Future-proofing operational data at scale
summary
CockroachDB demonstrates concrete resilience and speed: a public performance under adversity benchmark injects disk stalls, asymmetric network partitions and region failures and shows autonomous failover, zero RPO and smoothing via leader leases and write‑ahead log failover. A customer survey of ~1,000 executives reports 86 outages/year and 324 average minutes downtime. Releases add measurable gains: 25.2 is 50% faster than 24.3, and a new distributed vector index (C‑SPAN) supports billions of vectors with immediate freshness, SQL integration, plus migration (molt fetch from Oracle), improved Kubernetes operator and AD/LDAP.
AI-written summary · about
chapters (19)
- 0:00Welcome
- 1:57Keynote and insights
- 3:01Resilience as design
- 4:25High availability replication
- 5:50Fault tolerance edge cases
- 7:14Disaster recovery defense
- 8:17Outage costs and frequency
- 10:44Benchmarking chaos resilience
- 13:08Internal stress testing
- 14:10Gray failures storage stalls
- 16:17Network node simulations
- 17:20Zone region recovery
- 18:46Six-month performance
- 20:10Benchmark vs Oracle GDD
- 22:40Performance and hardware
- 23:43AI agents scalability
- 26:10Vector search embeddings
- 27:55Distributed vector indexes
- 30:03CockroachDB 25.3 enhancements
related talks
48mBuilding mission-critical applications with CockroachDB
22mBuilding CockroachDB for performance at scale
16mStraight from the Field
27mBeyond the Roadmap and Into the Future of CockroachDB
29mHow City Storage Systems designed its Data Infrastructure to minimize toil
24mNavigating Distributed Databases: Challenges & Opportunities
52mCockroachDB
01h 11mFearless global transactions with