GSMA: Network Management of Encrypted Traffic [pdf]

Mobile operators directly encouraging cleartext so they can fuck with traffic? That's the exact case made for encryption. (You can see how that went down at IETF 92.)

Interesting note about the Google data compression proxy. Can we get that bug fixed?

Again, an indicator that SNI really needs to be at the very least not cleartext in TLS 1.3.

It would be interesting to go through this and highlight everything that is or would be (for the hypothetical examples) or only serves to enable net neutrality violations.

The real kernel of technical interest seems to be the open problem of how to do QoS for multiple flows of different types being carried through the same encrypted tunnel.