Trade offs/Tech choices

Database SQL vs NoSQL - I choose NoSQL based on following reasons:

High scalability, can do horizontal scaling.

The data is just a pair of URLs, no stroing relation with other data.

Supports fast read and write.

More affordable for large volumn data.

Cache Stratergy:

Cache aside: If there's a cache miss, get data from DB and add an entry for the cache, then return the result.
Pros: Very fast if we hit the cache, but it depends on how the invalidation mechanism goes.
Cons: A cache miss will introduce extra steps, which costs more time.
Write through: Whenever a new entry coming in, it writes into the cache and database simultanously.
Pros: High consistency.
Cons: Slower than cache aside.
Write back: write to cache first, then async write to DB.
Pros: high write speed for write-heavy system.
Cons: Data might be inconsistency, and if cache server down before sending the data to DB, it will cause data loss.

Based on the pros and cons, I'd like to use cache aside combine with a suitable invalidation mechanism in this system.

Cache invalidation mechanism:

Least-recently use
Least-frequently use

It kind of depends on the user habit, if the analytics shows that some of the short_urls are very often to request, we might want to apply Least-frequently use. If the analytics shows that user prefer to request recently generated urls, then we might want to apply LRU.

Submissions

Displaying Solutions with Scores Ranging from 1 to 10

Submissions

Displaying Solutions with Scores Ranging from 1 to 10