In The past five days, our CEO Ming Chan and I have had the opportunity to visit China, Hong Kong, and Shanghai for a series of back to back conferences and events. Our team also continued to work on our usual tasks of research, development, DEVcon planning, and administrative matters.
Both conferences had a positive impact on the advancement of the Ethereum ecosystem. At the Hong Kong event, we took part in semi-closed sessions on topics such as identity, reputation, and decentralized governance. Robin Hanson proposed the use of prediction markets as a form of low-cost microgovernance. Anybody with access to private information can participate in the prediction markets, making it possible to get an accurate prediction of what the underlying court or arbitration system would usually come up with.
The Shanghai Conference was highly productive as well, with the first day featuring speeches and panels, and the second day devoted to topic-focused sessions. Moderating technical discussions about consensus and scalability was a great experience. We were impressed by the participation of some traditional industry players, such as Intel and Huawei.
Research and Protocol Development
We believe there are four main areas of research that need to be undertaken in order to move Ethereum from what it is today to its ultimate potential. These include integration with zk-SNARKs, Casper, scalability, and updates to the EVM.
- Integration with zk-SNARKS– Leveraging EVM’s existing 256 bit modular arithmetic, it is possible to enable either a new opcode or use the existing opcode to verify succinct zero-knowledge proofs. This would allow for the creation of an ultra-private cryptocurrency mixer and help to preserve privacy. The initial goal is to use it for two-party financial contracts, and the long-term goal is to implement Hawk in Ethereum.
- Casper: Ethereum’s Proof-of-Stake Algorithm is being developed by Vlad Zamfir with help from Lucius Greg Meredith and others. Key components include consensus by blocks instead of consensus through chains, “economic consensus by bet”, finality, and a much lower power usage. The initial goal is to reduce block time to 4 seconds, and the team is currently working to formalize and implement the non-economic part of the algorithm in order to mathematically prove its convergence properties, followed by optimization of the economics.
- Scalability– Through a combination of random sampling, sharding, heavy use of Merkle proofs, and asynchronous calls, it should be possible to increase transaction throughput from 10-20 transactions per minute to more than 100000 (or, theoretically, unlimited if you use super-quad versions). Our research team is confident about the general approach to scalability, but the details of how to make optimal tradeoffs while preserving as much functionality as possible remain a mystery. We plan to upgrade the scalability paper to make it more concise, readable, and up-to-date with the latest thinking.
- EVM Updates: Martin Becze has been exploring WebAssembly as a potential upgrade to the Ethereum Virtual Machine. WebAssembly allows for the creation of an EVM compiler, and environment opcodes can be removed by creating an ABI over a generic interface for external functions. This leaves open research questions regarding middleware “on” Ethereum, such as on-chain and formal services, decentralized governance, Solidity verification, and prediction market implementations.
The internal goal is to launch Homestead when the network has been running without any major issues for the past 4 weeks. We are almost there, and we plan to soon unveil a more detailed strategy. DEVcon is still scheduled for November 9-13 in London and we look forward to seeing you all there!