Abstract:

Allocations and related lack of spatial locality can be DAMN EXPENSIVE

Quote:

During the times of Ancient Gamers (also known as times of Atari, ZX Spectrum, and Commodore64), CPUs were running at clock speed of a single-MHz (and one operation usually took several clocks). At the same time, memory run at about the speed comparable to that of CPUs.

Quote:

Encoded data size is especially important for Client-Server communications, and less important (within reason) for Server-to-Server ones and for locally-stored format.

Another Quote:

On the other hand, due to being optimized for CPU operation, Flatbuffers are NOT optimized space-wise; even compared to not-so-optimal-space-wise Google Protocol Buffers, Flatbuffers can lose additional 1.5x in size

Quote:

I am not saying that this architecture is the only viable one, but it does work for TCP for sure (and performs reasonably well too)

Another Quote:

The whole task of optimizing performance beyond, say, 20-50K packets/second per box tends to be Quite Elaborated, and involves quite a few things which are platform- and hardware-dependent.

Quote:

TCP is a byte stream, the whole byte stream and nothing but the byte stream

Another Quote:

if all packets reach Client, there isn’t that much difference between TCP and UDP