Deep dive into data streaming security
summary
Real incidents include a health provider leaking 14 million messages via an unauthenticated Kafka broker and a food delivery cluster exposing ~20,000 customer records over one hour after years of misconfiguration, illustrating worst bets. The breakdowns are concrete: missing TLS, unpatched CVEs like CVE-2019-12399, leaked credentials, wild‑card ACLs, and unsafe connectors. Recommended controls are practical: enforce topic-level encryption or field encryption for PII, use ACLs without wildcards, rotate keys, patch dependencies fast, segment networks with private link, rate‑limit producers, and audit activity.
AI-written summary · about
chapters (15)
- 0:00Data streaming security intro
- 2:28Security bloopers awards
- 9:01Stream defenders checklist
- 10:27Payment fraud detection use case
- 15:21Apache Kafka explained in two minutes
- 20:45Kafka Flink Iceberg overview
- 24:04Security principles for streams
- 24:46Encryption for streaming security
- 32:14Access control in streaming
- 37:38Network segmentation for streams
- 39:27Data exfiltration risks
- 43:21DoS threats in streaming
- 46:21AI/LLM security challenges
- 50:40Situational awareness
- 54:30Final checklist
related talks
45mDeep Dive into Data Streaming Security
43mAttacking (and Defending) Apache Kafka
33mUnderstanding the Threats and Attacks on Data Science Applications and Models
32mDo This, Not That – Lessons from 7 Headline Grabbing Security Breaches
46mApplication security from start to finish
40mKeeping data private in real-time pipelines
59mInside the Breach: Secrets Exposed
57mLessons From Billions of Breached Records