From SQL to Cassandra
summary
VK migrated parts of its data layer from SQL to Apache Cassandra and explains why simply copying tables and swapping SQL for CQL doesn’t work. The talk shows real-world schema and query design practices, describing per-table schema changes driven by business queries and Cassandra implementation details to meet tight latency requirements (99th percentile < 10 ms). It targets engineers already using Cassandra and those planning a migration.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 1:45Cassandra experience
- 3:36Cassandra pros and cons
- 5:01Primary key and partition key
- 7:50Secondary indexes and limits
- 10:38Query limitations in Cassandra
- 11:42SQL table migration example
- 17:12Data consistency maintenance
- 23:41Table design summary
- 24:49OLAP vs OLTP queries
- 26:17CDC data sync
- 28:14Key-based query optimization
- 30:50Phone number prefix search
- 34:51Data deletion and tombstones
- 38:01Alternatives to updates
- 40:12Cassandra results and benefits
- 41:15Q&A
related talks
35mThe What and Why of Cassandra Datamodelling at Scale
45mCassandra Architecture: Understanding, Scaling and Optimizing
29mCassandra - an introduction
35mA Customer’s Journey: Migrating from NoSQL to a Cloud Native Distributed SQL Database
24mMigrations under Production Load: How to Switch...
32mMigrating Applications Between Cassandra Clusters with No Downtime
40mLatency, Throughput & Fault Tolerance: Building a Modern Database
40mApache Cassandra in High-Performance Systems Development