System design interview question

Design Pastebin

Pastebin, GitHub Gist, any “share this text” tool: “design Pastebin” is a close cousin of the URL shortener with two twists. The content can be large, and one link must never let anyone find other people's pastes. It tests identifiers, storage tiers and read-heavy traffic. Here is how to walk into it, and then how to practise it until the answer is yours.

Practise Pastebin

Multiple choicePick the design from four
Fill the gapsFill the gaps in one drawn for you
Build from scratchBuild it on a blank canvas

Practise it on a canvas

Free with an account.

The question, as an interviewer would set it

A team is building a paste-sharing tool. A user submits a block of text — anything from a few lines to a large log file of several megabytes — and gets back a link that others can open to read it, unchanged. Each paste needs its own identifier that cannot be guessed or stepped through in sequence, so one paste's link never reveals another's. Pastes are written once and never edited, are read far more often than they are written, and most traffic lands on a small number of popular pastes. Storing many large text blobs cheaply matters, and the tool should stay fast when a paste goes viral.

What it is really testing

Every phrase that matters in the brief is a requirement in disguise. Spot them before you draw a single box.

“cannot be guessed or stepped through in sequence”
An identifier that cannot be guessed or enumerated in order.
“Storing many large text blobs cheaply matters”
Object storage for large, write-once blobs.
“stay fast when a paste goes viral”
An in-memory cache for the hottest pastes.

Back of the envelope

New pastes: 1 million a day → ~12 a secondAn assumption to say out loud. If the interviewer has a number in mind, they will give it to you.
Size: ~10 KB on average, some several MB → ~10 GB a dayFar too much to keep in rows of a database comfortably.
Five years: ~18 TBMost of it text nobody reads again after the first week.
Reads: ~10 per paste → ~120 a secondThe average. A paste that goes viral takes thousands a second on its own.

Questions to ask first

How large can a paste be?
Can pastes expire, or be deleted by their author?
Public, unlisted or private — who may read one?
How many pastes a day, and how many reads each?
Is there processing on the server, such as syntax highlighting?

The trade-offs you will be pushed on

1.What makes a paste's link impossible to guess or step through?
2.Where do megabytes of text live, if not in the database?
3.One paste goes viral. What keeps it fast?
4.Most pastes go quiet after a week. What happens to them?
5.What has to stay true when 50,000 people read the same paste in one second?

Now answer it

The same design always gets the same score, with the reason behind every requirement — then defend it against the trade-offs above.