Keynote | Flink First: Designing Modern Trading Systems
summary
A Flink-first trading platform replaces a decades-old patchwork of tools with a unified stream backbone so every component is a Flink job, handling roughly 8 million CUSIPs and millions of real‑time ticks from multiple venues. The prototype becomes production-grade and is built in 18 months, integrates legacy services, enforces architectural governance, reduces redundant logging, and improves fault tolerance and scaling; tradeoffs include memory limits for full-stream state and choosing keyed pub/sub patterns versus product-type silos.
AI-written summary · about
chapters (10)
- 0:00Introduction and keynote
- 1:30Legacy system challenges
- 6:55Consumer tech meets finance
- 13:37Tool proliferation pitfalls
- 17:27Trading system design dimensions
- 23:19Flink as a foundational shift
- 28:16System development and benefits
- 31:03Indicative data solutions
- 33:48Handling back pressure
- 35:57Acknowledgments and conclusion
related talks
29mStreaming & Eventing for Top Management
32mFlink's SQL Engine: Let's open the engine room!
23mAdvancing Data Mesh: Building A Stream Processing EcoSystem of Reusable Processors and Datasets
34mCustom, Complex Windows at Scale using Apache Flink
43mStructure and Interpretation of Stream Processing
37mNot So Big – Flink as a true Application Framework
47mEvent Driven UIs
31mWriting an interactive streaming SQL engine and pre-parser using Flink