Daily Tech News Show What Caused A Third of the Internet to Go Down - DTNS 5128
29 snips
Oct 20, 2025 A major outage at AWS left countless sites offline, revealing the vulnerability of DNS and load-balancers. The hosts discuss how this reliance leads to recurring issues in the US East-1 region. They also speculate on whether the outages stemmed from code errors or coordinated attacks. Meanwhile, Samsung’s promising Exynos 2600 chip is highlighted, along with its strategy for a unified flagship chip. Other topics include intriguing sales data on the iPhone 17 and X's new link handling features, plus Alibaba's advancements in GPU technology.
AI Snips
Chapters
Transcript
Episode notes
DNS Failures Can Disable Large Parts Of The Internet
- A DNS or routing failure can make the internet feel down even when data is fine.
- AWS US-East-1 routing/load-balancer issues can cascade to many services and continents.
Load Balancers Amplify Small Failures
- Load balancer health checks and routing propagation failures amplify outages.
- Mis-marked healthy nodes or failed routing updates can cause cascading service degradation.
Design For Cascades And Isolate Regions
- Expect cascading effects and prepare for knock-on failures during cloud incidents.
- Design for isolation and avoid single-region dependencies to reduce blast radius.