Skip to content

tom_notes.md

Latest commit

 

History

History
39 lines (31 loc) · 1.44 KB

tom_notes.md

File metadata and controls

39 lines (31 loc) · 1.44 KB

NoSQL

Name typically explained as "Not Only SQL" or "Non-relational"

Relational Databases:

  • Great with data consistency
  • Hard to scale
  • Resource intensive

Basically, at some point it won't be able to handle the load, can scale vertically but not horizontally.

NoSQL Databases

  • Can scale vertically and horizontally.
    • Vertically would be like adding space by adding stories to an existing building.
    • Horizontally can be thought of as building new buildings.
  • Do away with relationships
  • Key/Value stores, basically. Every item stands on it's own.
  • Partitions are different servers (horizontal).
  • Primary Key is converted to Hash value within the keyspace.
    • Keyspace ->
      • Each partition (server) is assigned a range within the keyspace range
      • How it knows where to store the data, and where to find the data.
  • Can only be retrieved by the primary key.
  • Schema-less
  • Partitions are mirrored to other servers to prevent loss from hardware failure
    • May cause data to not be immediately returned.
    • Eventually consistent, it will be returned after milliseconds.

Who's using them?

  • Cloud providers
    • For scalability
    • DynamoDB, BigTable, CosmosDB
      • Prime Day 2019 -> Amazon's NoSQL database peaked at 45 million requests per second.
      • Apple -> 75,000+ servers (largest known) NoSQL database.
  • Self-hosted
    • Cassandra, Scylla, CouchDB, MongoDB