Course contentsShow
← System Design
Lesson 1400 of 1,91935. Backup and Disaster RecoveryFree lesson

What Are Backups and Why They Matter

Understanding backups as point-in-time copies of data for recovery from loss, corruption, or disasters.

What Are Backups and Why They Matter

You'll learn: Why backups are essential safety nets in system design and what they actually protect against.

The Core Idea

A backup is a point-in-time copy of your data stored separately from the primary system. Think of it like taking a photograph of your entire data state at a specific moment — if something goes wrong later, you can restore that frozen-in-time version.

Why Backups Matter

Despite all the replication, redundancy, and fault tolerance we've discussed (active-active systems, multi-region deployments, conflict resolution), backups serve a fundamentally different purpose: protecting against data loss scenarios that replication cannot solve.

Consider these situations:

1. Data Corruption or Bugs
If a software bug corrupts data and replicates that corruption across all your replicas, replication won't save you. A backup from before the corruption occurred will.

2. Accidental Deletion
A developer accidentally drops a production table. Replicas faithfully replicate the deletion. Only backups let you recover.

3. Security Breaches
Ransomware encrypts your database. Synchronized replicas are encrypted too. Clean backups become your lifeline.

4. Compliance and Legal Requirements
Many industries require maintaining historical data copies for auditing, even if the live system evolves.

Backups vs Replication

Replication provides high availability — keeping your system running during failures. Backups provide recoverability — getting back lost or damaged data. You need both. Replication handles hardware failures and network partitions; backups handle logic errors, malicious actions, and disasters that affect all replicas simultaneously.

Key Takeaway: Backups are point-in-time snapshots that protect against data loss from corruption, deletion, or disasters — scenarios where replication alone cannot help. They're your "undo button" for catastrophic data events.