In the quest for a reliable proof of stake algorithm, IOHK CEO and Ethereum Classic developer Charles Hoskinson believes his company has made a breakthrough.
Charles Hoskinson was one of the original founders and the initial CEO of Ethereum, but left the project before Ethereum's official launch in 2015. Along with another Ethereum veteran, Jeremy Wood, Hoskinson went on to Input/Output, or IOHK, a research and development company dedicated to cryptographic research and the architecture of digital currencies.
Since Ethereum Classic split from Ethereum, Hoskinson became involved with the former. He recently announced “the Grothendieck Team,” a new development team dedicated to the digital currency and smart platform project.
Meanwhile, IOHK, collaborating closely with university academics, has been working on a proof of stake algorithm dubbed “Ouroboros.” Whereas most digital currencies — including bitcoin — rely on proof of work by miners to reach consensus over the state of the ledger, proof of stake systems reach a consensus voted on by coin holders.
Proof of stake has been considered and experimented on for several years now, as some believe it may be superior to proof of work.
“The obvious benefit of proof of stake over proof of work is that it doesn’t require miners to burn energy, which makes it much more efficient,” Hoskinson told Bitcoin Magazine. “But there are more advantages to it. It also allows for next-generation governance systems, for example, allowing stakeholders to indicate their preference on hard or soft forks.”
Ouroboros
Ouroboros was developed by professors Aggelos Kiayias, Alexander Russel and Roman Olynykov, as well as Bernardo David.
In Ouroboros, new blocks are produced by randomly selected coin holders, based on a probability proportional to their stake. This is realized through a sort of lottery, which places the coin holders in dedicated timeslots that indicate their turn to produce a new block to update the blockchain. If a coin holder does not produce a block within their designated timeslot, their turn is simply skipped.
The main challenge of any proof of stake system is know as the “nothing at stake problem.” If the blockchain forks to two or more competing chains, the optimal strategy for any coin holder is to extend every fork. That way, it gets its block reward no matter which fork wins or multiple block rewards if several chains continue to exist.
But this is also at odds with a central design goal of blockchains: the incentive for everyone to converge onto a single chain.
Speaking to Bitcoin Magazine, Kiayias said this is to a large extent solved by a novel solution to achieve randomness for the lotteries. This should make the probability of stakeholders successfully forking the chain negligible.
“Imagine that you have an initial population of stakeholders,” Kiayias said, explaining how this randomness is achieved. “Shortly prior to the system initialization, an initial lottery is conducted to choose a committee. All of this information becomes part of the genesis block and the protocol starts. The initial committee is responsible for advancing the blockchain the first time around and running a secure multiparty computation protocol that realizes a publicly observable ‘beacon.’ It’s this beacon that generates the randomness for the lottery.”
Once this lottery has selected an initial round of winners, and the blockchain is in motion, this process is repeated. This is called a new “epoch.”
“At the end of the first epoch, the seed for the second epoch is computed and a new ‘genesis’ information is calculated,” Kiayias said. “This includes the multi-party computation protocol computed seed, and a stakeholder distribution reflecting stake shifts from the initial stakeholder set. The protocol continues in the same fashion moving from epoch to epoch.”
While noting that Ouroboros is provably secure — a mathematical method of proving a protocol works as intended — Hoskinson did add that this does not mean all challenges pertaining proof of stake are completely solved.
Scaling Ouroboros to a global user base will require additional research and engineering, he said, while it's also not always possible to drop out of the consensus process to securely rejoin later.
Ethereum Classic
Ethereum has always planned to switch to proof of stake at some point in the future, using a yet-to-be-finalized protocol dubbed “Casper.”
Although the initial plan for Ethereum Classic was to copy Ethereum in every step of the development process, it soon became clear that many in the community preferred for the project to take on its own path. The preferred consensus algorithm, whether it be proof of stake or proof of work, has always been a central topic in that debate.
For now, it seems the Ethereum Classic community is generally in favor of proof of work, and will most likely stick to that proven and tested consensus algorithm for the foreseeable future. The project recently even hard forked to diffuse the so-called “difficulty bomb,” which ensured sticking to a proof of work consensus algorithm remains an option.
Despite Hoskinson’s involvement with Ethereum Classic, it seems unlikely the project will adopt IOHK’s Ouroboros.
“And that’s fine. We shouldn’t fall in love with one idea,” Hoskinson said. “While I do think Ethereum Classic’s consensus algorithm, ‘Ghost,’ needs to be removed or improved, sticking to proof of work will help us differentiate from Ethereum, so the two projects can complement each other instead of competing. This could also help Ethereum, as miners will have a place to go when they make the switch, which should make their transition easier.”
For more information on Ouroboros, visit the IOHK website and white paper and its GitHub page.