Distributed Fault-Tolerant Financial Transactions Service
summary
A description of building a fault-tolerant, distributed, and consistent financial transaction service capable of 100,000 transactions per second. Covers why such a system seems impossible under the CAP theorem, how they worked within that limitation, use of test-driven development and numerical methods to rigorously prove the system’s behavior, and potential applicability of the same technologies to cryptocurrency development.
AI-written summary · about
chapters (15)
- 0:00Introduction
- 1:10Standard architecture issues
- 3:17CAP theorem limits
- 6:09Bitcoin example discrete limits
- 9:23Financial service requirements
- 11:11Sharding and batch processing
- 13:00Consensus algorithm overview
- 15:52Paxos modification
- 19:05Distributed system testing
- 27:33Cycles problem probabilistic approach
- 33:34Malicious admin role
- 34:58Blockchain solution transaction history
- 36:03Implementation recommendations
- 38:14Summary
- 39:03Q&A
related talks
34mDesigning, Building and Testing Deterministic HFT Systems
22mDistributed Transaction Management in Microservice Architecture
36mFlexible Byzantine Fault Tolerance
36mThe FIDS Theorems: Tensions between Multinode and Multicore Performance in Transactional Systems
01h 12mScaling distributed ledgers: Achieving arbitrary throughput and sub-second latency
01h 05mPBFT and Blockchains
18mEngineering Resilience at Scale; Lessons from a Distributed Systems Transformation
28mPractical Approaches To Scaling Distributed Transactions