Sequoia, Baidu Back Turing Award Winner’s Blockchain Project
Announced Tuesday, the Conflux Foundation is registered in Singapore and will use the proceeds to fuel work on its network. Backers include venture capital firms Sequoia China and Baidu Ventures, crypto mining firm F2Pool and exchange Huobi, as well as Metastable and IMO Ventures.
With the new capital, the foundation said it’s now looking to expand its 10-person development team. The goal: roll out a testing environment in February and officially launch a new public blockchain around the third quarter of next year.
The project was co-founded by scholars from the University of Toronto and China’s Tsinghua University, including Dr. Andrew Chi-Chih Yao, a Turing Award winner and an information sciences professor at Tsinghua. Named for renowned mathematician Alan Turing, the Turing award is given out yearly by the Association for Computing Machinery (ACM) and is considered a kind of Nobel prize for computer science.
According to an academic paper published in May, the group conducted a month-long experiment for the Conflux prototype late last year. Their efforts utilized the code of the bitcoin blockchain but changed the protocol from the Nakamoto consensus model, named for bitcoin’s creator, to Conflux’s own design.
“We implemented a prototype of Conflux and evaluated Conflux by deploying up to 20k Conflux full nodes on 800 Amazon EC2 virtual machines,” the paper stated. “The throughput is equivalent to 6,400 transactions per second for typical bitcoin transactions.”
In an interview with CoinDesk, Dr. Fan Long, a co-founder of the project and an assistant professor of computer science at the University of Toronto, said the key design decision that he believes could help to scale public blockchains is to change the way blocks are ordered.
Currently, most blockchains work in a rather linear way in that only one block can be produced at a time and added to the end of a chain. In the event of two blocks being created at the same time, the network relies on nodes to agree on which one will continue the chain.
Long said this feature of blockchain results in a bottleneck that is problematic for most public networks, especially those like ethereum that aim to power smart contracts and decentralized applications.
“Some may have a higher scalability, but to some extent at the expense of a complete decentralization,” Long said.
The Conflux protocol, however, is designed to allow blocks to be produced concurrently to boost the volume of transactions.
But to avoid forks, or the emergence of two competing versions of the ledger, Long said the group designed an ordering algorithm based on the concept of a directed acyclic graph (DAG) that is able to make sure every block will be eventually arranged in a sequence by the network.
According to Long, while the foundation utilized most of the original bitcoin blockchain code during its lab experiment, it will develop its own infrastructure for the public launch next year as the team hopes to include smart contract features.
Further, the foundation said its investors, such as Baidu, will commit to using the network for future development.
Long went on to say that the end goal is to create a scalable public blockchain with smart contracts that can deliver decentralized applications – and he made a sly dig at ethereum, whose scaling challenges were laid bare a year ago when CryptoKitties clogged that public blockchain network.
“That goal sounds easy and similar to what every other project is saying. But it’s actually really difficult,” Long said, concluding:
“Because what we need is a highly productive public network for high-value applications, not just several blockchain-based games.”
Andrew Yao image courtesy of Conflux Foundation