roadmap.sh: System Design

What it is

roadmap.sh is a community driven site of visual learning roadmaps, and this is its system design roadmap. Each topic appears as a node in a dependency map. Clicking a node opens a short explanation and links to free articles and videos. With a free account you can mark nodes as done, in progress or skipped and track progress across devices.

The project is open source: its main GitHub repository has about 368,000 stars (September 2026), and roadmap.sh reports more than 3.2 million registered users. It also offers an AI tutor feature for asking questions about topics.

What the roadmap covers

The map moves roughly from foundations to patterns:

• Performance versus scalability, latency versus throughput, availability versus consistency and the CAP theorem.
• Consistency patterns (weak, eventual, strong) and availability patterns (failover, replication).
• DNS, CDNs, load balancers, reverse proxies and application layer design.
• Databases (SQL and NoSQL, replication, sharding, federation, denormalisation), caching strategies and asynchronism (message queues, task queues, back pressure).
• Communication protocols (HTTP, TCP, UDP, RPC, REST, gRPC, GraphQL).
• Performance antipatterns, monitoring, and cloud design patterns grouped by concern (messaging, data management, design and implementation, reliability).

What it is good for

Auditing your knowledge. The most common problem in self directed learning is not knowing what you do not know. Mark off what you already understand, and the remaining nodes become your study plan. That takes about half an hour and is worth doing before you choose a course or book.

It is also useful for tracking progress across a long preparation and for explaining your learning plan to a mentor.

What it is not

It is not a course. The node explanations are brief and the linked material varies in quality. There is no practice, no feedback and no worked end to end designs. It tells you what to learn and largely leaves the learning to you.

How to use it well

1. Go through every node and honestly mark what you could explain to a colleague.
2. Group the remaining nodes into three or four weekly themes.
3. For each theme, pick one solid source (a book chapter, a course module, or an open source guide) rather than clicking every link.
4. After each theme, practise one design question that uses those concepts.
5. Come back to the map monthly and update it.

Common questions

Is it free? Yes. An account is free and only needed to save progress.

Is it enough to learn system design? No, but it is a very good map to pair with a real learning source.

Pair it with: a readable guide such as Karan Pratap Singh's open source system design course for the explanations, and Designing Data-Intensive Applications for depth.

Best for: finding gaps before choosing what to study, and tracking progress across a long system design preparation.