The Lightning Network and other variants of payment channel networks are said to be the future of scaling Bitcoin. A newly published paper by Christian Decker, Conrad Burchert, and Roger Wattenhofer introduces another layer between the blockchain and payment channels, which has the potential to overcome the last limits of scaling.
It is possible that Bitcoin will be able to scaler further on chain. However, after reaching a specific volume, some fundamental properties of the system will have to be given up, to increase capacity further. To prevent this, scientists and developers outlined the concept of using off-chain payment channels; transactions are no longer processed on the blockchain, where they need space and resources, but in payment channels between sender and receiver.
If you are not familiar with the concept of payment channels, you can read about it here. Roughly said, it is like updating a bank transfer form, instead of giving it to your bank. Participants in a payment channel lock funds and exchange updated unconfirmed transactions. Originally, these channels connect only two parties. However, they can connect to a network, like the Lightning Network, through which a sender can route his payment via several hops. You can learn more about how Lightning builds a network from the channels here.
Limits of Lightning
Currently, there exists no network of payment channels, and it is questionable if the networks will make enough sense economically to ever evolve on the free market. But this is no reason not to explore the limits of Lightning and research methods to overcome them; this is the topic of a newly published paper from Christian Decker, Blockstream employee and maybe the world’s first Ph.D. in cryptocurrency, as well as Conrad Burchert and Roger Wattenhofer from the ETH Zurich.
As a starter, the scientists explain two of the problems of Lightning and other off-chain networks. First, scalability is still restricted:
“Even with increases in block size, it was estimated that the blockchain capacity could only support about 800 million users with micropayment channels due to the number of on-chain transactions required to open and close channels.”
Broad adoption of micropayments, like the Internet of Things could require at some point, would go beyond the capacity of Lightning, since every off-chain channel needs his opening and closing transactions on chain.
Second, both parties of a payment channel find itself in a dilemma when locking in funds in a channel. The authors write:
“The locked-in funds should be sufficient to provide enough capacity for peaks of transactions. There is a conflict of the two aims to have a low amount of funds locked up in a channel, while at the same time being flexible for these peaks.”
Most people who have dug into the topic agree that these problems are hard to overcome. Decker and his colleagues, however, found a potential solution which eliminates both issues at the same time.
When Layer 2 becomes Layer 3
The idea of Decker, et al., is simple; if the opening and closing of payment channels require too much space on the blockchain, just do it off chain. “Payment channels will not appear in the blockchain, except in the case of disputes,” the authors explain their solution, “Users will be able to enter the system with one blockchain transaction and then open many channels without further blockchain contact.”
Instead of using one transaction to open each payment channel, users can be able to use one transaction to enter a lot of payment channels. How is this possible? The key is that the funds are not locked with a single partner, as you do with a payment channel, but on a multisig address owned by a group of people. So it gets possible to participate in any payment channel which is entertained by any user participating in the funding address.
The scientists name their innovation ‘Channel Factories.’ These are “a new layer between the blockchain and the payment network, giving a three-layered system.” You can imagine it as an airport. Up until now, everyone has used a direct flight from their home to their friend, and hoped, that from there they find another direct flight to their destination. With ‘Channel Factories,’ it is as if you drive to the airport, and on arrival, you can choose your flight to your destination.
Scaling without Limits
The architecture of putting the factory layer between the blockchain and the payment channels solves another problem of off-chain networks like Lightning.
If you create a channel, you have to fund it with the amount of money you want to send through it. If you create more channels, you have to fund each one. For this, it is expected that most users will create a few, or just one, payment channels with a major hub which does the routing for them which could result in a highly centralized topology of the network.
The channel factories of Decker and his co-authors could be a game changer; they enable the user to enter a variety of channels with one single transaction, which can help to keep the network decentralized, if you do not consider the factories themselves as centers. However, the extent of decentralization will depend on the precise implementation and organization of such factories.
If you look at the details, channel factories obviously increase the complexity of off-chain payment networks, which already is immense. Most likely, the factories are not designed for an immediate deployment, but rather as an optimization, after the first nodes and ties of the Lightning Network have been established. Nevertheless, the advantages the factories provide can be vast, according to the calculations in the paper.
If only three parties come together, to co-use three payment channels, the required space on blockchain is reduced by 50 percent. If you assume that 20 parties cooperate and use 100 channels, the ‘channel factories’ can save 90 percent. If you further assume that you use Schnorr signatures to aggregate input signatures, you can even save 96 percent. In other words, if you put another layer between the blockchain and payment channels, there might be enough capacity to settle all transactions in the world, by humans and by machines, on the blockchain.
The concept, as it is outlined in the paper, is in a rudimentary state. The scientists only describe the mechanism by which the ‘channel factories’ can be built. They remain vague when it comes to the topology of the network. Also, they do not mention economic issues or talk about the organization of the factories, but this is not what the paper is about. It demonstrates, rather, that there is the possibility to break the long-term restrictions of off-chain payment networks.