Rui Carmo
Back when I worked in deep packet inspection, there was a parallel thread involving spoofing the relevant bits of TCP window negotiations to mitigate this. I think some of it surfaced in commercial packet core gear, but it bears thinking about whether or not people should be worrying on compensating on the server side (Microsoft’s IIS cheats a bit and shoves as much data as it can towards the device when you try to reach microsoft.com, for instance.)