While RTP is meant for use in latency-vital scenarios, it isn't going to inherently offer any options that ensure QoS. Rather, it only features the knowledge necessary to make it possible for QoS to get implemented somewhere else inside the stack. This means that even when an attacker managed