Rongjian Lan was a search infrastructure engineer for Play Store at Google. He published over 10 academic papers on spatio-temporal querying and map-based visualization. Rongjian started researching decentralized protocols in early 2017.
Rongjian is the co-chair of ABC Blockchain Foundation, with more than 100 engineers from Google, Facebook, LinkedIn as members. He was a doctoral candidate of computer science at the University of Maryland College Park and obtained his bachelor’s degree from the University of Science and Technology Beijing.
- Monthly Deliverables: 2021 Oct
- Optimize consensus block verification logic to reduce block time by ~200-400ms
- Ramp up common prefix team on harmony technology and help propose research directions/topics for research DAO
- Finish design proposal and specification for cross-shard transaction with mutual knowledge system team.
- Quarterly Milestones: 2021 Q4
- Reduce mainnet block time down to 1 second with: 1) commit signatures collection optimization; 2) Parallel message processing in leader node; 3) Upgrade with blst library
- Work with mutual knowledge system team to finish the PoC implementation of cross-shard transaction
- Hire 1-2 core protocol engineers
- Yearly Planning: 2022 Q1 - Q3
- Q1: Leader rotation and malicious leader behavior detection integrated with view change: failure of proposing cross-links, txn censoring etc. Launch cross-shard transaction feature.
- Q2: Cryptography for zk-based light client and resharding and private election etc.
- Q3: Finish implementation of resharding: random resharding, fast state sync etc. Launch resharding and make Harmony protocol 100% decentralized on mainnet.
Past 6-Month Achievements
- Implemented and launched Harmony VRF the natively integrated randomness source for Dapps: https://medium.com/harmony-one/introducing-harmony-vrf-4fc51e175c2
- Implemented and launched 3 validator governance proposal: 1) Setting validator minimum fees; 2) EPoS parameters change; 3) Minimum delegation reduced from 1000 to 100.
- Added various optimization to bring down shard 0 block time from ~3.8s down to 1.7s.
- Resolved with the team and publicly addressed the major RPC outage incident: https://talk.harmony.one/t/update-on-the-recent-mainnet-stability-issues-june-2021/1527
- Prepared and gave talk "Play2Earn: the future of gaming" at our Shanghai metaverse event at Palo Alto.
- Analyze consensus bottleneck on block processing and reduce unnecessary processing of blocks through result caching.
- Meeting with research Dao and answer their questions on harmony protocol.
- Figured out how to update the base point for blst library with Ganesha
- Looked into BLE12-381 curve specification https://hackmd.io/@benjaminion/bls12-381
- Found the place in blst for base point setting: https://github.com/supranational/blst/blob/master/src/e1.c#L23. But it needs more help from the blst team on how to set it correctly.
- Fixed gas estimation rpc issue on transaction with payload: https://github.com/harmony-one/harmony/pull/3900
- Change consensus block verification logic: optimistically verify new blocks in prepare phase so it won't be a bottleneck in the commit phase.
- Add block content in the announce message of the leader
- Refactor block verification logic as a utility to be used by both announce and prepared message handling in validator node.
- Sync with Dimitris from common prefix for more harmony tech clarification.
- Worked on developer request to support pending txns in event subscription and allow fetching of pending txns in rpc.
- Had a tech intro session with Dionysis, Zeta and team for them to ramp up on Harmony tech and potential research ideas. They've started looking into a few interesting research topics and we will have weekly sync to clarify more questions.
- Looked into Solana's tech again and most of their tech speedup is due to their unique PoH consensus and a faster VM. So not much directly transferable tech we can immeditely borrow. We could think about speedup evm if possible or switching to other vm if speed is the most concern.
- Working on a consensus optimization which could potentially reduce block time by another 200ms-400ms. ETA this week.
2021/10/06: Week: 1.5 second block time Month: cross-shard proposal, research dao onboarding Quarter: work towards 1 second finality, hire 2 core protocol engineers