Using Rails to communicate with the New York Stock Exchange
summary
Claim: a Rails app handles real-time NYSE traffic by decoding JSON-over-TCP through a VPN, queuing updates in RabbitMQ and using RPC to map responses back to HTTPS requests. Built: an architecture that broadcasts WSS price updates, batches rapid ticks to reduce frontend noise, caches charts in Redis, and uses feature flags, tests and an internal Sentry clone. Numbers: RabbitMQ sustains 44,000 messages per second. Tools: RabbitMQ RPC, WSS updates and HTTPS transactions, feature flags for safe deploys.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 0:35Takeaways
- 2:07Connecting to the Exchange
- 6:58Messages
- 14:45Risk Mitigation
- 15:44Documentation
- 16:52Validation
- 18:54Performance
- 20:04Project Stability
- 21:12Feature Flags
- 22:12Flipper
- 23:28Summary
- 25:29Conclusion
related talks
39mBuilding a highly scalable exchange - To react or not to react
29mRealtime web applications with streaming REST
28mReal-Time Rails
30mSome Funny Things Happened on The Way to A Service Ecosystem
34mIndustrial Rails
07mBuilding Real-Time App with React/Redux/Rails/RethinkDB
38mHigh Concurrent Ruby Web Development Without Fear
23mImplications of the Realtime Web