Consistency Levels, Replication Factor, and the Write and Read Path
Learn the distributed consistency model that makes Cassandra powerful and why these settings matter deeply in real deployments.
Inside this chapter
- Replication Factor and Availability
- Consistency Levels
- Read and Write Path Awareness
- Tradeoffs in Practice
Series navigation
Study the chapters in order for the clearest path from beginner Cassandra concepts to advanced distributed operations. Use the navigation at the bottom of each page to move through the full series.
Replication Factor and Availability
Cassandra replicates data across nodes. The replication factor determines how many copies are stored. This is central to availability, durability, and read/write behavior across failures.
Consistency Levels
| Level | Meaning | Typical Use |
|---|---|---|
ONE | One replica response is enough | Low latency, less strict reads or writes |
QUORUM | Majority of replicas must respond | Balanced correctness and availability |
ALL | All replicas must respond | Strict but less available under failure |
LOCAL_QUORUM | Majority within local data center | Common multi-data-center pattern |
Read and Write Path Awareness
Advanced learners should understand that Cassandra writes involve commit log and memtable activity, while reads may require coordination across replicas depending on consistency choices and data freshness. This is very different from a simple single-node mental model.
Tradeoffs in Practice
Cassandra lets teams choose consistency levels based on workload needs. That flexibility is powerful, but it also means engineers must understand the tradeoff between latency, availability, and read correctness instead of assuming one universal default is always right.