Lecture 28: Wide Tables: HBase and Cassandra

Content locked

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

Overview

Learning Objectives

The Hadoop Ecosystem

Architectural Layering

HBase

Partitioning the Row Space

Single-Row Transactions

HBase Data Model: Versioned Sparse Tables

Fault Tolerance: RegionServer Failover

Storage Layout: Memory + HDFS

Writes Go to the Memstore

Sort and Flush

Subsequent Writes Start a New Memstore

HBase Reads

Compaction

Cassandra Data Model

Clusters and Rings

Keyspaces

HBase: “Wide Row” Design

Cassandra: “Wide Partition” Design

Advantages

Disadvantages

Column Roles: Partition Keys, Cluster Columns, Static Columns

Schema Example: Weather Data

Advantages of This Schema

Challenges

Conclusion

Source