Running a Lightning Node is Hard—Users Need Good UX
The lightning network has been growing at an exponential pace since its inception in 2018. Growth rates year-over-year have been at over 127%. With the number of clearnet nodes peaking at just over 20,000 earlier in March of 2022, the number has since fallen and stabilized at just over 17,000. Why?
No doubt a significant portion of node growth can be attributed to the rise of Umbrel, Voltage, Start9, and other companies that made it easier than ever to spin up a lightning node. However, there’s more to running a node than meets the marketing, and many node operators are familiar with this.
Problems with Operating a Node
Nodes need to be connected to a stable power source with a stable network connection 24/7 else risk a loss of funds. If they submit an outdated channel state to the network, their counter-party can revoke the entire balance locked in that channel. Throughout the Global South, where many Bitcoiners see the largest potential of adoption, this is a huge risk. Power and internet are not guaranteed 24/7. Of course, running a node in a datacenter with enterprise-grade technology and security solves this, but that always comes at the cost of self-sovereignty—you no longer have control over the hardware that your node is running on.
Finding high quality channel peers can be intimidating for a neophyte, and requires some searching on Amboss, 1ML, or any number of lightning telegram groups. There are now services offering inbound liquidity, but if you want inbound from nodes other than the most centralized in the network, you're going to need to go on a search through a fragmented market. For the plebs who enjoy the work of diving deep into a new technology, the search is rewarding. For the users who simply want to utilize the network in a self-custodial manner and receive the benefits (making/receiving payments or routing for yield), the user experience is far from optimal. The founders at Wavlake learned firsthand how difficult it was to onboard artists with their own lightning node. Spoiler alert: it was very difficult, and they've now pivoted their model. The process of running a node currently appeals to those who love the technology and the vision, not to those who need a lightning node for their business or to facilitate their payments.
Even for highly technical individuals, it's difficult to maintain a node. Liquidity management is something that requires frequent attention, and the more channels you have, the more channels you need to monitor. Zero fee routing recently shut down his node due to the time and energy it costs to maintain even despite his love for the lightning network. There's a reason why companies like River, Bitfinex, or Kollider have full-time staff dedicated to the management of their node(s). It's tough to run a lightning node, and many people dove in too deep too early. In my understanding, this is the reason for those 3,000 nodes going offline since March of 2022.
For the uninitiated non-technical folk who simply want to make or receive payments on lightning, the solutions available are only custodial. That begs the question: are we simply living in an in-between space where custodial lightning solutions are the bridge to a truly self-sovereign payments infrastructure, or is the future of lightning custodial? Another "decentralized" network that will grow under the purview of an oligopoly, under the will of the executives and board of directors?
On Network (De)centralization
This paper analyzes the lightning network in its current state, and has a bleak outlook on its decentralization. As most people familiar with the lightning network already know, a small number of nodes hold the vast majority of connectivity and effectiveness. With 30% of nodes on AWS and 11.8% on LunaNode, one can wonder what might happen if a policy shift happened in the headquarters of big tech that presents Bitcoin Lightning in an unfavorable light. The “they fight you” stage is already here, so it seems to be only a matter of time before some real heat from the state is brought to the table.
Networks tend towards centralization—it's a law of nature. In nearly every city of the world, you'll be able to get a sim card, purchase mobile data, download a ride hailing app, and book a cab to a hotel where you'll have a bed, shower, and possibly something to eat. These ways of living have coalesced around a centralized process facilitated by familiar institutions (ride-hailing apps, the hospitality industry, telecommunications infrastructure, payments, and so on). The motif of technology companies gobbling up services, as a16z puts it, is not far off. However, it's not just technology, it's people, it's mimetic, and it’s natural. For a network to maintain itself as decentralized, it must begin and grow in a decentralized fashion. The myth of centralization paving the way to decentralization is just that—a myth. The United States congress dissolving Standard Oil into a multitude of smaller companies didn’t make the oil industry any less centralized. Look at any crypto company or protocol in existence. None of them will become decentralized as time goes by—they're all already operating as centralized corporate entities, and that's because they started out that way.
Lightning doesn’t need to end up this way. The authors of this paper simulated the lightning network with and without hubs or service providers, and the results look promising! The lightning network is resilient to the loss of hubs, and can facilitate payments reliably with passively balanced channels. However, if lightning continues on its path of highly technical node operation, the only players that will be able to offer good UX to users will be well funded tech companies who have the knowledge and experience to do so. As Bitcoiners, we should ask ourselves: do we want that future?
The Future of Lightning
Despite all of this, I'm optimistic for the future of lightning. Whenever I use lightning, it truly feels like magic: a teleportation of value across space and time—a near instantaneous settlement of value across the globe without trusted third parties. And even beyond that, lightning is another layer of the internet that enables authentication, identity, and commerce under cryptography, not trusted third parties. I see a future where it is possible to automate away the granular technicalities and release this technology to a wide array of people. However, before the lightning network is ready for that growth, it's necessary to lower the bar to operate a node in a self-custodial manner. Users need good UX. That's the only way we maintain the cypherpunk security promises of Bitcoin.
Nothing rises higher than its source. Bitcoin is decentralized because it was decentralized at its source. If the lightning network is to be decentralized in the future, it must grow in a decentralized fashion moving forward. That means more users need to become lightning node operators. For that to happen, we need to lower the bar for lightning node operations. That means more automation, and better user experience. That’s what we’re working on at Birkeland.