I am making a webserver with scapy, which is going pretty well. However, it's a pain in the butt for scapy to maintain different connections at the same time. So I want the client to make a persistent connection with the webserver that servers a html page with an image.
I have the client succesfully iniating a tcp handshake and obtaining the html page, however, it opens a new connection to download the image. Which I do not want.
I understand that in HTTP/1.1 it is not necessary to send the keep-alive header, as it's a default. How come Chrome and Firefox still open more connections to download seperate files?
I am not sending a Connection: close header whatsoever, so I think it's weird that they do not maintain the same connection for all files on the webpage.
EDIT: Tried to use the actual Keep-Alive: timeout=n, max = n header. Still no result.
What could be the problem? Feel free to ask for details!
Persistent connections do not forbid to use parallel connections, they only allow to re-use the same connection for more requests. But, with persistent connections you can only do multiple requests within the same connection one after the other. This means to get lots of resources it is usually faster to open multiple connections in parallel and use each of these connections to get multiple resources, e.g. using 4 connections in parallel to get 12 images (3 images with each connection) is faster then getting all the 12 images one after the other using a single connection.