Best of Both Worlds: Building a Mirror-trading Platform on Top of ...
summary
Builds a mirror‑trading platform that actually runs: FIX connectivity via QuickFixJ, RabbitMQ for message passing, Redis cluster for high‑frequency quotes and MySQL with the core as sole writer. JRuby handles thousands of messages per second in plain Ruby threads, ZooKeeper performs leader election, and the Tanuki service wrapper plus JMX monitoring detect and recover JVM stalls. Shows pragmatic tradeoffs (use MRI for some rake tasks, Passenger threaded frontends), a few JRuby quirks, and overall a working, scalable system.
AI-written summary · about
chapters (15)
- 0:00Introduction
- 1:50What this talk is about
- 3:13Latency
- 3:52Magic Box
- 8:39Why JRuby
- 9:59Out of Memory
- 11:25Robustness
- 12:10Tanuki
- 13:47Service Rep igloo
- 15:08Service Rep teardown
- 15:34JVM Monitoring
- 16:48JConsole
- 17:32Passenger
- 18:17Issues
- 20:56Conclusion
related talks
18mRuby Messaging Patterns
27mRuby Messaging Patterns
30mRack based web apps with JRuby and TorqueBox
46mJRuby on Rails - RoR's Simplicity Meets Java's Class
39mJRuby: Insights from Six Years in Production
28mDeploy, Scale and Sleep at Night with JRuby
44mWhy JRuby?
30mSome Funny Things Happened on The Way to A Service Ecosystem