How to properly organize bff servers for a websocket application?

Question

Software design

I am writing a microservice project for educational purposes, its scheme is presented here:



On the left - a front-end application that works with a server via a websocket. It contains "microfrontends", each of which has access to a websocket and communicates with the server. A separate bff is created for each microfrontend, and it already aggregates data from other services via http. I had a number of questions, the answers to which I unsuccessfully try to google.

The most important thing is whether the organization in the picture is correct, or is it not done in real combat? Is it correct that the websocket server, after receiving the message, starts communicating directly via http with bff servers? Is it possible to develop a separate microfrontend locally without waiting for the deployment of the entire backend structure? And what about avotrization,