This post is part of a series of posts with notes as I’m studying for Google’s Professional Data Engineer Certification.
This particular post covers Spanner in a nutshell.
Please read this disclaimer.
- Horizontally scalable
- Only change needed is to add nodes
- Add another node if CPU > 70-75%
- Each node can handle 1.5 - 2 TB
- Can create linked/embeded tables
- Helps avoid the cost of joins for normalized tables
- Two payment methods:
- Pay-by-node: standard method
- Pay-by-processing-unit: let’s you pay for more granular operations/scale
- 1 node = 1,000 processing units