Strategy

• Identify the most frequently accessed data

• Identity the data that requires low latency access

• Identify the data that are computationally expensive to generate

• Deploy cache where data like above is accessed

• Determine the time to live (TTL) for cached data

• Choose the right cache type

  • Application in-memory cache

    Local to instance, fast but not consistent across instances

    Caffeine, Cache2k, etc.

  • Distributed cache

    Shared across instances ensuring consistency

    Redis, Memcached, etc.

  • CDN / Edge Cache

    For static assets, closer to users