System Design Template

1. Features

   • Use cases
   • Features we don’t care about
   • Who will use this service?
   • How many people will use this service?
   • How will this service be used?

2. Estimations

   • Throughput (QPS for R/W)
   • Maximum tolerated latency for (R/W)
   • R/W ratio
   • Traffic estimate
     • Write (Memory usage)
     • Read (Memory usage)
   • Storage estimates
   • Memory estimate
     • Cache usage
     • RAM usage
     • Disk usage

3. Design Goals

   • Latency and Throughput requirements
   • Consistency vs Availablility (Weak/strong/eventual consistency | failover/replication availability)

4. High Level Design

   • APIs for Reads and writes
   • Database Schemas
   • Algorithms
   • High level design for R/Ws

5. Deep Dive

   • Scaling Algorithm
   • Scaling components
     • Availability and Consistency scaling for each component

• Components:
  • DNS
  • CDN (Push vs Pull)
  • Load Balancers (Active-Passive, Active-Active, Layer 4, Layer 7)
  • Reverse Proxy
  • Application layer scaling (Services)
  • DB (RDBMS, NoSQL)
    • RDBMS:
      • Leader-Follower, Leader-Leader, Federation, Sharding, Denormalization, Indexing, Tuning
    • NoSQL:
      • Key-Value, Wide-Column, Graph, Document
      • Fast Lookups:
        • in RAM: (Bounded size): Redis, Memcached
        • Available: (on Disk): Cassandra
        • Consistent: DynamoDB
  • Caching:
    • Client caching, CDN caching, Webserver caching, Database caching, Application caching, Cache @Query level, Cache @Object level
    • Eviction policies: - Cache aside - Write through - Write behind - Refresh ahead
  • Async
    • Message Queues
    • Back pressure
    • Task Queues
  • Communication Protocols
    • TCP
    • UDP
    • REST
    • RPC

6. Justification
   • Throughput of each layer
   • Latency caused by each layer
   • Overall latency justification