gvinicius / ip_tracker Goto Github PK

Track and rank requesting IPs with Ruby.

• What would you do differently if you had more time? If had more time, I would also write a small Sinatra Web API just for providing a more real-life example and shipping this tracker as ruby gem to be imported in the Sinatra API.

• What is the runtime complexity of each function? request_handled(ip_address]) => O(1)

count_entries(ips) => O(n2)

top100() => O(log n)

clear() => O(n)

• How does your code work? My code work as object-oriented entity that is stores all requested ips in an attirbute called ips, and has instances methods implementing each one of the requirements. Besides, the method for ranking the 100-most frequent ips follows a concurrent approach in batches to deal with millions records in a quikier way. Moreover, added in auxiliary method to the raking for a better separation of concerns.

• What other approaches did you decide not to pursue? I decided to not pursue an entirely concurrent-approach, which means that I could have made the ip storing and clearing as concurrent as the raking method is, but the gains in doing it seemed to be not much significant.

• How would you test this? I would test this by writing benchmark comparisions. I could implement tests like that by writing specific test cases with RSpec.

# install ruby 3.0.2

bundle install

bundle exec rspec spec/tracker_spec.rb # for testing

irb
require_relative './tracker'
tracker = Tracker.new
tracker.request_handled('145.87.2.109')
tracker.top100()
tracker.clear()
tracker.top100()