Chapter 5 Data Storage

5.1 Theory

https://www.percona.com/live/17/sites/default/files/slides/Heaps%20B-trees%20log%20structured%20merge%20trees%202017-04-25.pptx_.pdf

Heaps – Good for low write rate, high read rate
Catalogs, Wikis, Content management systems

B-Trees – Works for most everything, Good for high concurrency transactional workloads
Bank accounts, inventory, Payroll processing, Stock trading, etc

Log structured Merge Trees (LSM) – Good for high write rate with low to moderate read rates
IOT, Time Series, Customer Service logs, etc.

Fractal Trees – Better write latency than traditional B-Trees, worse sequential operations
Most of the same places you would want to use B-Trees but heavier write work loads

Column stores – Not good for concurrent operations, Store data in columns not rows
Use for BI and Data Warehouses

5.1.1 BTree

5.1.2 LSM-Tree