Database

SQL / RDBM

• standardized languages and grammar
• can be scaled mainly vertically by increasing power of existing hardware
• data is relational and tabular/row-oriented
• ACID-conform

Databases

H2

• Local Database
• Often used for development purposes
• Completely in memory (?) but can be persisted in a file

SQLite

• No classical Client-/Serer-Architecture
  • Every client has the whole database and functions available
• tiny size
• wide spread among embedded and mobile systems
• lacking some feataure (e.g. for complex queries)
• lacking some security
• single file approach has lot of benefits but prevents multi user
• optimal for client side software - no additional installation needed because it's embedded in your application and the single file approach is easily backupable and movable

MariaDB

• Real OpenSource MySQL successor
• extremely close to MySQL Standard
• rapid Security Patches
• can become bloated
• has some caching issues
• offers In-Memory Databases

PostgreSQL

• extremely performant on huge datasets
  • not the best on small datasets
• database functions can be written in Python, Perl, Java, Ruby, C and R
• huge community support
• no In-Memory-Databases but very efficient caches

MySQL

• since it was bought from Oracle it is kinda discontinued

Migration

Liquibase

• Web

Flyway

• Web

NoSQL

• diverse languages and grammar
• can be scaled horizontally with hirarchical node structure
• CAP-conform

Document oriented

• great for dynamic data
• data mostly stored in JSON

MongoDB

• Mongo DB in Kotlin
• regarding CAP favors consistency and partition tolerance over availability
• has Atlas managed hosting with a Free Tier and has manual backup and restore

Google Firestore

• Used until Q1 2023 for Jarvis AI
• Not really available for Java/Kotlin Serverside ==> really made for direct access from fronted clients
  • that is fine for data that is just that - data - but not so great for data that we need to centrally react on
  • this could in term be done by firestore functions, but dependencies over dependecies over dependencies.. :c

CouchDB

• regarding CAP favors availability and partition tolerance over consistency

Couchbase

Column oriented

• often used for data analysis
• e.g. Apache Cassandra

Key-Value store

• fast and scalable
• not very customizable
• Couchbase Server
• Redis

Graph databases

• Neo4J
• Oracle NoSQL

References

• Difference Between NoSQL and MongoDB (Lithmee)