Hi, I'm running FreeSwitch on an AWS EC2 with config as per the wiki EC2 page. external_rtp_ip & bind_server_ip are both set to the box's public AWS Elastic IP - as is ext_rtp_ip on all profiles. I'm running a WebRTC gateway also in AWS. I can see that SDPs from Freeswitch to the gateway all include Freeswitch's public IP as desired, however when the call is answered Freeswitch always sends RTP traffic from it's private IP. Then a second or 2 later Freeswitch sees incoming RTP from the gateway to it's public IP address and RTP auto swtiching behaviour kicks in, resulting in a line like this in FS logs: switch_rtp.c:6954 Auto Changing audio port from 172.31.x.x:31196 to 52.17.x.x:31196 As it happens, the gateway is able to receive media from either private or public IP as the box is in the same subnet, but in around 30% of cases at the point this switching occurs the call audio drops. (I'm guessing the gateway has issues stitching together the rtp streams from 2 different sources.) I've tried experimenting with the disable_rtp_auto_switch parameter, but seems to me a cleaner solution would be to have Freeswitch send RTP from it's public IP in the first place, and I can't understand why it's not doing that given it's negotiated to use the public IP in it's SDP. Can anyone provide any explanation for the behaviour I'm currently seeing, or suggest how I can get FS to set the initial outgoing RTP port correctly? Thanks, Sean.