feels more centralized, since there are fewer connections. The question is, can you open more nodes, and who controls the code. When looked at from these 2 questions it is as decentralized as ever.
Good take for sure. Although it is interesting yo think of a system that has the potential for decentralization and yet nobody uses those properties. Is that system "decentralized"? Sort of, maybe
Oddly, it's the power that it can be decentralized that is key here. As with any standing army that is ready to fight. It's not that the army needs to fight when the opponent has sufficient fear of it. In this case, having the full ability to go more decentralized at a moments notice can be sufficient deterrent for those who would abuse their power. Also, those who would abuse their power, in some cases, know enough to not provoke and instead to diminish it, making it seem unimportant in the hopes that it will fade and get forgotten. After which the abusers will resume with their intended goals. The power of open source means the code will live on after all of us, hence the reason to ensure the sharing of knowledge about what code there is, how to use it, what it is for, etc.
Lightning's main use case is as an open&permissionless financial B2B integration interface. The other use case is helping amateur node runners lose money.
What’s been your average rate of return on that liquidity over the years?
LOL, negative. You only really need the rare force close to totally wipe out any fees earned from routing.
That’s what I thought… Do you think we’ll get to a point where this makes more sense?
It will make sense for liquidity providers who have other services that naturally drive capital in their direction.
Got it. do you think the optimal solution for an individual then is just to use custodial lightning wallets for small amounts and otherwise stay on-chain?
For now, seems like it.
As above, so below.
As long as your node's degree centrality in the lightning network's graph doesn't go up, it's fine. If more and more nodes have to find paths that go through your node in order to make a payment, it becomes a problem. I don't think the number of channels (edges) matter, as long as payment routes are found.
Centralization? It might come to that. Not everyone is willing to lose money running a lightning node. I didn't do the maths, can it be profitable? Good evening #nostr nostr:nevent1qqsz2k6eheph2c9pvtmt7sq90t2xj6ll0v3y7wrc2utceqj2pzpfmsqpz3mhxue69uhkummnw3ezummcw3ezuer9wcpzpaegm8nwwpyrtrnsjv84efjtp9mhpkvfenvxs487vx8d48y28qgxqvzqqqqqqypyvnc7