RainBlock: Faster Transaction Processing in Public Blockchains
summary
RainBlock removes the I/O bottleneck in public proof-of-work chains by decoupling state into sharded storage nodes, adding io helpers that prefetch compact witnesses, and storing a miner-side top layer using DSM-TREE. A single RainBlock miner hits ~27.4K tx/s (27× a single Ethereum miner) with 21 servers; geo-distributed deployment across four regions sustains ~20K tx/s with 32 servers, and cross-layer optimizations cut network traffic by up to 95%.
AI-written summary · about
chapters (13)
- 0:00Intro
- 0:23Blockchains: Decentralized Databases
- 0:46Public Blockchains: Proof-of-work consensus
- 1:36Public Blockchains - Low throughput
- 2:09Prior work: Modify Pow or New Consensus
- 2:30Can we Increase the Throughput of Public Blockchains Without Modifying Pow?
- 4:29Impact of I/O on Overall Throughput
- 5:43RainBlock: Faster Transaction Processing
- 6:26Transaction Processing in Ethereum
- 8:22RainBlock: Architecture for Public Blockchains
- 9:17RainBlock: Challenges
- 9:57Handling Concurrent Operations
- 11:37Life of a Transaction in RainBlock Miners do not perform VD in the critical path
related talks
52mScaling Bitcoin to Physical Limits
30mFireside Chat
36mDecrypting Tech Hype for the Busy Coder
11mBuilding the First Real-Time Blockchain
28mDistributed Ledger
23mScaling Wars: The Race to Speed Up On-Chain Infrastructure
39mHow Teranode Delivers Blockchain Scaling for Enterprise Use
22mA Decentralized Blockchain with High Throughput and Fast Confirmation