Real-time Trade Data Processing and Machine Learning at Giga-Sca...
summary
Claim is that Hazelcast runs real-time trade pipelines and ML on a single runtime and modest hardware, reaching billions of events per second in benchmarks and delivering sub-millisecond reads and writes. Demo shows a Kafka trade stream ingested into a five-node cluster keeping running totals for ~3,000 NASDAQ tickers, joined by SQL to in‑memory reference data and pushed into FINOS Perspective for live charts and Slack alerts; a real query hits ~30 ms 99th percentile in the demo so the system actually works.
AI-written summary · about
chapters (8)
- 0:00Start
- 2:44Hazelcast Vision
- 3:56Real-time in Financial Services
- 6:08Why Does Real-time Matter?
- 9:23Anti-Pattern for Real-time - "Delta" Architecture
- 10:32Architecture - Scale Ingest/Data/Compute/ML - Together
- 16:34Deployment Architecture vs Traditional
- 23:02Hazelcast Continuous Query with Drilldown Process Flow
related talks
30mFive Challenges in Real-time Stream Processing and Five Solutions
47mReal-time Fraud Detection Challenges and Solutions
51mReal-time Fraud Detection: Challenges and Solutions
19mSupercharging Kafka Applications with Real-time Contextual Insights
08mReal-Time Data Systems: Speed, Scale, and Resilience
49mEnriching Kafka Applications With Real-time Contextual Data
27mBeyond Logs: Real Time Anomaly Detection with Machine Learning
37mBig Data Analytics, Machine Learning, Real Time Processing