Loading...
Estimate the scale of the system you are going to design...
1 million DAU at 10 req / day = 1000 req/s
Define what APIs are expected from the system...
POST /urls { longUrl: str } - creates a short url for a given long URL
GET /urls/:id - returns a long URL for a short URL
GET /urls/:id/stats - returns stats for a given url
Defining the system data model early on will clarify how data will flow among different components of the system. Also you could draw an ER diagram using the diagramming tool to enhance your design...
Url [
userId
shortUrl
longUrl
]
User [
id
name
]
You should identify enough components that are needed to solve the actual problem from end to end. Also remember to draw a block diagram using the diagramming tool to augment your design. If you are unfamiliar with the tool, you can simply describe your design to the chat bot and ask it to generate a starter diagram for you to modify...
Write service will generate a shortUrl for a given url and return the object by way of a hashing function. If we use a 12 character shortUrl and base64 encode it, we will have 12^64 combinations. Collisions can be handled by using a counter that increments for each new url and is stored in redis and uses the counter to generate the shortUrl instead of the longUrl.
Read service will convert a shortUrl to a longUrl using the url database. It will also increment the view count for a given url in redis.
Stats service will read view counts from redis using the url's id as the hash key.
Explain how the request flows from end to end in your high level design. Also you could draw a sequence diagram using the diagramming tool to enhance your explanation...
Creating a shortUrl. Url goes through API GW, to write service, write service increments counter, hashes it and returns.
Resolving a short Url goes to read service. Read service looks up the url in the url database and returns the hashed short url. It also uses the url's id to increment the view count in the stats redis.
Fetching stats goes to stats service and fetches a url's stats from redis via the url in the params.
Dig deeper into 2-3 components and explain in detail how they work. For example, how well does each component scale? Any relevant algorithm or data structure you like to use for a component? Also you could draw a diagram using the diagramming tool to enhance your design...
Write service should scale fine.
Read service will easily scale up to 1000 requests per second.
Stats service will easily scale up to 1000 req/s.
All components could be vertically scaled to handle more load.
Explain any trade offs you have made and why you made certain tech choices...
Stats are available for total time only. If we wanted 'over time' stats, we would have to replace redis with a time series dataset.
Also, the counter creates predictability in the hashes. We would need to obfuscate the url if we wanted to remove this issue.
Try to discuss as many failure scenarios/bottlenecks as possible.
We could use failover for the api gateway and service instances.
Databases could use high availability with a master/slave for the url database and redis.
What are some future improvements you would make? How would you mitigate the failure scenario(s) you described above?
Store stats outside of redis in a timeseries database so we could show stats over time.
Use a obfuscation function to decrease predictability.