System Design

System design questions ask you to design something big, out loud. Here each design grows like a real system: a simple v1 for 100 users, then v2 and v3, each forced by a real problem.

Want to try it now? Build a design by drag and drop and get instant feedback.

How to use this roadmap

Topics are in teaching order. “Builds on” links show what to know first, even from other tracks.
New? Start at the top. Know some already? Jump to a level, or skip any topic you can already explain.
Short on time? Tick “Asked often only” to see what interviews ask most.
Your progress is saved in this browser.

Best after: SQL. Do SQL foundations first. Design answers keep coming back to databases.

34 topics in 4 levels, about 12 h of reading. With practice, plan 2 to 3 times that. 11 of 34 topics have lessons so far.

34 practice problems from this track’s lessons, in one list.

Foundations

Know this level already? Check what you know: 11 questions, about 6 minutes. Topics you get right are marked Known.

1. Client–server, HTTP and HTTPSNot started
How a browser asks a server for something, and how HTTPS keeps it private.
- My page takes two seconds, and I can’t say where the time goes
Start
15 min read, about 50 min with practice
2. DNSNot started
How a name like example.com turns into a server's address.
- I moved my site, but some users still reach the old server
Start
10 min read, about 40 min with practice
3. REST vs GraphQL vs gRPCNot started
Three ways for programs to talk to each other, and when each fits.
- The phone app makes five calls to draw one screen
Start
20 min read, about 1 h with practiceBuilds on: Client–server, HTTP and HTTPS
4. Latency numbers worth knowingNot started
Rough speeds of memory, disk and network, for quick estimates.
- My page makes 50 database calls and takes four seconds
Start
10 min read, about 40 min with practice
5. Vertical vs horizontal scalingAsked oftenNot started
A bigger server, or more servers?
- On sale day, our one server hit 100% and the shop went down
Start
15 min read, about 1 h with practice
6. Load balancers: L4 vs L7 and algorithmsAsked oftenNot started
Spread traffic across servers, and the ways to pick which server gets a request.
- One server is at 95% while the other two sit idle
Start
20 min read, about 1 h with practiceBuilds on: Vertical vs horizontal scaling
7. Caching: cache-aside, write-through, write-back, eviction, TTLAsked oftenNot started
Keep popular data close and fast, and stop it going stale.
- I added a cache, and now the shop shows the old price
Start
25 min read, about 1 h 15 min with practice
8. SQL vs NoSQLAsked oftenNot started
Tables or documents? How to choose a database for the job.
- Our orders are fine, but the activity feed has two billion rows
Start
20 min read, about 1 h with practiceBuilds on: Tables, rows and keys (SQL)
9. CDNsNot started
Copies of files on servers near users, so pages load fast anywhere.
- Users in Sydney wait three seconds for images from Virginia
Start
10 min read, about 40 min with practiceBuilds on: Caching: cache-aside, write-through, write-back, eviction, TTL
10. Back-of-the-envelope estimatesAsked oftenNot started
Rough maths for users, storage and traffic, done out loud.
- The interviewer asks how much storage a photo app needs, and I freeze
Start
20 min read, about 1 h with practiceBuilds on: Latency numbers worth knowing

Core

1. ReplicationAsked oftenComing soon
Keep copies of data on several servers, for speed and safety.
20 min read
2. Partitioning and consistent hashingAsked oftenComing soon
Split data across servers, and add servers without moving everything.
25 min readBuilds on: Hashing and hash maps (DSA)
3. CAP and PACELCAsked oftenComing soon
When the network breaks, choose between always-correct answers and staying up.
15 min readBuilds on: Replication
4. Consistency modelsComing soon
How fresh is the data you read? From "always the latest" to "eventually".
20 min readBuilds on: CAP and PACELC
5. Queues and pub/subAsked oftenComing soon
Hand work to a queue so users don't wait, and send one event to many listeners.
20 min read
6. Rate-limiting algorithmsAsked oftenComing soon
Cap how many requests each user can make, fairly.
20 min read
7. API design: pagination, idempotency, versioningAsked oftenComing soon
Design endpoints that page through big results and are safe to retry.
20 min readBuilds on: REST vs GraphQL vs gRPC
8. Authentication and authorization: sessions, JWT, OAuth2, OIDCComing soon
Who are you, and what may you do? Sessions, tokens and "sign in with Google".
25 min readBuilds on: Sessions, JWT and OAuth2 (Frameworks)
9. Object storageComing soon
Store big files like photos and videos cheaply, outside the database.
10 min read
10. Search and inverted indexesComing soon
Find records by their words fast, the way search engines do.
20 min read
11. Unique ID generationComing soon
Make IDs that never clash, even across many servers.
15 min read
12. ProxiesComing soon
Middlemen between clients and servers, and what each kind is for.
10 min read

Advanced

1. Distributed transactions: two-phase commit and sagasComing soon
Keep data correct when one action touches several services.
25 min readBuilds on: Transactions, ACID and isolation levels (SQL)
2. Consensus (Raft), leader election and distributed locksComing soon
How servers agree on one leader or one value, even when some of them fail.
25 min read
3. Event sourcing and CQRSComing soon
Store every change as an event, and keep reads and writes in separate models.
20 min readBuilds on: Queues and pub/sub
4. Batch vs stream processingComing soon
Process data in big scheduled chunks, or as soon as it arrives.
20 min read
5. Observability: logs, metrics, tracesComing soon
See what a live system is doing, and find out why it is slow.
20 min read
6. Resilience: timeouts, retries with jitter, circuit breakers, bulkheadsComing soon
Stop one slow service from taking everything else down with it.
20 min read
7. Monolith vs microservicesComing soon
One app or many small ones, and the real costs of splitting.
15 min read
8. Multi-region and disaster recoveryComing soon
Keep running when a whole data centre goes down.
20 min readBuilds on: Replication

Interview-ready

1. The system design interview frameworkAsked oftenComing soon
A step-by-step plan for the 45-minute design round.
20 min readBuilds on: The system-design-round method and its time budget (Interview Playbook)
2. Case studies: URL shortener, rate limiter, chat, news feed, video, ride-hailing, payments and moreAsked oftenNot started
Classic designs built from v1 to v3, each step forced by a real problem.
- Design a URL shortener
Start
90 min read, about 2 h with practiceBuilds on: Back-of-the-envelope estimates · The system design interview framework
3. Low-level design: SOLID and the patterns that matterComing soon
Design clean classes for a small system: the other kind of design round.
25 min readBuilds on: Classes: inheritance, MRO, super() and dunder methods (Python)
4. LLD problems in Python: parking lot, elevator, vending machine, expense splitter and moreComing soon
Classic class-design questions, solved in Python.
45 min readBuilds on: Low-level design: SOLID and the patterns that matter