What Makes Bitcoin Great? One Scientist is On a Quest to Find Out
Visa, PayPal, bitcoin. The last, it seems, is not like the others.
You might be thinking, well, of course. It’s unique compared to older institutions, ones that first made digital currency possible by storing payment data in centralized databases.
But, that might be only one way of looking at it.
Computer scientists and developers are quick to note that the reason bitcoin has succeeded at decentralizing its monetary system is because it improves on past computer consensus protocols, a point that Cornell associate professor Elaine Shi stressed in her presentation at the recent Stanford blockchain security conference.
Even after 30 years of research, Shi explained, classical consensus protocols fail under certain conditions. But she believes bitcoin is different because it’s more “robust”.
However, defining and mathematically spelling out these differences isn’t so easy to do.
Shi told CoinDesk:
“The protocol’s success is kind of ahead of the scientific understanding.”
Despite the challenge, the academic seems determined to catch up.
Sitting in the cold after a long day of security presentations, she chatted excitedly about bitcoin’s unique properties.
She noted that other recent research has sought to develop a formal security proof for bitcoin, and that thinkers from IC3 and elsewhere are now looking to help flag potential vulnerabilities and to inform future research into the protocol.
Shi’s curiosity was first piqued in 2010 or 2011 while she was working for the technology company Xerox PARC in Palo Alto.
It was then that her friend, a hobbyist and miner, showed her the bitcoin white paper. They read through it together, fascinated.
“We tried to understand why bitcoin took off,” she said.
From her point of view, it was a big deal that the currency saw so much use compared to ecash, a technology put into the world by long-time cryptographer David Chaum in the mid-1990s.
“At that time, they adopted more sophisticated crypto. But there wasn’t so much traction,” she said.
She added that she was impressed that bitcoin, in contrast, saw more rapid adoption and uses simple cryptography – public key encryption, signatures and hash functions.
“One big thing for bitcoin was that it made the incentives right. It gave incentives to early adopters. There are various other aspects that maybe it did right in terms of incentives and possibly helped with adoption and how it gained popularity,” she added.
Later on, Shi moved to University of Maryland, where she continued her bitcoin research, and then to Cornell’s Initiative For CryptoCurrencies Contracts (IC3), the university’s center for study on all things blockchain.
Her presentation at Stanford, “Rethinking Large-Scale Consensus,” discussed her new research, aimed at rethinking how bitcoin might work differently, but retain its unique properties. The result is her proposed ‘sleepy‘ model of consensus.
She noted that when she asked why people were exploring the use of a blockchain rather than a long-studied classical protocol, such as PBFT, people would typically respond “because it’s more robust”.
This is the common wisdom. But, she noted that from an academic perspective, it’s been difficult to even define what ‘robust’ means exactly.
In this light, ‘sleepy consensus’ explores a specific piece of bitcoin’s robustness: sporadic participation, where nodes can leave and enter the system as they please. It further examines whether a system can be as robust without proof-of-work, the algorithm that leads to one agreed-upon transaction history
In Shi’s model, there are ‘sleepy’ nodes (that are offline) and ‘awake’ nodes (that are online and active).
Shi displayed images of Snow White to show each state, and to demonstrate that that nodes can shift between these two states.
“For example, when the prince kisses Snow White, she wakes up and continues to participate,” she said. “Snow White is a very robust princess.”
New protocol, new problems
One way to test the robustness of the system is to see whether it can come to agreement when 51% of the online nodes are ‘honest’ (and therefore will not accept an invalid transaction), even with this property of sporadic participation.
Classical models fail here. In fact, Shi went as far as to say that no classical protocol, whether synchronous or asynchronous, holds up. Not even when 99% of the online nodes are honest.
She concluded that bitcoin, as conventional wisdom says, is indeed robust. It’s a system that’s been up and running for eight years, and that continues to work as long as 51% of nodes are honest.
‘Sleepy’ consensus builds on that robustness, but rearranges the protocol in a way that ditches bitcoin’s proof-of-work.
The research team found that the tweaked system was more robust in some ways, but with the new construction, new security problems also sprung up.
Work is ongoing here, and Shi said that, for now, the protocol is suitable for consortium blockchains along the lines of those released by the Linux-led Hyperledger.
Though, again, there are perhaps other elements to bitcoin’s ‘robustness’.
Another project from Shi and IC3, FruitChains, explores bitcoin’s game theoretical component, or how it incentivizes participants to act in a way that ultimately benefits everyone.
The result of the research is a proposal for a ‘fair blockchain’, where block rewards and transaction fees are evenly distributed and there’s less variance in rewards.
Analyzing each piece on its own could lead to something bigger.
“In general, we need a new scientific foundation for all of this,” Shi said.
Yet, Shi emphasized that this research is not about defining things just for the sake of academic curiosity.
Once people understand the protocols better, there are different, perhaps unexpected, directions to go. Broadly, researchers will have a better understanding of how public blockchains can be improved.
Proof-of-work is expensive, for example, as powerful computers from around the world are currently hashing puzzles at dizzying rates to secure blockchains like bitcoin and ethereum. Many researchers, such as those working on proof-of-stake for ethereum, are trying to develop a way around these massive electricity demands.
More research could help determine whether or not those efforts are in vain.
Furthermore, Shi argues that it’s important to work on understanding the security of the protocol, and writing up mathematical proofs that could potentially bring to light hidden protocol flaws.
“People have somehow developed these very nice intuitions, but it’s still very, very difficult to like design a provably correct protocol. That’s very, very important when you’re dealing with something like cryptocurrency, because if the protocol is broken then your money is at stake,” she said.
A ‘provably correct’ protocol, on the other hand, is one that satisfies certain mathematical requirements.
She mentioned that such a protocol could help ward off future situations along the lines of The DAO – the ethereum project that ended in failure.
“It’s very easy to make a mistake unless you go through this whole process,” she said. “I think that both in academia and in industry there’s this huge need for these protocols, including both consensus and cryptography.”
She also argued that smart contracts require more advanced cryptography protocols.
“IC3 would like to help make these secure by constructing protocols. And deploy them in the real world,” she added.
Beyond all that, Shi has other research ideas.
One potential project is to design a programming language that would let coders with little knowledge of cryptography create more secure apps. Programmers could state vaguely what security properties they need, and the programming language itself would decide what consensus protocol would be best used under the hood.
To Shi, the ability to combine disciplines in such a way is partly what’s so exciting. And, bitcoin is a rich area to experiment with cryptography in particular, she said.
“This is like the goldmine of problems.”
Bitcoin maze image via Shutterstock
You want the latest news about Crypto? Sign up to our weekly Newsletter!