NoSQL Databases You Have to Know
summary
Concrete comparisons of Hazelcast, MongoDB, HBase and Cassandra, with the actual tradeoffs and knobs: Hazelcast is an in‑memory data grid using consistent hashing for automatic sharding, configurable backup count and an entry processor to run logic on the node; MongoDB stores JSON‑style docs (practical max ~60MB), primary/secondary replication and tunable readPreference; HBase is a column‑family store with region servers and automatic sharding used for Facebook messaging; Cassandra combines BigTable+Dynamo, offers linear scale, a replication factor and tunable consistency levels.
AI-written summary · about
chapters (10)
- 0:00Introduction and context
- 1:33RDBMS and web-scale
- 5:31Scaling and distributed systems
- 9:44CAP theorem and trade-offs
- 13:45Real-world scalability lessons
- 17:03Hazelcast key-value store
- 30:15MongoDB document database
- 38:48Apache HBase
- 45:01Apache Cassandra
- 48:37Conclusion
related talks
01h 01mNoSQL: An Analysis
34mData Storage: NoSQL Toasters and a Cloud of Kitchen Sinks
39mFramewars: the battle between NoSQL and Java in the cloud arena
29mCassandra - an introduction
47mNoSQL Endgame.
05mStartup Showcase: Hazelcast
38mNoSQL Databases: Its not a free lunch
45mEasy Scaling with Open Source Data Structures