Art of Codeartofcode
Lesson 6 of 15

31. Design a URL Shortener

1
The interview prompt
2
31.1 What do you ask first?
3
31.2 How big is this, really?
4
31.3 What are the doors?
5
31.4 What do we store?
6
31.5 What does the whole system look like?
7
31.6 How do you generate the short code?
8
31.7 301 or 302?
9
31.8 How do you keep redirects fast at 12,000 a second?
10
31.9 How do you count clicks without slowing redirects?
11
31.10 What would you say in the last five minutes?
12
Score yourself
13
Variations to practise
14
Sources
15
Code lab
31. Design a URL Shortener

31.5 What does the whole system look like?

3 min read+10 XP

Step 4: High-level design (10 minutes)

Start simple, then add only what the numbers asked for.

the final high-level design: redirects are served from Redis, writes go through ID blocks, clicks are counted asynchronously.

Figure: the final design. Notice the redirect (green) path never touches the ticket server or the analytics pipeline.

Redirect, step by step:

  1. The user opens sho.rt/aB3x9Qz. The load balancer sends it to any app server (all stateless, Chapter 7).
  2. The app server checks Redis for aB3x9Qz. On a hit, done.
  3. On a miss, it reads a read replica of the database, puts the answer in Redis with a TTL, and continues (cache-aside, Chapter 12).
  4. It replies 302 with the long URL, and drops a click event on a queue without waiting.

Create, step by step:

  1. POST /api/links reaches any app server.
  2. The server takes the next ID from its own block of IDs (leased from the ticket server, 31.6), turns it into a 7-character code, and inserts the row into the primary database.
  3. It returns the short URL.

Interviewer pushback:

  • "Walk me through what happens if Redis is down." → Every redirect becomes a miss and goes to the replicas. They must be sized to survive that for a while, with timeouts on the Redis call so a dead cache doesn't make every request hang (Chapter 29).
  • "Why read replicas?" → Misses and the occasional cache restart shouldn't load the primary, which handles all creates. Replicas also give us a failover target if the primary dies (Chapter 14).

Memory code: "Draw it simple, walk it through." Then add only what the numbers asked for: a cache for reads, replicas for safety, a queue for counting.

The interviewer usually picks one or two. A strong candidate offers: "The two interesting parts are generating unique codes and keeping redirects fast. Which would you like?"