A scalable taxi dispatch system
Last semester, I took a class in distributed systems. One of the assignements was to design and implement a system which integrates as a component in the backend of an uber-like ( mobile taxi dispatch ) system. As specifications go, it was about as vague as it gets, but the words ‘middleware’ and ‘message-queue’ were said quite a bit.
Ubers current system they use to connect and organise mobile network of consumers and drivers is called DISCO, which is an anagram for something about dispatch optimisation. Theres a pretty detailed blog post about it here, but essentially its supported by a giant infrastruture and terabytes of data.
For this piece of coursework, I essentially developed my own version of this system using gRPC and a the chord protocol (a distributed hash table). Alongside each piece of coursework we must submit a report detailing how to build and run the program, along with our design decisions and possible improvements. Below is my offical report for the coursework, and I have open sourced the project also:
https://github.com/NFS002/Cabaret
