Overview

Cassandra Architecture

• Distribution Model - Dynamo
• Storage Model -BigTable
• Peer-to-peer
• Column-oriented
• Commitlog
• Partition
• Replication strategy

Write Path

Read Path

Data Model

Storage Model

Column-oriented

• Map<RowKey, SOrtedMap<Column, ColumnValue>>
• RowKey: unique
• ColumnKey
• Schemaless

Logical Data Structure

Column: Name, Value, Timestamp, TTL
Row: Row Key, several columns Column Family: contains serveral Rows Keyspace: serveral column families, Super column family

Caveats for Relational Developers

• No normalization
• Focus on query
• Physical topology of cluster
• Collections (set, list, map)
• No foreign keys, no referential integrity, no joins
• Denormalization
• No sequence
• Secondary index
• Tradeoff in ACID