Exploring the Centralization Risks of Bitcoin’s Lightning Network

The Lightning Network is seen by the vast majority of Bitcoin Core contributors as the best option for scaling Bitcoin to bring millions of new users to the network. But the system has yet to be implemented on the Bitcoin blockchain, leading many to wonder whether it is wise to place so much hope and trust in an untested solution.

At a recent event at Coinbase’s offices in San Francisco, Lightning Network creators Joseph Poon and Tadge Dryja presented their design for scaling bitcoin through generalized network payment channels. During their talk, Dryja and Poon responded to audience concerns that the Lightning Network could eventually become centralized through something called supernodes.

Some nodes are better connected than others

Coinbase co-founder Fred Ehrsam raised the issue of supernodes during the Lightning Network presentation, and Poon responded:

“I think some nodes are better connected than others. It’s important to make those nodes unidentifiable and easily replaceable. When it comes to highly connected nodes, it’s not going to be cheap, it’s not going to be a supernode like BitTorrent or Skype (the way Skype worked five or ten years ago), just because of the cost involved.”

The costs Poon mentions here are mainly the costs of “locking” bitcoins in order to make payments on the network. From a consumer’s perspective, Dryja and Poon argue that locking money is inappropriate because users can still use the money on the Lightning Network. For those who want to earn money through network payments, there is a time value component.

While many Bitcoin holders hope to get a return on their holdings through the Lightning Network, Dryja and Poon explained that the network will not be the gold rush that some people expect. Both also believe that fees on the network will be zero.

Poon: There is no benefit for these super nodes

The main reason Ponn isn’t too concerned about these potentially well-connected nodes is that there isn’t much economic incentive to create them. “It’s not that it’s impossible, it’s about what you get out of it,” he explained.

Poon explained this point:

“If you have infinite money and you want to reach everyone, then yes, it’s possible, but it’s not really beneficial for you as a node. That’s why I’m not really concerned about the node topology becoming centralized, because there’s no network effect that’s in your favor. Because they’re relatively mutable, so if you start charging a certain rent for it, everyone else will do the same.”

Poon mentioned that well-connected nodes in the network could potentially be used for micropayments. This is because the lower upfront fees associated with routing payments are worth around a penny so could be used for something like buying a cup of coffee.

Dryja said such direct payment channels may be more advantageous between two counterparties, but there is not much difference when such payment methods are between two parties and three or four parties.

Further analysis from BitGo’s Jameson Lopp

BitGo engineer and Statoshi.info creator Jameson Lopp has recently been thinking and writing about the technical challenges facing the Lightning Network. Bitcoin Magazine reached out to Lopp for his thoughts on Poon’s comments. Regarding the incentive (or lack thereof) to run a supernode, Lopp opined:

“If you consider having routing fees to be a positive (because in some cases they are a negative), then there is an incentive to run a highly connected routing node in order to collect the fees. On the other hand, there is a disincentive to run a normal routing node because it requires keeping the private key ‘exposed’ when the machine is connected to the network, which is quite risky.”

Lopp said: "It remains to be seen how the balance between these opposing dynamics will stabilize."

Lopp also suggested that fears about the Lightning Network are not based solely on centralization issues, but rather the risks associated with a particular type of centralization that could go wrong. “The problem with any centralization debate is you need to explain the fear of the potential outcome,” he said.

Lopp went on to explain that the possible centralized aspects of the Lightning Network that worry him is the topological centralization of the node network, as it would make the network more fragile and prone to collapse under stressful circumstances.

Lopp said that a few entities route most payments on the network, and if those entities ran many nodes there might not be as many problems. However, he went on to say:

“There would be a concern that some entity could choose to shut down all nodes, wreaking havoc until lightning network users could ‘heal’ the network by reestablishing new channels through on-chain transactions.”

Lopp also mentioned the potential for censorship and privacy issues associated with centralization in a small number of entities:

“Concentration in a small number of entities will lead to censorship and privacy issues, making it more difficult to route data to uncooperative entities or those that collect it for the purpose of de-anonymization.”

Blockstream’s Rusty Russell, who is working on his own Lightning Network implementation, said he believes the layer-2 protocol will likely feature a Tor-style arms race between those who want to preserve privacy and those who want to abolish it. Russell and other Lightning Network developers are also working on Tor-style onion routing for the network.

Lopp told Bitcoin Magazine that many of the concerns about the Lightning Network only become apparent in conjunction with mempool storage, which can make it difficult to re-establish channels on the Lightning Network.

Coinbase is concerned about centralization risks

Coinbase representatives were more concerned about potential centralization risks in the Lightning Network during Dryja and Poon’s presentation. The Lightning Network’s creators were grateful for the recognition of the importance of decentralization in Bitcoin exchanges.

At least one Coinbase representative mentioned that the company isn’t necessarily looking toward making money from the Lightning Network. For now, it’s simply seen as a way to improve its products and services.