What is Database?
A database is a structured system for persisting, querying, and managing data. It is the backbone of practically every web application, CRM, e-commerce platform, or SaaS product — no database means no user accounts, no orders, no analytics. Two primary classes exist: Relational databases (SQL) store data in tables with fixed schemas and relationships — PostgreSQL, MySQL/MariaDB, SQL Server, and Oracle dominate this space. They guarantee ACID properties (Atomicity, Consistency, Isolation, Durability) and are the right choice for transactional workloads like accounting, banking, or shop systems. NoSQL databases drop fixed schema and relational joins in favor of flexibility and horizontal scaling — MongoDB (documents), Redis (key-value, in-memory), DynamoDB (distributed), Neo4j (graph), and Elasticsearch (full-text) are prominent representatives. The most important design decisions: Normalization (avoid data duplication, secure integrity via foreign keys) vs. denormalization (avoid joins, prioritize read performance). Indexing (fast queries via B-tree or hash indexes, cost: slower writes + storage). Backups (point-in-time recovery, automated snapshots — without backup, the question is not if but when data loss happens). Migrations (version-controlled schema changes, never manually in production). For EU companies, GDPR compliance is central: data location, deletability of personal data, encryption at rest and in transit. Modern cloud offerings (Supabase, Neon, PlanetScale, AWS RDS) provide managed databases with backup, replication, monitoring, and regional data residency — often the better choice than self-hosting for teams under 50 people. At goldenwing.at, for example, the CMS runs on PostgreSQL 16 locally on a VPS with daily snapshots.
Key Points
- Two worlds: relational DBs (PostgreSQL, MySQL) for transactional consistency vs. NoSQL (MongoDB, Redis) for flexibility and scaling
- ACID properties guarantee that transactions are atomic, consistent, isolated, and durable — indispensable for finance, orders, and accounts
- Indexing makes queries orders of magnitude faster — but every index costs on writes; choose strategically
- Normalization vs. denormalization: integrity vs. read performance — find the right balance per workload
- Backups are mandatory, not optional: point-in-time recovery, automated snapshots, regular restore drills
- Ship schema migrations versioned (e.g., via Drizzle, Prisma, Flyway) — manual DB changes in production are a recipe for drift and chaos
- GDPR aspects: document data location, make personal data deletable, encrypt at rest and in transit
- Managed services (Supabase, Neon, RDS) save ops effort — for small teams almost always the better choice than self-hosting
- Connection pooling (PgBouncer, Prisma Accelerate) stops being a luxury around ~50 concurrent users and becomes a necessity
Practical Example
“Our Payload CMS uses PostgreSQL 16 with localized translations in locale tables, foreign-key integrity, and daily pg_dump snapshots for point-in-time recovery.”
Need professional help with Database?
Our GoldenWing team offers strategic Web & App Development services for Austrian and international clients. From initial consulting to implementation — we deliver measurable results.