It’s possible bitcoin’s Lightning Network will fall short of its lofty ambition – at least, that’s what a group of researchers are arguing.
Often trumpeted as the future of bitcoin, Lightning’s particular twist on payments channels has served as one of the foundational ideas for scaling all public blockchains. By proposing to push transactions off of the bitcoin blockchain, but keeping strong cryptographic guarantees intact, it’s believed by many developers to be the best way to increase the number of transactions the network can handle.
But the pressure for Lightning may be on. With bitcoin’s debate over its own scaling roadmap now paused, some businesses, upset by the decision, are watching and waiting to see how Lightning develops. Others are packing bags – threatening to migrate to alternative blockchains with fewer users (and thus less expensive transactions).
But, if new research is any indication, there could be some bumps in the road ahead.
Specifically, computer science researchers Simina Brânzei and Eral Segal-Halevi and assistant professor Aviv Zohar have been exploring what a Lightning Network economy might look like. At issue is whether people will actually use it, or if they will simply send normal bitcoin transactions.
In this way, Zohar argues enthusiastically that, although much of bitcoin’s scaling debate has centered on technological theory (whether Lightning transactions could be achieved), not much has been done in the way of determining how consumers and businesses will actually react to it.
He told :
“Sometimes people talk about Lightning as if it’s a done deal. They point out there are early implementations and all the work that’s been done so far. But I think the fate of it will be determined partly by economic forces.”
But Zohar and his colleagues are also making assumptions. As part of their research, yet to be formalized in a white paper, they are building off of a generalization that there are two main types of transactions.
The first are small transactions, which most people make on a frequent basis, say, to buy a cup of coffee en route to work. The second group consists of larger transactions, say investing in a laptop or a hot-air balloon. These purchases are made less often, but are still important.
With this in mind, the researchers guess Lightning will be used for the smaller transactions, while on-chain transactions, which come with the security of having been finalized and recorded on bitcoin’s global ledger, will be used for the larger ones.
In other words, users will probably pick the cheapest option. And the cheapest option varies.
As such, Zohar believes there could be some competition between Lightning transactions and on-chain transactions. Notably, that’s directly related to bitcoin’s limited block size.
Only those that pay enough in fees will be able to get their transaction in the blockchain, and that’s significant, since users (and businesses) are already complaining about fees today.
So, will one transaction type be cheaper overall and thus crowd out the other? Or will the two types coexist? Put another way, how well does Lightning fare against on-chain transactions?
That’s the question the group is attempting to answer with economic experiments.
The researchers used the above assumptions to build two simple bitcoin economy models. The first “toy model” is biased toward Lightning, Zohar said. The economy has a Lightning Network, where users are inclined to make smaller transactions.
“What we want to see is how many Lightning transactions would happen and how many big transactions. What’s the fraction of each?” he asked.
What they found, in this case, is that most transactions are conducted using the Lightning Network. An overwhelming percentage, in fact. Yet, as far as the transaction goes, roughly two-thirds still occurs on the blockchain.
These volume results might seem “disappointing” in that some bitcoin developers envision Lightning Network replacing most, if not all, bitcoin transactions, Zohar remarked.
The second simple model – another “extreme” case that Zohar said doesn’t model what people really do – sees if this changes at all when every transaction in the economy is the same size: Users just pay 1 BTC for everything.
The takeaway, said Zohar, is that the results look different in each model. Lightning thrives in the first model, while it’s “destroyed” in the second because users would rather pay the on-chain fees.
But models aren’t reality – they’re simply the best available guess at the outcome.
“I do have to say, this is not the world of Lightning Network in 10 years,” Zohar said, adding that there are many variables the researchers couldn’t capture in their model, such as spending habits among different groups of people.
Plus, better wallet user experiences might one day hide these sorts of questions. Just like most internet users don’t think twice about their tweets are sent across the web, users won’t think about how exactly their payments are routed over the bitcoin network, whether it be on a top or bottom part of the stack.
In other words, it’s impossible to predict the future. “Our model is very naive,” he said. “Don’t take these numbers too seriously.”
Other Lightning Network developers argued similarly.
Fabrice Drouin, the CTO of Paris-based Lightning startup ACINQ, thinks the research could be “very important,” even while he doubts that these models will be very accurate at such an early stage.
“Lightning is not live yet and it is impossible to predict how it will be used and what the network will look like,” Drouin said. “[Because] there is no real money involved, [the models’] behavior cannot really help when it comes to predicting what will happen on [the bitcoin blockchain].”
But even if the research can’t capture everything, Zohar is asking questions to set more appropriate expectations for the technology.
He told :
“Our main message is not whether Lightning succeeds or fails. But that there are economic effects based on who wants to pay who and how.”
Now, Zohar and other researchers are working on yet another model that falls in between the two “extremes” described above, with plans to release their full findings in a paper soon.