Building Planet-Scale Payment Platforms
summary
The session names specific targets and tradeoffs: handling 150 billion card transactions annually, absorbing holiday peaks of 65,000 transactions per second, and avoiding $31,000 lost per second downtime. It spells out concrete latencies—100–200 ms authorization and 10–20 ms fraud checks—plus tools and patterns used: horizontal scaling, service‑oriented isolation, distributed caching and cache‑warming, HSMs and tokenization, model serving with hot swaps, continuous reconciliation, adaptive load balancing and chaos engineering, and orchestration for settlement and cross‑border FX hedging.
AI-written summary · about
chapters (14)
- 0:00Introduction and Speaker Background
- 0:17Understanding Global Card Payment Ecosystems
- 0:45Challenges in Designing Payment Platforms
- 1:09Scale and Complexity in Payment Platforms
- 2:43Transaction Processing and Fraud Checks
- 4:31Distributed System Architecture for Payments
- 7:15Load Balancing and Traffic Management
- 8:55Fraud Detection and Risk Management
- 10:55Settlement and Clearing Processes
- 11:57Cross-Border Payment Complexities
- 13:53Security in Payment Platforms
- 17:22Observability and Monitoring at Scale
- 19:56Future Directions and Emerging Challenges
- 21:34Conclusion and Final Thoughts
related talks
51mThe Tie That Binds: Unique Cyber Tools in the Payments Industry
38mPerformance Engineering At MasterCard
30mScaling crypto payments for gaming volumes
49mPayment System: Survival Guide
32mEnhancing Security, Reliability, and Global Scalability in FinTech
54mPayment System Architecture: Almost Enterprise
22mBlack Friday Brilliance: Managing a Billion Transactions with Tech, and Teamwork
35mCustomer Panel