Lecture 30: Cassandra Partitioning

Content locked

The catalog is public. Reading this content requires approved access.

Overview

Learning Objectives

Cassandra Background

Partitioning Approaches

Mapping Data Structure

Hash Partitioning

Consistent Hashing

Scalability Concerns

Elasticity and Incremental Scalability

Consistent Hashing

The Number Line

Step 1: Place Workers on the Line

Step 2: Place Rows on the Line

Step 3: Map Rows to Workers via Ranges

Wrapping Range

Token Rings

Adding a Node

Cost of Scaling

Collisions

Virtual Nodes (vnodes)

Heterogeneity

Token Map Storage

Every Node Holds a Copy

Bad Approach: Update Everyone Synchronously

Better Approach: Gossip

Coordinators

Conclusion

Source