Please enable JavaScript.

Coggle requires JavaScript to display documents.

(Scalability & System Design, Distributed Systems, Lead-Level System…

- - - - Cannot guarantee all three
- - - - Fixed intervals
    - - Overlapping intervals
    - - Based on user activity
- - - - Payment system
      - Notification system
      - File storage system
    - - Must justify every decision
- - - - Controller → Service → Repository → Entity
      - DEF: Logical separation in same deployment
    - - Each layer deployed independently
    - - Too many tiers → latency
    - - Max 3–4 tiers
  - - - Financial systems, audit logs
    - - Full history
      - Replay capability
    - - Hard to query current state
  - - - Scalability
      - Team autonomy
    - - Complexity + network failures
- - - - Financial systems
    - - Hard to scale horizontally
      - Schema rigidity
  - - - Redis, DynamoDB
      - Use case
        
        Caching, sessions
    - - MongoDB, Cassandra
      - Flexible schema
    - - Neo4j, Amazon Neptune
      - Relationships-heavy data
    - - Eventual consistency issues
  - - - Amazon S3
      - GCP Cloud Storage
      - Azure Blob
    - - Store large immutable objects
    - - Media, backups
    - - Not for low-latency access
  - - - Google Drive
- - - - Payments (avoid double charge)
      - Retry mechanisms in distributed systems
    - - Idempotency keys (UUID per request)
      - Store request result in DB/cache
    - - Missing idempotency → duplicate operations
    - - Always required in financial / critical APIs
  - - - gRPC (modern, binary, fast)
      - REST (HTTP-based)
    - - Microservice communication
    - - Tight coupling between services
    - - REST (flexible) vs gRPC (performance)
  - - - Non-blocking calls
    - - High throughput APIs
    - - Hard debugging + tracing