bdahon/bcm

This is a solution to the exercise: https://github.com/BCM-ENERGY-team/bcm-backend-interview

The solution is in the form of a simple nodejs server with one callable route.

Usage

To run:

$ npm install
$ npm start

This will launch a server on port 8080.

To run the tests:

$ npm run test

Notes

• to load a different route file, do so in .env and restart the server
• this solution also handles [Bonus #1]
• @todo: tests will fail if server is running. They cannot run at the same time.

Architecture

• This is a simple nodejs server which uses express
• There are a few config options in the .env file
• The tests are run with mocha
• We did not use a database as the quantity of data for the exercise is easily handled by a slow computer
• The response of the route is of the form

[
    {
        operatingAirlines: [AF, BA],
        sellingAirline : AF
        from: CDG,
        to: JFK,
        stopOvers: [HKG]
    },
    ...
]

Bonus

Cache and reuse results

A solution could be to use a simple key-value cache such as redis. It can be integrated into nodejs in a few lines and is easily shared between multiple instances. Caching per {airlineCode-from-to} triplet would be the obvious first choice.

Endpoint security

Endpoint security can be handled in multiple ways and it all really depends on who will use the API.

If the API is used by a known small group of users, the IAM solutions by google cloud or AWS should be sufficient.

For a public url without an explicit whitelist my preference is for a json web token authentication combined with a saas solution such as auth0 or equivalent. This can be handled at the API level or at the API gateway/ reverse proxy level. The advantage with the first solution is that there are less servers to maintain/ configure but each new API built in this way will require its own security configuration. However, the second solution decouples authentication/ authorisation from the API and allows for greater flexibility on how the API gets used (internally/ externally).

Rate limiting

For API route rate limits, there are multiple nodejs libraries such as express-rate-limit which can control the rate. Alternatively, AWS or google cloud will allow you to configure the ACL.