Cassandra Cheatsheet

• Wide Column Store Database
• Cluster Key
• Sort Key

It's recommended that read and write of the same data go to the same partition

Leaderless Replication

• Using a quorum (can be configured), can even use 1 which will read and write from 1 partition

Read Repair Anti Entropy (merkle tree)

Write Conflicts?

• Last write wins, can lead to lost writes

• as opposed to Riak uses CRDTs to resolve write conflicts

Single Node

• Memtable + SSTables. Fast write and slow read

• only row level locking, no ACID transaction