HTTP/1.1 can reuse a TCP connection with keep-alive, but responses on one connection arrive in request order. Browsers historically opened several connections per origin to fetch resources in parallel.

HTTP/2 represents messages as binary frames. Several independent streams share one connection, so a slow application response does not have to block every response behind it.

HTTP/2 also compresses headers, which helps when many requests repeat the same cookies and metadata.

The application semantics do not change. A GET is still a GET; a 404 is still a 404. Browsers and servers negotiate the protocol, usually without changes to your routes.

Because every stream still shares one ordered TCP connection, packet loss can temporarily pause them all. HTTP/3 addresses that transport-level limitation.