I worked on personal project to learn how to implement the back end of a stock trading web app. I decided on a relational database since stock data would benefit from transactional integrity. After generating and seeding 10 million fake records, I was able to choose PostgreSQL as it had the fastest query times for my data.
After deploying the service on an AWS EC2 T2-micro server, I used Nginx to load balance between four instances in order to bring the RPS from 600 to the requested 1000.
Chris Doornbos 