What is an In-Memory Database

Every computer has two types of storage:

Analogy: RAM is your desk — fast access, limited space. Disk is your filing cabinet — slow access, unlimited space.

Redis stores all data in RAM. This means:

• No waiting for disk read/write operations
• Data access is in nanoseconds to microseconds
• It is 100x to 1000x faster than disk-based databases

flowchart LR
    subgraph RAM[RAM — Memory
      ~100 nanoseconds
      Volatile (lost on restart)]
        R1["📝 Redis<br/>1M ops/sec
        < 0.1ms per lookup"]
    end

    subgraph Disk[Disk — SSD/HDD
      ~100 microseconds
      Persistent (survives restart)]
        D1["🗄️ MySQL / PostgreSQL<br/>~5K ops/sec
        ~1-5ms per lookup"]
    end

    RAM -->|100x - 1000x faster ⚡| Speed[Redis wins on speed]
    Disk -->|Safe storage ✅| Safety[DB wins on durability]

    style RAM fill:#7c3aed,color:#fff
    style Disk fill:#3b82f6,color:#fff
    style R1 fill:#7c3aed,color:#fff
    style D1 fill:#3b82f6,color:#fff
    style Speed fill:#059669,color:#fff
    style Safety fill:#f59e0b,color:#fff

Analogy: RAM is your desk — fast access, limited space. Disk is your filing cabinet — slow access, unlimited space.

Since Redis stores data in RAM, if the server crashes, the data is lost. Redis solves this with persistence mechanisms (RDB snapshots and AOF logs) discussed in Section 12.

Tip: For use cases like session caching or rate limiting, losing data on restart is acceptable. For critical data, enable persistence or use Redis alongside a durable database.