Overview of Database Types

Overview: This guide covers 9 popular database types (note: DB solutions can cover multiple categories, and the following is not an exhaustive list of database solutions and how many categories they overlap in):

Table of Contents:

1. Popular Database Types
2. Database Use Cases
3. Scaling Features

---

1. Popular Database Types

• Columnar Databases
  • Examples Include: Amazon Redshift, Apache Cassandra, Apache HBase, ClickHouse, Google Bigtable.
• Document-Based Databases
  • Examples Include: Amazon DocumentDB, CouchDB, MongoDB, RavenDB, RethinkDB.
• Graph-Based Databases
  • Examples Include: Amazon Neptune, ArangoDB, Azure Cosmos DB, Neo4j, OrientDB.
• In-Memory Databases
  • Examples Include: Memcached, Microsoft SQL Server In-Memory (OLTP), Redis, SAP HANA, VoltDB.
• NoSQL Databases
  • Examples Include: Apache Cassandra, Apache HBase, Couchbase, MongoDB, Redis.
• Object-Oriented Databases
  • Examples Include: db4o, GemStone/S, ObjectDB, Versant Object Database, ZODB.
• Relational Databases
  • Examples Include: CockroachDB, Microsoft SQL Server, MySQL, Oracle (various products), PostgreSQL, SQLite.
• Time-Oriented Databases
  • Examples Include: Druid, InfluxDB, OpenTSDB, Prometheus, TimescaleDB.
• Vector-Based Databases
  • Examples Include: Chroma, Milvus, Pinecone, Qdrant, Weaviate.

---

2. Database Use Cases

Before diving into specific use cases for each type of database, several factors should be considered. These include: cost of operations (can the organization cover it, and is there a more efficacious alternative?), employee/team expertise (will it take significant training to implement the database solution?), whether it makes sense to host services in-house or via a cloud provider, whether scalability is a significant concern (such as it is for streaming services and some embedded/IoT environments), and if one's organization is prepared to back up and recover data, or install and maintain backup/recovery solutions corresponding to the selected database solution.

Use cases for these nine database types include...

Database Type	Use Cases
Columnar	Business intelligence (BI), data warehousing, significant analytics and enterprise-level reporting needs.
Document Based	Content management, JSON and JSON-like data, fast development deadlines, need for flexibile schematics.
Graph Based	Social bookmarking, networking, and recommendation-based sites, interconnected data needs, pattern recognition.
In Memory	Need for instant/real-time performance and processing, data caching, and temporary storage of data.
NoSQL	Quick response times/low latency, instant/real-time analytics, constantly-evolving data, server scalability needs.
Object Oriented	Integration with object-oriented languages (e.g., Java), hierarchy-based data objects, long-term object storage/persistence.
Relational	ACID compliance (important for large enterprises and essential services), data integrity, sophisticated joins/transactions of multi-table data, structured data.
Time Oriented	Time-series analytics, monitoring and metrics, sensor-related and IoT/embedded system data, data sequencing/patterns.
Vector Based	Quick retrieval, machine learning (including complex tasks), similarity searching.

---

3. Scaling Features

Four of the unique ways that databases scale (and can handle increasing traffic/data requests) include...

Feature	Explanation
Indexing	This is the use of 'pointers', in which keys in memory (indexes, or indices) are mapped to rows/records in a table.
Caching	By searching for data in a memory cache, performance of reads can be improved (no need to request data from a database, preserving time and memory).
Sharding	This splits rows/records and columns/fields across multiple servers/devices, potentially providing load balancing.
Replications	This also provides load balancing by replicating data (for reading purposes) across multiple servers/devices. Write responsibilities can still be allocated to one database (the 'primary' database) to preserve data integrity.