Building mission-critical applications with CockroachDB
summary
CockroachDB is a distributed SQL database that provides serializable consistency across global clusters by splitting one sorted keyspace into automated raft-backed ranges (default 512 MB ranges) with three replicas per range. Demoed behaviors: adding nodes rebalances thousands of replicas in seconds to minutes, lease transfer completes in ~3s after a node crash, full replica recreation after a permanent failure completes on the order of an hour, and P99 latency blips last ~3s; supports follower reads and regional-by-row sharding.
AI-written summary · about
chapters (16)
- 0:00Introduction to CockroachDB
- 1:48Evolution of databases
- 3:44What is Cockroach Database
- 7:57Scalability in CockroachDB
- 12:09Raft consensus groups
- 16:05Live demo of node addition
- 20:21Resilience to failure
- 25:16Demo of node failure
- 27:49Data locality and latency
- 32:50Regional by row data
- 33:31Global tables and follower reads
- 37:44Serializable isolation levels
- 40:41Follower reads and time travel
- 42:28CAP Theorem and availability
- 44:36Mission-critical applications
- 45:19Deployment options
related talks
31mFuture-proofing operational data at scale
52mCockroachDB
01h 11mFearless global transactions with
22mBuilding CockroachDB for performance at scale
42mThe hows and whys of a distributed SQL database
14mBuilding CockroachDB for performance at scale
24mNavigating Distributed Databases: Challenges & Opportunities
14mHow CockroachDB can enable Multi-Region Deployments and Data Locality