This one-way audio issue is driving me absolutely insane! I am a network engineer and I am trying to support our remote agents who are all using the WebRTC phone. We are seeing cases where the agent can hear the customer, but the customer cannot hear the agent. I have analyzed the WebRTC diagnostics and I am seeing that the Genesys Cloud media service is sometimes providing an ICE candidate from a different AWS region than the agent’s actual location. This is causing our local firewall to drop the outbound UDP traffic because it looks like a spoofing attack. How do I force the WebRTC phone to only utilize ICE candidates from the agent’s assigned regional edge?
Good afternoon. From an accessibility standpoint, ensuring clear two-way audio is essential for our hearing-impaired callers who rely on high-fidelity sound. While I am not a network specialist, I know that Genesys Cloud handles regionality based on the ‘Media Region’ setting in your ‘Site’ configuration. You should ensure that each remote agent is assigned to a Site that correctly reflects their geographical proximity. Furthermore, ensure that your ‘TURN’ server settings are not overriding the local peer-to-peer negotiation. Providing a standardized regional configuration for every remote worker is the best way to meet our WCAG audio quality benchmarks.
Hey! I’ve been researching the undocumented behaviors of the WebRTC media stack and the ‘Media Region’ is definitely not a guarantee! The platform uses a global ‘Media Optimizer’ that can route traffic to the ‘least busy’ edge, which might be in a different region if your primary region is spiking! If your firewall is that strict, you are going to have a hard time. One thing I found is that if you explicitly whitelist the entire Genesys Cloud CIDR range for your specific AWS region, the firewall will stop being so picky about the source IP. It’s a bit of a security trade-off, but it’s better than agents dealing with one-way audio every ten calls!