DPI and the QoS in the Russian Federation
It's six in the morning, I haven't gone to bed yet.
P.S. My notes are not recommendations or technical advice. These are my thoughts, not a call to action. For those who read between lines: under "Wars" I mean exclusively unequal fight with packet loss and QoS curves on the mainways. It's not Hyde to bypass anything, but the Chronicles trying to force UDP to work where he doesn't want to work.
At the moment, I have several servers besides these and (XD) raised on them, while looking at how actively we are
in the network i thought we should look for a fallback solution. Of the most well-known (theoretically working) protocols, we just have Hy2, which was successfully installed and working stable on my wifi network with "half a bump" (maybe I misinterpreted this phraseology into English), but obviously not on my mobile-network. Here's what you need to make it work on a
network as well... As a "self-proclaimed network engineer," I decided to try to get it.
In theory, the problem is that despite the fact that SNI costs VK, which actively drives QUIC and HTTP/3 (hy2 works on it), the problem remains in MTU and SNI which does not calls.vk.com, but just vk.com which rolls on
because I have it mostly xHTTP, gRPC and raw TCP.
Purely theoretically... The solution has been found, but in fact it is still unknown. I'll try tomorrow.
So far, judging by the log, the problem is not exactly in the patterns of behavior and port-hopping, but just in the lack of disguise. After all, UDP traffic is moderated much more rigidly than regular TCP.
Of course, all this is for the privatization of my personal informationn...