How should I balance memory with network latency on a paginated table?

Question

A web application I'm working on has a tool that queries a database and returns up to several thousand rows. When considering pagination, at what point does it make more sense to get the entire result set at once, and store it in memory, as opposed to making asynchronous calls each time a user changes pages on the table?The app is written using Angular 1.X, so I'm trying to use browser memory as sparingly as possible. For the API to return a typical result set takes about 2 seconds. The API is written in node, and the database is PostgreSQL. Should I consider returning a readable stream from the API call?

Answer 1

You mentioned you're using Angular 1.x on the front end and using Nodejs and PostgreSQL on the back-end.

Do you expect your application to scale? If so, I'd avoid using memory. Also, are you running your application behind a load balancer? That can get tricky if using in-memory storage. Also within a docker container if you go that route.

My recommendation would be to pull the data from the back-end as you need it. I find oData calls to be quite speedy and flexible for pagination type calls.

I'm assuming you are using Express or Koa as your middleware and creating restful endpoints? If you're serializing your data using JSON then the user experience should not diminish and you should still be within your 2 second response time.

I'd also recommend moving off of Angular 1.x.