Is my estimate of the minimum payload size for a fast transfer over HTTP correct?

Question

Computer networks

I'm trying to figure out a reasonable minimum useful size in bytes for an HTTP request based on the fact that the HTTP request goes over TCP packets.
Therefore, fast transmission means that 1 HTTP_request fits into 1 TCP packet.
On the one hand, the TCP packet must be less than 65kB (IP frame limit), and on the other hand, the size of the TCP packet is the TCP HEADER (20 ... 60 bytes) + TCP BODY of any length, but in total no more than MTU .
The MTU values, as I understand it, may be different on the route, but according to RFC 879, the minimum size is 576 bytes.
Therefore, it turns out that the minimum payload size in an HTTP request is 576 bytes minus 60 bytes per TCP packet header, minus the length of the HTTP header.
And since the minimum HTTP header, according to my very subjective feelings, is about 300-400 bytes on average, it turns out that about 100-200 bytes remain for useful data.

Are there any errors in reasoning?

Answer 1

The Big Question™: What do you want to do? Why all this reasoning? Interesting theories? Then go on.

Answer 2

there is. from the second sentence onwards

Answer 3

> IP frame limit
frames are operated by the link layer protocol, ip - network, tcp - transport.
so the expression ip-frame is incorrect.
> but in total no more than MTU
MTU is the concept of the link layer, so at least the IP header must be enabled.
Well, in general, there is the concept of MSS (Maximum Segment Size):
MSS = MTU - sizeof (TCPHDR) - sizeof (IPHDR)
> the minimum HTTP header, according to my very subjective feelings, is on average about 300-400 bytes
much less.