System requirements

Functional:

• tiny URL should be 32 amount of characters
• tiny URL should expire after 6 months if it is not used
• same URL should not be provided for different long URLs
• URL should be ready to access as soon as the URL is generated

Non-Functional:

• service is highly available
• if there are customers from different regions it is provided
• the service does not have storage issue

Capacity estimation

• it will have daily 1000 write requests and 500.000 read requests
• for 1000 write requests, if average number of chars are 150, then 1000 * 150 byte storage is required

API design

Write:

POST:

request: co.io/url

responseCodes: 200 (OK), 500 (INTERNAL_SERVER_ERROR), 400 (BAD_REQUEST)

response: the shorten URL

GET:

request: co.io/url/{URL}

responseCodes: 200, 500, 400, 404

response: the long url

Database design

columns: id, short_url, long_url, created_at, last_access_date, tombstoned (true, false)

High-level design

• client will cal load-balancers and they will call the servers
• then servers will access to cache or database based on some checks

Request flows

• client will make POST or GET request
• for GET
  • if the URL is in cache return it directly
  • if not in cache write to cache and write to the database
• for POST
  • write to cache and database

Detailed component design

• For client I will use ReactJs and it will run in a standalone web server
• backend will run in standalone server
• cache will be distributed cache and it will be write-thgrough cache
• backend will generate short options and will check the database if there is a collision or not
  • if there is a collision it will generate as long as collision does not happen (not sure which algorithm to use here)
• backend will use Kotlin and SpringBoot
• Authentication will not be required
• the request URl should not be longer than 500 characters
• cache will have TTL of 6 hours
• database will be Postgress

Trade offs/Tech choices

Explain any trade offs you have made and why you made certain tech choices...

Failure scenarios/bottlenecks

Try to discuss as many failure scenarios/bottlenecks as possible.

Future improvements

What are some future improvements you would make? How would you mitigate the failure scenario(s) you described above?