archinario.
Free course3 modules · 9 lessons

System design foundations

Learn to distribute load, handle failures and scale systems. Build diagrams and test solutions on working models.

System design means designing a system around load, data and failures. Change components and connections in the course, then compare throughput, latency and errors.

Why queues grow and when to add workers

  • Find bottlenecks using load, latency and errors.
  • Choose between more capacity, a cache and a queue.
  • Evaluate trade-offs between resilience, data freshness and cost.

Your progress

Checking progress…

Tests passed: Checking progress…

Open the first lesson

Load balancing

Progress is stored in this browser. All lessons are open: follow the sequence or choose a topic.

Course syllabus

9 practical lessons
Module 01

Load and bottlenecks

Distribute requests and let the system handle the work.

  1. 01

    Load balancing

    Why can't one server cope?

    Available
    Check your understanding
  2. 02

    Caching

    Why did the database become a bottleneck?

    Available
    Check your understanding
  3. 03

    Queues and background tasks

    How do you survive a task burst?

    Available
    Check your understanding
Module 02

Fault tolerance

Learn when to retry requests, disconnect a dependency and limit flow.

  1. 04

    Timeouts and retries

    Why do retries amplify failures?

    Available
    Check your understanding
  2. 05

    Circuit breaker

    How do you stop a cascading failure?

    Available
    Check your understanding
  3. 06

    Rate limiting

    Where should excess requests be rejected?

    Available
    Check your understanding
Module 03

Data and scaling

Distribute data and size capacity for changing load.

  1. 07

    Database replication

    How can database load drop while data stays fresh?

    Available
    Check your understanding
  2. 08

    Sharding and load imbalance

    Why don't new shards save a hot key?

    Available
    Check your understanding
  3. 09

    Autoscaling

    Why does autoscaling respond late?

    Available
    Check your understanding